tag:blogger.com,1999:blog-9622110573146002032024-03-14T19:02:18.813+09:00Motoo Tange's BlogMotoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.comBlogger344125tag:blogger.com,1999:blog-962211057314600203.post-78206898880423555672023-05-30T18:00:00.001+09:002023-05-30T18:00:00.135+09:00数学リテラシー2(第1回)<p> <span style="color: orange; font-family: inherit; font-size: 16px; orphans: 2;">[場所:3A308(火曜日15:15〜16:30, 16:45〜18:00</span><span style="color: orange; font-family: inherit; font-size: 16px; orphans: 2;">)]</span></p><div style="-webkit-text-stroke-width: 0px; font-family: &quot; font-size: 16px; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; orphans: 2;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div></div></div></div></div></div></div></div></div></div></div></div><div style="-webkit-text-stroke-width: 0px; font-family: &quot; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="font-size: 16px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/22/literacy2.html">数学リテラシー2のHP</a></div><div style="font-size: 16px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="font-size: 16px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">数学リテラシー2が始まりました。</div><div style="font-size: 16px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><b style="font-size: x-large;"><br /></b></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><b style="font-size: x-large;"><u>実数直線の部分集合</u></b></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">今回は 実数直線上の部分集合 と数列の収束の定義について行いました。</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">実数全体のことは </span><span lang="EN-US" style="font-size: 16px;">${\mathbb R}$ </span><span style="font-size: 16px;">と書き表します。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">また、有理数全体のことは ${\mathbb Q}$ と書き表します。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">まず実数上の部分集合について考えます。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">まずは部分集合として区間というものを定義しておきましょう</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><b style="font-size: 16px;"><br /></b></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><b style="font-size: 16px;">例1.3.1</b></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span lang="EN-US" style="font-size: 16px;">$\{x\in {\mathbb R}|a<x<b\}$ </span><span style="font-size: 16px;">を開区間といい、</span><span lang="EN-US" style="font-size: 16px;">$(a,b)$ </span><span style="font-size: 16px;">と書き</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span lang="EN-US" style="font-size: 16px;">$\{x\in {\mathbb R}|a\le x\le b\}$ </span><span style="font-size: 16px;">を閉区間といい、</span><span lang="EN-US" style="font-size: 16px;">$[a,b]$ </span><span style="font-size: 16px;">と書き</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">そして半開区間として</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$\{x\in {\mathbb R}|a\le x<b\}$</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$\{x\in {\mathbb R}|a<x\le b\}$</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">をそれぞれ </span><span lang="EN-US" style="font-size: 16px;">$[a,b)$ </span><span style="font-size: 16px;">と </span><span lang="EN-US" style="font-size: 16px;">$(a,b]$ と表しま</span><span style="font-size: 16px;">す。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">これらの記号を使っていきましょう。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">また、この記号は</span><span lang="EN-US" style="font-size: 16px;"> $a,b$ </span><span style="font-size: 16px;">が </span><span lang="EN-US" style="font-size: 16px;">$-\infty$
</span><span style="font-size: 16px;">や </span><span lang="EN-US" style="font-size: 16px;">$\infty$ </span><span style="font-size: 16px;">であっても定義可能です。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">ただし、</span><span lang="EN-US" style="font-size: 16px;">$\pm\infty$ </span><span style="font-size: 16px;">は実数ではないので、</span><span lang="EN-US" style="font-size: 16px;">$(-\infty, 2]$ </span><span style="font-size: 16px;">のように使いますが、</span><span lang="EN-US" style="font-size: 16px;">$[-\infty,2]$ </span><span style="font-size: 16px;">のようにはならないので注意してください。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">次に実数の部分集合 </span><span lang="EN-US" style="font-size: 16px;">$A$ </span><span style="font-size: 16px;">が有界であるということについて定義して行きたいと思います</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><b style="font-size: 16px;"><br /></b></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><b style="font-size: 16px;">定義1.3.1</b></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span lang="EN-US" style="font-size: 16px;">$A$ </span><span style="font-size: 16px;">を </span><span lang="EN-US" style="font-size: 16px;">${\mathbb R}$
</span><span style="font-size: 16px;">の部分集合とします。実数</span><span lang="EN-US" style="font-size: 16px;"> $M$ </span><span style="font-size: 16px;">に対して任意の実数 </span><span lang="EN-US" style="font-size: 16px;">$a\in
A$ </span><span style="font-size: 16px;">に対して </span><span lang="EN-US" style="font-size: 16px;">$a\le M$ </span><span style="font-size: 16px;">が成り立つとき </span><span lang="EN-US" style="font-size: 16px;">$A$ </span><span style="font-size: 16px;">は</span><b style="font-size: 16px;">上に有界</b><span style="font-size: 16px;">といい、このような</span><span lang="EN-US" style="font-size: 16px;"> $M$ </span><span style="font-size: 16px;">のことを</span><span lang="EN-US" style="font-size: 16px;"> $A$ </span><span style="font-size: 16px;">の<b>上界</b>と言います。次に実数 </span><span lang="EN-US" style="font-size: 16px;">$m$ </span><span style="font-size: 16px;">に対してある </span><span lang="EN-US" style="font-size: 16px;">$a\in A$ </span><span style="font-size: 16px;">に対して </span><span lang="EN-US" style="font-size: 16px;">$m\le a$ </span><span style="font-size: 16px;">が成り立つとき</span><span lang="EN-US" style="font-size: 16px;"> $A$ </span><span style="font-size: 16px;">は</span><b style="font-size: 16px;">下に有界</b><span style="font-size: 16px;">といい、このような </span><span lang="EN-US" style="font-size: 16px;">$m$ </span><span style="font-size: 16px;">のことを </span><span lang="EN-US" style="font-size: 16px;">$A$ </span><span style="font-size: 16px;">の</span><b style="font-size: 16px;">下界</b><span style="font-size: 16px;">と言います。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">集合 </span><span lang="EN-US" style="font-size: 16px;">$A$ </span><span style="font-size: 16px;">が上に有界かつ下に有界のとき</span><span lang="EN-US" style="font-size: 16px;"> $A$ </span><span style="font-size: 16px;">は<b>有界</b>と言います。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">つまり </span><span lang="EN-US" style="font-size: 16px;">$A$ </span><span style="font-size: 16px;">が有界のとき上界と下界存在するので ある実数の有限の部分に収まっている部分集合と言うこともできます。集合の言葉で書けば、ある実数 $m,M$ が存在して</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$$A\subset \{x|m<x<M\}$$</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">ということができます。もちろん十分大きい実数 $K$ を取れば、</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$$A\subset \{x|-K<x<K\}$$</span></div><div><span style="font-size: 16px;">とすることもできます。</span></div></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><b>例1.3.2</b></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">(1) $A=(-\infty,2]$ は上に有界だが、下に有界ではない。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">(2) $A=\{x\in {\mathbb Q}|-2<x<3\}$ は有界な部分集合である。</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">(2) の部分集合は $-2$ から $3$ までの有理数ではない数は全て入っていませんので${\mathbb R}$ の区間ではありませんが ${\mathbb R}$ の部分集合にはなっています。</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">次に、</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span lang="EN-US" style="font-size: 16px;">${\mathbb R}$ </span><span style="font-size: 16px;">の部分集合 $A$ に対して $A$ の上界全体を $U(A)$ とかき、</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span lang="EN-US" style="font-size: 16px;">${\mathbb R}$ </span><span style="font-size: 16px;">の部分集合 $A$ に対して $A$ の下界全体を $L(A)$ とかきます。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">この時、$U(A)\neq\emptyset$ であることは $A$ が上に有界であること同値であり、</span><span lang="EN-US" style="font-size: 16px;"> $L(A)\neq \emptyset$ </span><span style="font-size: 16px;">であることは $A$ が下に有界であることと同値になります。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">ここで実数の連続性公理について説明します。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><b>実数の連続性公理</b></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$A$ が実数の部分集合として $U(A)\neq \emptyset$ ならば、つまり $A$ が上に有界であるならば</span><span lang="EN-US" style="font-size: 16px;"> $U(A)$ </span><span style="font-size: 16px;">は最小値を持つ。また、もし</span><span lang="EN-US" style="font-size: 16px;"> $L(A)\neq \emptyset$ なら $L(A)$ </span><span style="font-size: 16px;">は最大値を持つ。このことを<b>実数の連続性公理</b>という。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">このことは実数の性質、もしくは実数を定義づける性質の1つになるのですが、性質として認めておくことにします。有理数はもちろん実数とは違いますが、連続性公理の観点からもそれとは異なることを最後に説明します。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">例えばこの性質を使うことで どのような実数の性質が導かれるか見てみましょう。例えば $\sqrt{2}$ に近づくような数列 $1,1.4,1.41,\cdots$ を取ります</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">この時 $A$ を</span><span style="font-size: 16px;">下のように取ります。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$$A=\{1, 1.4, 1.41, 1.414, \cdots\}$$</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">のようにとります。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">この集合 $A$ に対して $2$ が上界となるので</span><span style="font-size: 16px;">上に有界、つまり 実数の連続性公理の最初の仮定を満たします。ですのでこの集合 $A$ に対して $U(A)$</span><span lang="EN-US" style="font-size: 16px;"> </span><span style="font-size: 16px;">は最小値を持ち、</span><span style="font-size: 16px;">その最小値は実際に $\sqrt{2}$ になります。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">上の「</span><span style="font-size: 16px;">最小値を持つ」について説明します。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span id="docs-internal-guid-8920239b-7fff-ab03-d4f1-f34d9b29e0d2"><p dir="ltr" style="line-height: 1.38; margin-bottom: 0pt; margin-top: 0pt;"><span style="font-family: Arial; font-size: 12pt; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; vertical-align: baseline; white-space: pre-wrap;"><br /></span></p><p dir="ltr" style="line-height: 1.38; margin-bottom: 0pt; margin-top: 0pt;"><span style="font-family: Arial; font-size: 12pt; white-space: pre-wrap;">例えば集合 $(0,1)$ を考えてみます。</span></p><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-family: Arial; font-size: 12pt; white-space: pre-wrap;">この時、<u>この集合の最大値をもちません。</u></span></div><br /><p dir="ltr" style="line-height: 1.38; margin-bottom: 0pt; margin-top: 0pt;"><span style="font-family: Arial; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; vertical-align: baseline; white-space: pre-wrap;">しかし、集合 $(0,1]$ を考えた場合、<u>この集合には</u></span><span style="font-family: Arial; white-space: pre-wrap;"><u>最大値をもち</u> $1$ となります。 </span></p><p dir="ltr" style="line-height: 1.38; margin-bottom: 0pt; margin-top: 0pt;"><span style="font-family: Arial; white-space: pre-wrap;"><br /></span></p><p dir="ltr" style="line-height: 1.38; margin-bottom: 0pt; margin-top: 0pt;"><span id="docs-internal-guid-a4293296-7fff-220e-4b7b-ef6a2e7a1abe"><span style="font-family: Arial; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; vertical-align: baseline; white-space: pre-wrap;">大事なことは、<u>最大値というのは。その集合に含まれていないといけないということです。 </u></span></span></p><div><span style="font-family: Arial; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; vertical-align: baseline; white-space: pre-wrap;">最小値についても同じことです。</span></div><div><span style="font-family: Arial; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; vertical-align: baseline; white-space: pre-wrap;"><br /></span></div></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span id="docs-internal-guid-61cac5a0-7fff-0167-cd07-3a198919cd44"><div><span style="font-family: Arial; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; vertical-align: baseline; white-space: pre-wrap;">最大値を持つというのは、最大値が存在する、と言うこともあります。</span></div><div>$A$ の最大値のことを $\max A$ と書き、$A$ の最小値を $\min A$ と書きます。</div></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><b><u><span style="font-size: x-large;">上限・下限</span></u></b></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">ここで上限と下限について定義をしましょう。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$A$ が空集合でなく、 $U(A)$ </span><span style="font-size: 16px;">が空集合でないとき、$U(A)$ </span><span style="font-size: 16px;">の最小値(つまり $\min U(A)$ )は連続性公理から必ず存在し、その最小値のことを $A$ の<b>上限</b>と言います。それを $\sup A$ と書きます。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="font-size: medium; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="font-size: medium; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$A$ が空集合でなく、 $L(A)$ </span><span style="font-size: 16px;">が空集合でないとき、$L(A)$ </span><span style="font-size: 16px;">の最大値(つまり $\max L(A)$ )は連続性公理から必ず存在し、その最大値のことを $A$ の<b>下限</b>と言います。それを $\inf A$ と書きます。</span></div><div style="font-size: medium; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="font-size: medium; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$U(A)$ が存在しなければ、$\sup A=\infty$ と書いたり、$L(A)$ が存在しない場合、$\inf A=-\infty$ と書くことがあります。</span></div><div style="font-size: medium; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">例</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">$\max(0,1)$ は最大値はありませんでしたが、$\sup(0,1)$ は存在して $1$ となります。なぜなら、$U((0,1))=[1,\infty)$ であるから、$\sup (0,1)=\min([1,\infty))=1$ となるからです。</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">つまり、集合が上に有界ではない状況で最大がない場合は $\sup$ も $\max$ も同じ意味で最大が存在しないか、$\infty$ になるかとなります。</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">しかし、<u>上に有界であるとき、最大値が存在しない場合でもその最大に当たる値が上限として定義されているということになります。</u></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">それで上のような、上界の最小値というややこしい定義になってると言ってもよいかもしれません。</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">下限についても同じです。</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">また、$U([0,1])=[1,\infty)$ でもあるから、やはりこのときでも、</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">$\sup[0,1]=1$ であり、<u>最大値が存在するなら、上限はその最大値と一致します。</u></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">補題1.3.1</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">$A\subset{\mathbb R}$ とし、$A\neq \emptyset$ とする。$\sup A=\alpha$ なら</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">任意の $\epsilon >0$ に対して、 ある $a\in A$ が存在して、$\alpha-\epsilon<a$ が成り立つ。</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">(証明) 任意の $\epsilon>0$ に対して、$\alpha-\epsilon$ は $A$ の上界ではありません。なぜなら $\alpha$ が上界の最小値だからです。よって、$\alpha-\epsilon$ は上界の条件を満たさないことになります。$x$ が上界であるとは、任意の $a\in A$ に対して、$a\le x$ ですから、このことを否定することで、ある $a\in A$ が存在して、$\alpha-\epsilon<a$ を満たすことになります。 </div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; text-align: right;">$\Box$</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><b><u><span style="font-size: x-large;">${\mathbb Q}$ は連続性公理を満たさないこと</span></u></b></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span>上で ${\mathbb R}$ は連続性公理を満たさないと言いましたが、そうではない性質を持つ集合もあるので書いておきます。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span>それで上に書いた有理数 ${\mathbb Q}$ は連続性公理を満たしません。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span>先ほどの例を用います。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$$A=\{1, 1.4, 1.41, 1.414, \cdots\}$$</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">とします。このとき、もし ${\mathbb Q}$ が連続性公理を満たすなら、この $A$ は上に有界であるから上限が存在することになってしまいます。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">しかし、 $\sqrt{2}\not\in {\mathbb Q}$ ですから、これは矛盾するわけです。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">$\sqrt{2}$ が有理数ではない証明は省略します。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;">よって、${\mathbb Q}$ は連続性公理を満たさないということになります。</span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><span style="font-size: 16px;"><br /></span></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">よって、連続性公理とは、実数のように途中に穴が開いていないびっしりと数が詰まっていることを表しています。</div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div><div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;"><br /></div></span></div></div></div></div></div></div></div></div></div></div></div></div></div></div>Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-1328330320593902432022-02-14T11:22:00.003+09:002022-04-19T10:33:14.739+09:00トポロジー入門(第15回)<p> <span style="color: orange; font-family: quot; font-size: 16px;">[場所:オンライン</span><span style="color: orange; font-family: quot; font-size: 16px;">(月曜日3限)]</span></p><div style="font-family: quot; font-size: 16px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><br /></div></div></div></div></div></div><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="font-family: quot; margin: 0px;"><div style="font-size: 16px; margin: 0px;"><a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/21/top.html">トポロジー入門のHP</a></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">最終回は、完備距離空間についてやりました。</div><div style="font-size: 16px;"><br /></div><div><span style="font-size: large;"><u>完備距離空間</u></span></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義15.1</u></div><div style="font-size: 16px;">距離空間 $(X,d)$ において点列 $(x_x)$ が、$\forall \epsilon>0$ に対して、</div><div style="font-size: 16px;">$\exists N\in {\mathbb N}$ と、$\forall m,n>N$ に対して、$d(x_n,x_m)<\epsilon$</div><div style="font-size: 16px;">を満たすとき、$(x_n)$ を<b>コーシー列</b>という。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この定義は ${\mathbb R}$ 上のコーシー列の一般化になっています。</div><div style="font-size: 16px;">一般に、コーシー列は収束列とは限りません。</div><div style="font-size: 16px;">たとえば、$a_n=1/n$ とすると、$(0,1)$ において、$a_n$ は</div><div style="font-size: 16px;">コーシー列ですが、$(0,1)$ に収束先はありません。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"> つぎに、距離空間の完備性の定義をします。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義15.2</u></div><div style="font-size: 16px;">$(X,d)$ を距離空間とする。任意のコーシー列が収束するとき、</div><div style="font-size: 16px;">$(X,d)$ は完備という。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">先ほどの例 $(0,1)$ は完備距離空間ではないということになります。</div><div style="font-size: 16px;">完備距離空間の例は以下のものがあります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義15.1</u> ${\mathbb R}$ は完備距離空間である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理15.2</u> コーシー列 $(a_n)$ は有界である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $\forall \epsilon>0$ に対して、</div><div style="font-size: 16px;">$\exists N\in {\mathbb N}$ において、$n_0,m>N$ となる自然数で</div><div style="font-size: 16px;">$n_0$ を固定しておきます。</div><div style="font-size: 16px;">このとき、$d(a_{n_0},a_m)<\epsilon$ です。</div><div style="font-size: 16px;">$\{d(a_{n_0},a_{n})|n\le N\}$ は高々有限集合なので、その最大が存在して、</div><div style="font-size: 16px;">それを $\delta$ とします。よって、</div><div style="font-size: 16px;">$d(a_{n_0},a_n)<\max\{\delta,\epsilon\}$となります。</div><div style="font-size: 16px;">$n,m\in {\mathbb N}$ に対して、</div><div style="font-size: 16px;">$d(a_n,a_m)\le d(a_n,a_{n_0})+d(a_{n_0},a_m)\le 2\max\{\delta,\epsilon\}$</div><div style="font-size: 16px;">$\text{diam}(\{a_n|a\in {\mathbb N}\})\le 2\max\{\delta,\epsilon\}$</div><div style="font-size: 16px;">よってコーシー列 $(a_n)$ は有界となります。$\Box$ </div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">次の定義をしておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義15.3</u></div><div style="font-size: 16px;">$(a_n)$ を 位相空間 $X$ の点列 $(a_n)$に対して、</div><div style="font-size: 16px;">$${\mathbb N}\ni k\mapsto n_k\in {\mathbb N}$$</div><div style="font-size: 16px;">を単射とする。</div><div style="font-size: 16px;">このとき、$(a_{n_k})$ を $(a_n)$ の<b>部分列</b>という。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理15.3</u>(ボルツァーノ・ワイエルシュトラスの定理)</div><div style="font-size: 16px;">${\mathbb R}$ の任意の有界数列は収束する部分列をもつ。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(x_n)$ を${\mathbb R}$ の有界数列とします。</div><div style="font-size: 16px;">このとき、$\{x_n|x\in {\mathbb N}\}\subset [-M,M]$ とします。</div><div style="font-size: 16px;">拡大縮小、平行移動をして $\{x_n|n\in{\mathbb N}\}\subset [0,1]$ としておきます。</div><div style="font-size: 16px;">定理13.8 (<a href="https://motochans.blogspot.com/2022/02/13.html">こちらのページ</a>)と同様に、区間を半分にしていくことで、</div><div style="font-size: 16px;">$$[0,1]\supset[a_1,b_1]\supset [a_2,b_2]\supset \cdots $$</div><div style="font-size: 16px;">各 $n\in{\mathbb N}$ に対して、$[a_n,b_n]$ において、</div><div style="font-size: 16px;">点列 $(x_n)$ が無限個入るようにしておきます。</div><div style="font-size: 16px;">このとき、$x_{n_1}\in [a_1,b_1]$ とし、$n_1<n_2$</div><div style="font-size: 16px;">かつ $x_{n_2}\in [a_2,b_2]$ となるようにします。</div><div style="font-size: 16px;">同様に、$n_{k-1}<n_{k}$ であって、$x_{n_k}\in [a_k,b_k]$ を満たすように</div><div style="font-size: 16px;">します。 </div><div style="font-size: 16px;">そうすると、上の区間の減少列において、</div><div style="font-size: 16px;">$a_n,b_n\to x$ が成り立ち、$\{x\}=\cap_{n=1}^\infty [a_n,b_n]$ となります。</div><div style="font-size: 16px;">$(x_n)$ の部分列 $(x_{n_k})$ であって $a_k\le x_{n_k}\le b_k$ を満たします。</div><div style="font-size: 16px;">また、$a_k,b_k\to x$ を満たすので、$x_{n_k}\to x$ となります。</div><div style="font-size: 16px;">よって、$(x_n)$ の部分列 $(x_{n_k})$ は $x$ に収束する部分列になります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、次の定理が成り立ちます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理15.4 </u></div><div style="font-size: 16px;">コーシー列 $(a_n)$ が収束する部分列をもつなら、$(a_n)$ は収束列である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この証明は $\epsilon$-$N$ 論法を使って簡単に証明できるので、ここでは省略します。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この定理を用いることで、上の$ {\mathbb R}$ は完備距離空間であることがわかります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(定理15.1の証明)</div><div style="font-size: 16px;">$(a_n)$ を任意の ${\mathbb R}$ のコーシー列とします。</div><div style="font-size: 16px;">このとき、$(a_n)$ は有界数列なので、ボルツァーノ・ワイエルシュトラスの</div><div style="font-size: 16px;">定理により、収束する部分列を持ちます。</div><div style="font-size: 16px;">$(a_n)$ が収束する部分列をもつので、定理15.4から$(a_n)$ は収束列ということに</div><div style="font-size: 16px;">なります。よって、${\mathbb R}$ は完備距離空間になりました。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで以下を示しましょう。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理15.5</u></div><div style="font-size: 16px;">距離空間において以下が同値である。</div><div style="font-size: 16px;">・コンパクト空間</div><div style="font-size: 16px;">・全有界かつ完備</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">まず、上から下の条件を導きましょう。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理15.6</u></div><div style="font-size: 16px;">コンパクト距離空間は全有界かつ完備である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) コンパクト距離空間は全有界であることは既に示したので、</div><div style="font-size: 16px;">完備性を示そう。</div><div style="font-size: 16px;">$(a_n)$ を任意のコーシー列とします。</div><div style="font-size: 16px;">$A=\{a_n|n\in {\mathbb N}\}$ </div><div style="font-size: 16px;">とします。</div><div style="font-size: 16px;"><u>$A$ が集積点を持たないとします。</u></div><div style="font-size: 16px;">このとき、$B_d(x,\epsilon_x)\cap A=\emptyset $ または $\{x\}$ </div><div style="font-size: 16px;">であり、$\{B_d(x,\epsilon_x)|x\in X\}$ は $X$ の開被覆であり、</div><div style="font-size: 16px;">コンパクト性から、$\{B_d(x_i,\epsilon_{x_i})|i=1,\cdots, n\}$ が</div><div style="font-size: 16px;">部分開被覆となります。</div><div style="font-size: 16px;">よって、$A=\cup_{i=1,\cdots,n}(B_d(x_i,\epsilon_{x_i}\cap A)\subset \{x_i|i=1,\cdots,n\}$</div><div style="font-size: 16px;">より、$A$ は有限集合になります。</div><div style="font-size: 16px;">そうすると、ある $p\in A$ に対して、無限個の $(a_n)$ が存在して、</div><div style="font-size: 16px;">$a_n=p$ となります。</div><div style="font-size: 16px;">よって、収束する部分列を持ちます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">一方、</div><div style="font-size: 16px;"><u>$A$ に集積点を持つとします。</u></div><div style="font-size: 16px;">それを $x\in X$ とします。</div><div style="font-size: 16px;">このとき、$a_{n_1}\in B_d(x,1)\cap (A\setminus \{x\})$ </div><div style="font-size: 16px;">とします。このとき、$\delta_1=d(a_{n_1},x)/2$ とします。</div><div style="font-size: 16px;">条件から $\delta_1<\frac{1}{2}$ です。</div><div style="font-size: 16px;">$a_{n_2}\in B_d(x,\delta_1)\cap (A\setminus\{x\})$</div><div style="font-size: 16px;">として、$\delta_2=\frac{d(a_{n_2},x)}{2}$ とすると、</div><div style="font-size: 16px;">$\delta_2<\frac{\delta_1}{2}<\frac{1}{4}$</div><div style="font-size: 16px;">$a_{n_3}\in B_d(x,\delta_2)\cap(A\setminus \{x\})$</div><div style="font-size: 16px;">より、これを続けることで、$\delta_n<\frac{1}{2^n}$ です。</div><div style="font-size: 16px;">また、選び方から、$a_{n_1},a_{n_2}, a_{n_3},\cdots$ は全て違う点であるから、</div><div style="font-size: 16px;">$a_{n_k}$ は部分列であることがわかります。</div><div style="font-size: 16px;">また、$d(a_{n_k},x)<\delta_k<\frac{1}{2^k}$ であるから、</div><div style="font-size: 16px;">$a_{n_k}\to x$ であることが分かります。</div><div style="font-size: 16px;">どちらにしても、コーシー列 $(a_n)$ に対して 収束する部分列 $(a_{n_k})$ </div><div style="font-size: 16px;">が存在します。</div><div style="font-size: 16px;">よって、定理15.4から、$(a_n)$ は収束する列になります。</div><div style="font-size: 16px;">よって、$(X,d)$ は完備になります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">次に上の逆を示しましょう。その前に次を示しておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理15.7</u> 全有界な距離空間は、任意の点列は、コーシー列となる部分列をもつ。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(X,d)$ を全有界な距離空間とします。</div><div style="font-size: 16px;">今、$\forall n\in {\mathbb N}$ に対して、</div><div style="font-size: 16px;">有限点 $\{x_1^n,\cdots, x_{m_n}^n\}$ が存在して、</div><div style="font-size: 16px;">$$X=\cup_{k=1}^{m_n}B_d\left(x^n_k,\frac{1}{n}\right)$$</div><div style="font-size: 16px;">が成り立ちます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで、$(a_n)$ を任意の点列とします。</div><div style="font-size: 16px;">$B_1=B_d(x_{k_1}^1,1)$</div><div style="font-size: 16px;">$B_2=B_d(x_{k_1}^1,1)\cap B_d(x_{k_2}^2,\frac{1}{2})$</div><div><div style="font-size: 16px;">$B_3=B_d(x_{k_1}^1,1)\cap B_d(x_{k_2}^2,\frac{1}{2})\cap B_d(x_{k_3}^3,\frac{1}{3})$</div><div style="font-size: 16px;">$\cdots$</div><div style="font-size: 16px;">のようにして、$B_i$ には、$(a_n)$ のうち無限個の点列を含むようにすることが</div><div style="font-size: 16px;">できます。同じことですが、そのような $k_i$ を選ぶことができます。</div><div style="font-size: 16px;">そうすると、</div><div style="font-size: 16px;">$$B_1\supset B_2\supset B_3\cdots$$</div><div style="font-size: 16px;">となることがわかります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">こうすることで、$a_{n_1}\in B_1, a_{n_2}\in B_2,\cdots$ </div><div style="font-size: 16px;">のように点列を選ぶことができて、$n_1<n_2<\cdots$ と仮定することができます。</div><div style="font-size: 16px;">よって、部分列 $a_{n_k}$ をとることができます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">今、$\epsilon>0$ に対して、$\frac{1}{2N}<\epsilon$ となる自然数 $N$ が存在して、</div><div style="font-size: 16px;">$\forall k,l>N$ に対して、</div><div style="font-size: 16px;">$d(a_{n_k},a_{n_l})\le d(a_{n_k},x^N_{n_k})+d(x^N_{n_k},a_{n_l})<\frac{1}{N}+\frac{1}{N}<\epsilon$ </div><div style="font-size: 16px;">を満たします。</div><div style="font-size: 16px;">よって、$a_{n_k}$ はコーシー列となります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">では、先ほどの定理15.5の下から上を示しましょう。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理15.8</u></div><div style="font-size: 16px;">全有界かつ完備距離空間はコンパクト空間である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(X,d)$ が全有界完備であるとします。</div><div style="font-size: 16px;">このとき、$(X,\mathcal{O}_{d})$ は全有界かつ完備であるから、</div><div style="font-size: 16px;">$X$ はリンデレフ空間になります。</div><div style="font-size: 16px;">よって、任意の開被覆 $\mathcal{U}\subset \mathcal{O}_d$ に対して、</div><div style="font-size: 16px;">高々可算部分被覆 $\mathcal{V}\subset \mathcal{U}$ が</div><div style="font-size: 16px;">存在します。$\mathcal{V}$ が有限であれば、証明が終わるので、</div><div style="font-size: 16px;">可算(無限)集合であるとします。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$\mathcal{V}$ にどんな有限部分集合も $X$ を被覆しないと仮定しておきます。</div><div style="font-size: 16px;">どんな $n\in {\mathbb N}$ に対しても、</div><div style="font-size: 16px;">$x_n\not\in V_1\cup \cdots \cup V_n$ をとり、点列 $(x_n)$ を構成します。</div><div style="font-size: 16px;">$\forall i\in {\mathbb N}$ に対しても $V_i$ には高々有限個の $x_n$ のみしか</div><div style="font-size: 16px;">含みません。</div><div style="font-size: 16px;">この点列 $(x_n)$ には定理15.7 から部分コーシー列 $(x_{n_k})$ が存在します。</div><div style="font-size: 16px;">よって、完備性から $(x_{n_k})$ は収束し、その収束先を $x$ とおくと、</div><div style="font-size: 16px;">$x\in V_n$ に対して、$V_n$ は $x$ の近傍であるから、$V_n$ に含まれる</div><div style="font-size: 16px;">無限個の $x_{n_k}$ が存在することになります。</div><div style="font-size: 16px;">これは、$V_n$ には高々有限個の $x_{n_k}$ しか存在しないことに矛盾します。</div><div style="font-size: 16px;">よって、$\mathcal{V}$ は、有限部分被覆が存在します。</div><div style="font-size: 16px;">これは $\mathcal{U}$ の有限部分被覆でもあるから、</div><div style="font-size: 16px;">$X$ はコンパクトであることになります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">これにより、定理15.5のコンパクト距離空間が全有界かつ完備であること</div><div style="font-size: 16px;">と同値であることが証明できました。</div><div style="font-size: 16px;"><br /></div><div><span style="font-size: medium;"><u>ベールの定理</u></span></div><div style="font-size: 16px;">ここで、完備距離空間の性質としてベールの定理を示しておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">そのために以下の定義をしておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義15.4</u></div><div style="font-size: 16px;">位相空間 $(X,d)$ に対して、$A\subset X$ が</div><div style="font-size: 16px;">$\text{Int}(\text{Cl}(A))=\emptyset$ であるとき、$A$ は<b>疎集合</b>であるという。</div><div style="font-size: 16px;">また、疎集合の可算個の和集合のことを<b>第1類</b>という。</div><div style="font-size: 16px;">また、第1類ではない集合を<b>第2類</b>という。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">疎集合であることは、上と同値な条件に、</div><div style="font-size: 16px;">$\text{Cl}(\text{Int}(A^c))=X$ があります。</div><div style="font-size: 16px;">これは、以前示した、</div><div style="font-size: 16px;">$(\text{Int}(B))^c=\text{Cl}(B^c)$</div><div style="font-size: 16px;"><div>$(\text{Cl}(B))^c=\text{Int}(B^c)$</div><div>により導けます。</div></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、条件から、疎集合の部分集合は全て疎集合であることが分かります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">注意すべきことは、疎集合というのは、入っている位相空間に依存しますし、</div><div style="font-size: 16px;">どのように入っているかにも依存します。</div><div style="font-size: 16px;">たとえば、$[0,1]\subset {\mathbb R}$ は疎集合ではありませんが、</div><div style="font-size: 16px;">$[0,1]\subset [0,1]\times [0,1]$ は疎集合となります。</div><div style="font-size: 16px;">特に、疎集合かどうかは位相的性質ではありません。</div></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>例15.2</u></div><div style="font-size: 16px;">疎集合の例としては、$({\mathbb R},\mathcal{O}_{d_1})$ 上の</div><div style="font-size: 16px;">有限点集合は疎集合です。</div><div style="font-size: 16px;">よって、${\mathbb R}$ 上の可算集合は全て第1類ということになります。</div><div style="font-size: 16px;">例えば有理数全体の集合 ${\mathbb Q}\subset {\mathbb R}$ </div><div style="font-size: 16px;">は通常のユークリッド距離位相空間において第1類となります。</div><div style="font-size: 16px;">ただ、疎集合ではありません。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで、疎集合を特徴づけましょう。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>命題15.1</u></div><div style="font-size: 16px;">稠密開集合の補集合は疎集合である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $D$ を稠密開集合であるとします。</div><div style="font-size: 16px;">条件から、</div><div style="font-size: 16px;">$\text{Int}(D)=D$ かつ $\text{Cl}(D)=X$ であるから、</div><div style="font-size: 16px;">$\text{Int}(\text{Cl}(D^c))=\text{Int}((\text{Int}(D))^c)=\text{Int}(D^c)=(\text{Cl}(D))^c=X^c=\emptyset$</div><div style="font-size: 16px;">となります。</div><div style="font-size: 16px;">よって、$D^c$ が疎集合ということになります。$\Box$ </div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">次の命題を示しましょう。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>命題15.2</u></div><div style="font-size: 16px;">$(X,\mathcal{O})$ を位相空間とするとき、以下は同値。</div><div style="font-size: 16px;">・$A\subset X$ が疎集合である</div><div style="font-size: 16px;">・稠密開集合 $D$ が存在して、$A\subset D^c$ となる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(\Rightarrow)$ を示します。</div><div style="font-size: 16px;">$A\subset X$ を疎集合とします。</div><div style="font-size: 16px;">このとき、$D=(A^c)^\circ$ とおくと、$D$ は開集合であり、</div><div style="font-size: 16px;">$\text{Cl}(D)=\overline{(A^c)^\circ}=X$ となります。</div><div style="font-size: 16px;">よって、$D$ は稠密開集合となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、$D\subset A^c$ であるので条件を満たします。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">($\Leftarrow$) を示します。</div><div style="font-size: 16px;">$D$ を稠密開集合とし、$A\subset D^c$ とします。</div><div style="font-size: 16px;">このとき、命題15.1から、$D^c$ は疎集合であり、</div><div style="font-size: 16px;">疎集合の部分集合は疎集合であるから</div><div style="font-size: 16px;">$A$ が疎集合であることになります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで、ベールの定理を書いておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理15.9(ベールの定理)</u></div><div style="font-size: 16px;">$(X,d)$ を完備距離空間とする。</div><div style="font-size: 16px;">$D_i$ が可算個の稠密開集合のとき、</div><div style="font-size: 16px;">$\cap_{i=1}^\infty D_i$ は $X$ で稠密集合になる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この証明を考える前に、いくつか説明をしておきます。</div><div style="font-size: 16px;">まずこの定理が補集合では何を言っているか考えましょう。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">するとこうなります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">完備距離空間において、</div><div style="font-size: 16px;">$A_i$ を可算個の稠密開集合の補集合とする。</div><div style="font-size: 16px;">このとき、</div><div style="font-size: 16px;">$A=\cup_{i=1}^\infty A_i$ は、$A^\circ=\emptyset$ である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、その部分集合を取ると、</div><div style="font-size: 16px;">完備距離空間において、</div><div style="font-size: 16px;">$A_i$ を可算個の疎集合とする。</div><div style="font-size: 16px;">このとき、</div><div style="font-size: 16px;"><div>$A=\cup_{i=1}^\infty A_i$ は、$A^\circ=\emptyset$ である。</div><div><br /></div></div><div style="font-size: 16px;">つまり、ベールの定理は、</div><div style="font-size: 16px;">完備距離空間において、</div><div style="font-size: 16px;">$A\subset X$ が第1類ならば、 $A^\circ =\emptyset$ となる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">さらに、対偶を取れば、</div><div style="font-size: 16px;">$A$ が内点を持つとすると、$A$ は可算個の疎集合の和集合で書けない。</div><div style="font-size: 16px;">つまり、第2類集合である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">特に、$X$ が完備距離空間であれば、</div><div style="font-size: 16px;">$A=X$ は可算個の疎集合の和によって書くことができない。</div><div style="font-size: 16px;"><br /></div><div><span style="font-size: 16px;">評語的に言えば、</span></div><div><span style="font-size: medium;">「チリ</span>(疎集合)も<span style="font-size: medium;">つもれど</span>(可算和集合をとっても)<span style="font-size: medium;">山</span>(内点を持つ集合)にならない」</div><div style="font-size: 16px;">ということが言えます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>例15.3</u> 例えば、${\mathbb R}^2$ は完備距離空間ですが、</div><div style="font-size: 16px;">可算個の直線によって覆うことができない。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、</div><div style="font-size: 16px;"><u>例15.4</u> 有理数空間は、完備距離空間に同相ではない。</div><div style="font-size: 16px;">なぜなら、もし同相なら、1点集合は、疎集合であるから、${\mathbb Q}$ </div><div style="font-size: 16px;">は、疎集合の可算和集合になってしまうからです。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、一般化して、孤立点を含まなければ完備距離空間は非可算個の点を含みます。</div><div style="font-size: 16px;">もちろん、離散距離空間は、可算個でも完備距離空間になります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ではここで、ベールの定理の証明をしておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(X,d)$ を完備距離空間とします。</div><div style="font-size: 16px;">$D_i$ を稠密開集合とします。($i=1,2,\cdots$)</div><div style="font-size: 16px;">この時、$\cap_{i=1}^nD_i$ が $X$ において稠密であることを示します。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$\forall x\in X$ として、$\forall \epsilon>0$ に対して、</div><div style="font-size: 16px;">$\epsilon_0=\epsilon$ とし、$x_0=x$ とおきます。</div><div style="font-size: 16px;">このとき、</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$B_d(x_0,\epsilon)\cap D_i\neq \emptyset$ </div><div style="font-size: 16px;">であるから、$x_1\in B_d(x_0,\epsilon_0)\cap D_i$ を取ります。</div><div style="font-size: 16px;">また、</div><div style="font-size: 16px;">$0<\epsilon_1<\frac{1}{2}$ かつ、</div><div style="font-size: 16px;">$$\text{Cl}(B_d(x_1,\epsilon_1))\subset B_d(x_0,\epsilon_0)\cap D_1$$</div><div style="font-size: 16px;">とすることができます。</div><div style="font-size: 16px;">それは、$B_d(x,\epsilon)$ かつ $D_i$ がどちらも開集合であり、</div><div style="font-size: 16px;">距離空間が正則空間であることからわかります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、$B_d(x_1,\epsilon_1)\cap D_2\neq \emptyset$ であることから、</div><div style="font-size: 16px;">$x_2\in B_d(x_1,\epsilon_1)\cap D_2$ を取り、</div><div style="font-size: 16px;">$0<\epsilon_2<\frac{1}{4}$ を取り、</div><div><div style="font-size: 16px;">$$\text{Cl}(B_d(x_2,\epsilon_2))\subset B_d(x_1,\epsilon_1)\cap D_2$$</div><div style="font-size: 16px;">とすることができます。このようにして、</div><div style="font-size: 16px;">任意の $n\in {\mathbb N}$ に対して、</div><div style="font-size: 16px;"><div><br /></div><div>$$\text{Cl}(B_d(x_n,\epsilon_n))\subset B_d(x_{n-1},\epsilon_{n-1})\cap D_n$$</div></div><div style="font-size: 16px;">かつ $0<\epsilon_n<\frac{1}{2^n}$ を取ることができます。</div><div style="font-size: 16px;">よって、</div><div style="font-size: 16px;">$\forall m>n\in {\mathbb N}$ に対して、</div><div style="font-size: 16px;">$B_d(x_m,\epsilon_m)\subset B_d(x_{m-1},\epsilon_{m-1})\subset \cdots\subset B_d(x_n,\epsilon_n)$</div><div style="font-size: 16px;">となります。</div><div style="font-size: 16px;">このことから、$d(x_m,x_n)<\epsilon_n<\frac{1}{2^n}$ より、</div><div style="font-size: 16px;">$(x_n)$ はコーシー列であり、</div><div style="font-size: 16px;">完備性から、ある $x_\infty$ に収束します。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、$\forall n\in {\mathbb N}$ に対して $\forall m>n$ に対して、</div><div style="font-size: 16px;">$x_m\in B_d(x_n,\epsilon_n)$ であるから、</div><div style="font-size: 16px;">$x_\infty\in \text{Cl}(B_d(x_n,\epsilon_n))$ となります。</div><div style="font-size: 16px;">よって $x_\infty\in D_n$ かつ、$B_d(x_n,\epsilon)$ であることがわかります。</div><div style="font-size: 16px;">特に、$x_\infty\in B_d(x,\epsilon)$ であるから、$x_\infty\in B_d(x,\epsilon)\cap_{n=1}^\infty D_n$ </div><div style="font-size: 16px;">であり、</div><div style="font-size: 16px;">$x_\infty\in \overline{\cap_{n=1}^\infty D_n}$ であることがわかります。</div><div style="font-size: 16px;">よって、$\cap_{n=1}^\infty D_n$ は $X$ で稠密であることがわかります。 $\Box$</div><div style="font-size: 16px;"><br /></div><div><u><span style="font-size: medium;">完備化</span></u></div><div style="font-size: 16px;">最後に、完備化の話をして終わります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">完備化の定義をしておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$(X,d)$ を距離空間とします。</div><div style="font-size: 16px;">このとき、ある完備距離空間 $(\hat{X},\hat{d})$ が存在して、</div><div style="font-size: 16px;">$h(X)\subset \hat{X}$ が稠密となる単射連続写像 $h:X\to \hat{X}$ が存在するとき、</div><div style="font-size: 16px;">$(\hat{X},h)$ を $X$ の<b>完備化</b>と言います。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">例えば、以下の例ががあります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>例15.6</u></div><div style="font-size: 16px;">$({\mathbb R},d_1)$ における $({\mathbb Q},d_1)$ の通常の埋め込みは完備化である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、次の例を考えます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$C^\ast(X)$ を $X$ 上の有界な実数値連続関数全体とします。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$f,g\in C^\ast(X)$ に対して、</div><div style="font-size: 16px;">$$d_{\sup}^X(f,g)=\sup\{|f(x)-g(x)||x\in X\}$$</div><div style="font-size: 16px;">と定義すると、$(C^\ast(X),d_{\sup}^X)$ は完備距離空間となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">完備化についての定理を与えて証明することで終わります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"> <u>定理15.10</u> </div><div style="font-size: 16px;">$(X,d)$ を距離空間とする。</div><div style="font-size: 16px;">このとき、$X$ の完備化 $(\hat{X},h)$ が存在し、</div><div style="font-size: 16px;">完備化は等長写像を除いて一意的である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(X,d)$ を距離空間とします。</div><div style="font-size: 16px;">$a,x\in X$ に対して、 </div><div style="font-size: 16px;">$\varphi_x(z)=d(a,z)-d(x,z)$ </div><div style="font-size: 16px;">とすることで、$\varphi_z\in C^\ast(X)$ が構成します。</div><div style="font-size: 16px;">ここで、$\Phi(x)=\varphi_x$ とします。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$\Phi$ が単射であることを示しておきます。</div><div style="font-size: 16px;">$\Phi(x)=\Phi(y)$ であるとします。</div><div style="font-size: 16px;">このとき、任意の $z\in X$ に対して、$d(x,z)=d(y,z)$ であり、</div><div style="font-size: 16px;">特に $z=x$ とおくことで、$0=d(x,y)$ であるから、$x=y$ となります。</div><div style="font-size: 16px;">よって、単射</div><div style="font-size: 16px;">$$\Phi:X\to C^\ast(X)$$</div><div style="font-size: 16px;">を得ます。</div><div style="font-size: 16px;">このとき、$|\varphi_x(z)-\varphi_y(z)|=|d(a,z)-d(x,z)-(d(a,z)-d(y,z))|=|d(x,z)-d(y,z)|\le d(x,y)$</div><div style="font-size: 16px;">より、</div><div style="font-size: 16px;">$d_\sup^X(\varphi_x,\varphi_y)\le d(x,y)$ </div><div style="font-size: 16px;">が成り立ちます。この不等式から $\Phi$ が連続であることがわかり、同様に、$d(x,y)\le d_\sup^X(\varphi_x,\varphi_y)$</div><div style="font-size: 16px;">も成り立つことがわかります。</div><div style="font-size: 16px;">$\Phi:X\to C^\ast(X)$ は距離を保つ連続写像であることがわかります</div><div style="font-size: 16px;">$\hat{X}=\text{Cl}(\Phi(X))\subset C^\ast(X)$ とすることで、</div><div style="font-size: 16px;">$\hat{X}$ は完備距離空間であり、$X\subset \hat{X}$</div><div style="font-size: 16px;">は完備化となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">一意性に対してはここでは省略します。$\Box$ </div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで使ったのは、完備距離空間の中の閉集合はまた完備距離空間であることでした。</div></div></div></div></div></div></div></div>Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-48142036878190114502022-02-11T07:14:00.001+09:002022-04-19T10:33:43.790+09:00トポロジー入門(第14回)<p> <span style="color: orange; font-family: quot; font-size: 16px;">[場所:オンライン</span><span style="color: orange; font-family: quot; font-size: 16px;">(月曜日3限)]</span></p><div style="font-family: quot; font-size: 16px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><br /></div></div></div></div></div></div><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="font-family: quot; margin: 0px;"><div style="font-size: 16px; margin: 0px;"><a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/21/top.html">トポロジー入門のHP</a></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">今回は、コンパクト空間についての後半部分をやりました。</div><div style="font-size: 16px;">コンパクト空間の前半は<a href="https://motochans.blogspot.com/2022/02/13.html">こちら</a>を見てください。</div><div style="font-size: 16px;"><br /></div><div><span style="font-size: large;"><u>コンパクト空間</u></span></div><div style="font-size: 16px;">まず、次の定理を示しました。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.1(チコノフの定理)</u></div><div style="font-size: 16px;">$X,Y$ をコンパクト空間であれば、</div><div style="font-size: 16px;">$X\times Y$ もコンパクト空間</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">証明</div><div style="font-size: 16px;">$\mathcal{U}$ を $X\times Y$ の開被覆とします。</div><div style="font-size: 16px;">このとき、$(x,y)\in X\times Y$ に対して、$U_{(x,y)}\in \mathcal{U}$ </div><div style="font-size: 16px;">を選んでおきます。</div><div style="font-size: 16px;">また、$U_{(x,y)}$ に対して、$A_y(x)\times B_x(y)\subset U_{(x,y)}$</div><div style="font-size: 16px;">となる$x$ の開近傍 $A_y(x)$ と $y$ の開近傍 $B_x(y)$ が存在します。</div><div style="font-size: 16px;">この時、$\{B_x(y)|y\in Y\}$ は $Y$ の開被覆であり、コンパクト性から、</div><div style="font-size: 16px;">$B_x(y_1^x),\cdots, B_x(y_{n_x}^x)$</div><div style="font-size: 16px;">が存在して、それらは$Y$ の被覆になります。</div><div style="font-size: 16px;">ここで、$x\in X$ を固定して、$\{A_{y_j^x}(x)\times B_x(y_j^x)|j=1,\cdots, n_x\}$</div><div style="font-size: 16px;">は $\{y\}\times Y$ の被覆になっていて、</div><div style="font-size: 16px;">$V(x)=\cap_{i=1}^{n_x}A_{y_j^x}(x)$</div><div style="font-size: 16px;">は有限個の共通部分なので、$x$ の開近傍になります。</div><div style="font-size: 16px;">また、$V(x)\times Y\subset \cup_{j=1}^{n_x}A_{y_j^x}(x)\times B_x(y_j^x)$</div><div style="font-size: 16px;">となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで、$\{V(x)|x\in X\}$ は $X$ の開被覆なので、$X$ のコンパクト性から</div><div style="font-size: 16px;">$x_1,\cdots, x_s$ が存在して、</div><div style="font-size: 16px;">$\{V(x_1),\cdots,V(x_s)\}$ は $X$ の開被覆となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">今、$\mathcal{V}=\{U_{(x_i,y_j^{x_i})}|1\le i\le s,1\le j\le n_{x_i}\}$ </div><div style="font-size: 16px;">が $X\times Y$ の有限被覆であることを示します。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$(x,y)\in X\times Y$ に対して、</div><div style="font-size: 16px;">$x\in V(x_i)$ となる $x_i$ が存在して、</div><div style="font-size: 16px;">$y\in B_{x_i}(y_j^{x_i})$ となる $1\le j\le n_{x_i}$ が存在して、</div><div style="font-size: 16px;">$(x,y)\in V(x_i)\times B_{x_i}(y_j^{x_i})$ となります。</div><div style="font-size: 16px;">つまり、$V(x_i)\times B_{x_i}(y_j^{x_i})\subset A_{y_j^{x_i}}(x_i)\times B_{x_i}(y_j^{x_i})\subset U_{(x_i,y_j^{x_i})}$ がなりたつので、</div><div style="font-size: 16px;">$\mathcal{V}$ は有限開被覆になり、$\mathcal{V}\subset\mathcal{U}$ であるから、</div><div style="font-size: 16px;">$X$ はコンパクトになります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">このことから、このチコノフの定理を有限回繰り返すことで、</div><div style="font-size: 16px;">各 $X_i$ がコンパクトであるとき、有限直積位相 $X_i\times X_2\times \cdots\times X_n$ </div><div style="font-size: 16px;">もまたコンパクトであることがわかります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">例として、区間の直積 $(I^1,\mathcal{O}_{d_1})$ や $(I^n,\mathcal{O}_{d_n})$ はコンパクトであることがわかります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">実は、このチコノフの定理は、</div><div style="font-size: 16px;">有限だけではなく、任意個の直積に対しても、</div><div style="font-size: 16px;">同じ主張が成り立ちます。</div><div style="font-size: 16px;">ここでは証明はしません。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.2(チコノフの定理)</u></div><div style="font-size: 16px;">$X_\lambda$ ($\lambda\in \Lambda$)$ がコンパクト空間である時、</div><div style="font-size: 16px;">$$\prod_{\lambda\in \Lambda}X_\lambda$$</div><div style="font-size: 16px;">もコンパクト空間。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">このことから、</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$(I^{\mathbb N},\mathcal{O}_{d_1}^{\mathbb N})$ がコンパクトとなります。</div><div style="font-size: 16px;">この空間は、</div><div style="font-size: 16px;">$d_{\infty}((x_n),(y_n))=\sum_{n=1}^\infty\frac{1}{2^n}|x_n-y_n|$</div><div style="font-size: 16px;">を距離関数として距離空間になり、その時、</div><div style="font-size: 16px;">$\mathcal{O}_{d_1}^{\mathbb N}=\mathcal{O}_{d_\infty}$ が成り立ちます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、直積空間 $\prod_{\lambda\in \Lambda}X_\lambda$ がコンパクトであれば、</div><div style="font-size: 16px;">標準射影 $\text{pr}_\lambda$ が全射連続であることから、因子空間</div><div style="font-size: 16px;">$\text{pr}_{\lambda}(\prod_{\lambda\in \Lambda}X_\lambda)=X_\lambda$ </div><div style="font-size: 16px;">もコンパクトであることがわかります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">これ以降、コンパクト空間の性質についてまとめておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">まず、次の定理は今後よく使います。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.3</u></div><div style="font-size: 16px;">コンパクト空間の閉集合はコンパクト。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(X,\mathcal{O})$ をコンパクト空間とします。</div><div style="font-size: 16px;">$A\subset X$ を閉集合として、$A$ の開被覆を $\mathcal{U}\subset \mathcal{O}$ とします。</div><div style="font-size: 16px;">この時、$\mathcal{U}\cup\{A^c\}$ は $X$ の開被覆であるから、</div><div style="font-size: 16px;">$\mathcal{U}$ の有限部分集合 $\mathcal{V}$ が存在して、$\mathcal{V}\cup\{A^c\}$ は</div><div style="font-size: 16px;">$X$ の被覆になります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">このとき、$\mathcal{V}$ は $A$ の被覆になっていますので、</div><div style="font-size: 16px;">$A$ の任意の開被覆に対して、有限部分被覆が存在したことになります。</div><div style="font-size: 16px;">これは、$A$ がコンパクトであることを意味します。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">次の定理を示しました。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.4</u></div><div style="font-size: 16px;">$X$ をハウスドルフ空間とする。</div><div style="font-size: 16px;">$A\subset X$ をコンパクト集合とし、</div><div style="font-size: 16px;">$x\in X$ で、$x\not\in A$ となる $x$ に対してある開集合 $U,V$ が存在して、</div><div style="font-size: 16px;">$x\in U$ かつ $A\subset V$ かつ $U\cap V=\emptyset$ を満たす。</div><div style="font-size: 16px;">つまり、$x,A$ は開集合で分離できる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $\forall p\in A$ をとります。</div><div style="font-size: 16px;">$x,p$ を分離する開集合 $U_p,V_p$ をとり、</div><div style="font-size: 16px;">$x\in U_p, p\in V_p$ かつ $U_p\cap V_p=\emptyset$</div><div style="font-size: 16px;">を満たします。 </div><div style="font-size: 16px;">この時、$\mathcal{V}=\{V_p|p\in A\}$ は $A$ の開被覆となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">よって、$p_1,\cdots ,p_n\in A$ が存在して、$\{V_{p_1},V_{p_2},\cdots,V_{p_n}\}$</div><div style="font-size: 16px;">は $A$ は開被覆となります。</div><div style="font-size: 16px;">つまり、$A\subset \cup_{i=1}^{n}V_{p_i}=:V$ とする。</div><div style="font-size: 16px;">$\cap_{i=1}^nU_{p_i}=:U\in \mathcal{O}$ とおくことで、</div><div style="font-size: 16px;">$x\in U$ かつ、$U\cap V=\emptyset$ となります。</div><div style="font-size: 16px;">もし、$U\cap V\neq \emptyset$ となるとすると、</div><div style="font-size: 16px;">$y\in U\cap V$ に対して、 ある $i$ が存在して、$y\in U_{p_i}\cap V_{p_i}$</div><div style="font-size: 16px;">となるので、これは、一般に、$U_p, V_p$ が共通部分を持たないことに反します。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">これにより、$U,V$ は、$x,A$ を分離する開集合となります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">さらに、この定理を使って、次の定理を示すことができます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.5</u></div><div style="font-size: 16px;">$X$ をハウスドルフ空間とする。</div><div style="font-size: 16px;">互いに交わらないコンパクト集合 $A,B\subset X$ は</div><div style="font-size: 16px;">開集合によって分離できる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $A,B\subset X$ を互いに交わらないコンパクト集合とします。</div><div style="font-size: 16px;">この時、定理14.4に対して、$x\in A$ と $B$ に対して、交わらない開集合 $U_x,V_x$</div><div style="font-size: 16px;">が存在して、</div><div style="font-size: 16px;">$x\in U_x, B\subset V_x, U_x\cap V_x=\emptyset$ </div><div style="font-size: 16px;">となります。</div><div style="font-size: 16px;">ここで、$\{U_x|x\in A\}$ は $A$ の開被覆であり、</div><div style="font-size: 16px;">$A$ はコンパクトであるから、$x_1,x_2,\cdots,x_m\in A$ が存在して</div><div style="font-size: 16px;">$\{U_{x_i}|i=1,\cdots, m\}$ が $A$ の有限開被覆となります。</div><div style="font-size: 16px;">ここで、$B\subset \cap \{V_{x_i}|i=1,\cdots, m\}\in \mathcal{O}$ </div><div style="font-size: 16px;">となるので、</div><div style="font-size: 16px;">$$U=\cup\{U_{x_i}|i=1,\cdots, m\}$$</div><div style="font-size: 16px;">$$V=\cap\{V_{x_i}|i=1,\cdots, m\}$$</div><div style="font-size: 16px;"> </div><div style="font-size: 16px;">とすることで、$U,V$ は前と同じ議論により、$U,V $は $A,B$ を分離する</div><div style="font-size: 16px;">開被覆となります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この定理を用いると、</div><div style="font-size: 16px;">コンパクトハウスドルフ空間は、正規空間であることがわかります。</div><div style="font-size: 16px;">なぜなら、コンパクトハウスドルフ空間の任意の閉集合は定理14.3からコンパクトになり、</div><div style="font-size: 16px;">それらは開集合によって分離できるからです。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">次の定理を示しました。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.6</u></div><div style="font-size: 16px;">$X$ をハウスドルフ空間とする。</div><div style="font-size: 16px;">この時、任意のコンパクト集合 $A\subset X$ は閉集合となる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $A$ をコンパクト集合します。</div><div style="font-size: 16px;">$x\not\in A$ となる $x$ に対して、</div><div style="font-size: 16px;">$x\in U, A\subset V, U\cap V=\emptyset$ </div><div style="font-size: 16px;">となる開集合 $U,V$ が存在します。</div><div style="font-size: 16px;">特に、$x\in U\in A^c$ となり、 $x$ は $A^c$ の内点であることが</div><div style="font-size: 16px;">わかります。</div><div style="font-size: 16px;">よって、$A$ は閉集合であることがわかります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">距離空間はハウスドルフ空間なので、</div><div style="font-size: 16px;">距離空間のコンパクト集合は閉集合であることがわかりました。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">次に位相空間論で必ず習う次の定理を紹介しておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.7</u></div><div style="font-size: 16px;">コンパクト空間からハウスドルフ空間への連続写像は閉写像となる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明). $f:X\to Y$ をコンパクト空間からハウスドルフ空間への連続写像</div><div style="font-size: 16px;">とします。この時、$F\subset X$ を閉集合とします。</div><div style="font-size: 16px;">$X$ はコンパクトであるから、$F$ はコンパクト集合になります。</div><div style="font-size: 16px;">よって、コンパクト集合の連続像はコンパクトであるから、</div><div style="font-size: 16px;">$f(F)$ はコンパクトとなります。</div><div style="font-size: 16px;">定理14.6からハウスドルフ空間の中のコンパクト集合は閉集合であったから、</div><div style="font-size: 16px;">$f(F)$ は閉集合となります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この定理を用いると、次を証明することができます。</div><div style="font-size: 16px;">・コンパクト空間からハウスドルフ空間への全射連続写像は商写像となる。</div><div><div style="font-size: 16px;">・コンパクト空間からハウスドルフ空間への全単射連続写像は同相写像となる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで、コンパクト空間を少し一般化しておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">定理14.1</div><div style="font-size: 16px;">位相空間 $X$ の任意の開被覆は高々可算個の部分被覆をもつとき、</div><div style="font-size: 16px;">$X$ は<b>リンデレフ空間</b>という。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この時、リンデレフ空間の例を与えます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.8</u></div><div style="font-size: 16px;">位相空間は第2可算公理を満たすならリンデレフ空間である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(X,\mathcal{O})$ を $X$ の可算開基とします。</div><div style="font-size: 16px;">この時、$\mathcal{U}$ を任意の</div><div style="font-size: 16px;">開被覆とします。この時、$\forall U\in \mathcal{U}$ に対して、</div><div style="font-size: 16px;">$\exists \mathcal{B}_U\subset \mathcal{B}(U=\cup\mathcal{B}_U)$</div><div style="font-size: 16px;">となります。</div><div style="font-size: 16px;">$$\mathcal{V}=\{B\in \mathcal{B}_U|U\in \mathcal{U}\}$$</div><div style="font-size: 16px;">と置きます。このとき、$\mathcal{V}$ は $X$ の高々可算被覆となります。</div><div style="font-size: 16px;">このとき、$B\in \mathcal{V}$ に対して、$B\subset U_B\in \mathcal{U}$</div><div style="font-size: 16px;">をとると、</div><div style="font-size: 16px;">$$X=\cup_{B\in \mathcal{V}}B\subset \cup_{B\in \mathcal{V}}U_B=X$$</div><div style="font-size: 16px;">であるから、 </div><div style="font-size: 16px;">$\{U_B\in\mathcal{U}|B\in \mathcal{V}\}$ は $\mathcal{U}$ の可算部分被覆。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>例14.5</u></div><div style="font-size: 16px;">コンパクト空間は任意の</div><div style="font-size: 16px;">開被覆は高々可算被覆をもつから、リンデレフ空間の例になります。</div><div style="font-size: 16px;">また、可分距離空間は第2可算公理を満たしますので、リンでレフの例になります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">リンデレフ空間の閉集合もリンデレフ(定理14.3と同様に証明できる)になります。</div><div style="font-size: 16px;"> </div><div style="font-size: 16px;">また、次の定理が成り立ちます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.9</u></div><div style="font-size: 16px;">第2可算公理を満たすコンパクトハウスドルフ空間は距離空間となる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $X$ を第2可算公理を満たすコンパクトハウスドルフ空間とすると、</div><div style="font-size: 16px;">先ほど書いたことから、正規空間となります。</div><div style="font-size: 16px;">第2可算公理を満たし、正規空間であるなら、ウリゾーンの距離化定理から、</div><div style="font-size: 16px;">$X$ は距離空間になります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><br /></div><div><span style="font-size: large;"><u>コンパクト距離空間</u></span></div><div style="font-size: 16px;">距離空間のコンパクト集合についての性質について</div><div style="font-size: 16px;">考えます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">すぐわかることは、以下の定理です。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.10</u></div><div style="font-size: 16px;">コンパクト距離空間は有界</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$(X,d)$ をコンパクトな距離空間とします。</div><div style="font-size: 16px;">このとき、$x\in X$ をとると、$\cup_{n=1}^\infty B_d(x,n)=X$ </div><div style="font-size: 16px;">であるから、コンパクト性より、</div><div style="font-size: 16px;">$\cup_{n=1}^mB_d(x,n)=B_d(x,m)=X$ であるから、</div><div style="font-size: 16px;">$\text{diam}(X)=\sup\{d(x,y)|x,y\in X\}\le 2m$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">となり、$X$ が有界であることがわかります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、コンパクト性は、有界より強い性質を持ちます。</div><div style="font-size: 16px;">まず、以下の定義をしておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義14.2</u> </div><div style="font-size: 16px;">距離空間 $(X,d)$ は、任意の $\epsilon>0$ に対して、ある有限集合</div><div style="font-size: 16px;">$\{x_1,\cdots,x_n\}\subset X$ が存在して、</div><div style="font-size: 16px;">$$X=\cup_{i=1}^nB_d(x_i,\epsilon)$$</div><div style="font-size: 16px;">を満たすとき、$(X,d)$ は<b>全有界</b>と言う。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">実は次の定理が成り立ちます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.11</u></div><div style="font-size: 16px;">コンパクト距離空間は全有界である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(X,d)$ をコンパクト距離空間とします。</div><div style="font-size: 16px;">この時、$\epsilon>0$ に対して、</div><div style="font-size: 16px;">$\mathcal{U}=\{B_d(x,\epsilon)|x\in X\}$ は $X$ の開被覆であるから、</div><div style="font-size: 16px;">有限部分集合 $\{B_d(x_i,\epsilon)|i=1,\cdots, n\}$ が $X$ の被覆となります。</div><div style="font-size: 16px;">よって、これは $X$ が全有界であることを意味しています。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">実際、全有界であることと、有界であることは違っていて、</div><div style="font-size: 16px;">一般に、全有界なら有界ですが、逆は成り立ちません。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(全有界 $\Rightarrow$ 有界)</div><div style="font-size: 16px;">全有界性から、$\forall \epsilon>0$ より、</div><div style="font-size: 16px;">$X=\cup_{i=1}^nB_d(x_i,\epsilon)$ が成り立ちます。</div><div style="font-size: 16px;">ここで、$\epsilon$ を固定しておきます。</div><div style="font-size: 16px;">このとき、$\{x_1,\cdots, x_n\}$ は $\{d(x_i,x_j)|i,j=1\cdots, n\}$ が最大値を持つので、</div><div style="font-size: 16px;">有界です。それを $K$ としておきます。このとき、$\epsilon$ は固定された</div><div style="font-size: 16px;">実数なので $K$ も固定された実数です。</div><div style="font-size: 16px;">このとき、$p,q\in X$ に対して、</div><div style="font-size: 16px;">$p\in B_d(x_i,\epsilon), q\in B_d(x_j,\epsilon)$ となる $x_i,x_j$ が存在するから、</div><div style="font-size: 16px;">$d(p,q)\le d(p,x_i)+d(x_i,x_j)+d(x_j,q)\le \epsilon +K+\epsilon=K+2\epsilon$ が成り立つから</div><div style="font-size: 16px;">$X$ は有界となります。 </div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また全有界性は位相的性質ではありません。</div><div style="font-size: 16px;">$(0,1)$ は全有界ですが、それと同相な ${\mathbb R}$ は全有界ではありません。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">よって距離空間において、コンパクト集合は全有界閉集合ということになります。</div><div style="font-size: 16px;">もちろん全有界閉集合は有界閉集合になります。</div><div style="font-size: 16px;">${\mathbb R}$ ではこの逆が成り立ちます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">それを示していきます。それを<b>ハイネボレルの被覆定理</b>と言います。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.12(ハイネボレルの被覆定理)</u></div><div style="font-size: 16px;">${\mathbb R}$ において $A\subset {\mathbb R}$ がコンパクトであることと</div><div style="font-size: 16px;">有界閉集合であることは同値である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) ${\mathbb R}$ において、コンパクトなら有界閉集合であることは</div><div style="font-size: 16px;">これまで示していました。なので、今回は逆の、</div><div style="font-size: 16px;">有界閉集合ならコンパクトであることを示していきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$A\subset {\mathbb R}$ が有界閉集合であるとします。</div><div style="font-size: 16px;">このとき、$\exists M>0(A\subset[-M,M])$ となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$[-M,M]$ は ${\mathbb R}$ においてコンパクトであるから、</div><div style="font-size: 16px;">$A$ はコンパクトかつ閉集合であるから、</div><div style="font-size: 16px;">定理14.3 から $A$ はコンパクトであることがわかりました。$\Box$ </div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この証明は、${\mathbb R}^n$ においても同様にすることができるので、この定理を</div><div style="font-size: 16px;">一般化して、以下のようにいうことができます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.13(ハイネボレルの被覆定理)</u></div><div style="font-size: 16px;">${\mathbb R}^n$ において $A\subset {\mathbb R}^n$ がコンパクトであることと、</div><div style="font-size: 16px;">有界閉集合であることは同値である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">一般の距離空間の場合には、この同値性は成り立ちません。</div><div style="font-size: 16px;">つまり、コンパクトではないが、有界閉集合のものは存在します。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>例14.8</u></div><div style="font-size: 16px;">$({\mathbb R},d)$ を通常のユークリッド距離位相空間ではなく、離散距離位相空間とします。</div><div style="font-size: 16px;">つまり、</div><div style="font-size: 16px;">$$d(x,y)=\begin{cases}1&x\neq y\\0&x=y\end{cases}$$</div><div style="font-size: 16px;">となる距離を入れておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">このとき、</div><div style="font-size: 16px;">${\mathbb R}$ の全ての部分集合が閉集合であり、</div><div style="font-size: 16px;">さらに、全てのことなる点の距離は1なので、${\mathbb R}$ の直径は1です。 </div><div style="font-size: 16px;">${\mathbb N}\subset{\mathbb R}$ をとれば、${\mathbb N}$ も有界で、</div><div style="font-size: 16px;">閉集合となります。</div><div style="font-size: 16px;">一方、$\{\{n\}|n\in {\mathbb N}\}$ は ${\mathbb N}$ の開被覆ですが、</div><div style="font-size: 16px;">どの有限部分集合も被覆にはなりません。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">よって、この距離空間において、コンパクトではないが、有界閉集合になっています。 </div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">次に、全有界である距離位相空間の性質を述べておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.14</u></div><div style="font-size: 16px;">全有界な距離空間は可分である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $(X,d)$ を全有界であるとします。この時、$n\in {\mathbb N}$ に対して、</div><div style="font-size: 16px;">$$X=\underset{k=1,\cdots,n}{\cup }B_d\left(x_k^n,\frac{1}{n}\right)$$ </div><div style="font-size: 16px;">となり、$D=\{x^n_k|n\in {\mathbb N},1\le k\le m_n\}$</div><div style="font-size: 16px;">とおくことで、$D$ が $X$ の可算稠密集合になっています。</div><div style="font-size: 16px;">まず、可算であることはすぐわかります。</div><div style="font-size: 16px;">稠密であることを以下示します。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$\forall x\in X$ を取ります。この時、任意の近傍 $\forall V\in \mathcal{N}(x)$ に</div><div style="font-size: 16px;">対して、$B_d(x,\frac{1}{n})\subset V$ となる $n\in {\mathbb N}$ が存在します。</div><div style="font-size: 16px;">上記の被覆性から、</div><div style="font-size: 16px;">$x\in B_d(x^n_k,\frac{1}{n})$ となる $x^n_k$ が存在します。</div><div style="font-size: 16px;">また、$x^n_k\in B_d(x,\frac{1}{n})$ であることから、</div><div style="font-size: 16px;">$x^n_k\in V\cap D$ であることから、$V\cap D=\emptyset$ であることがわかります。</div><div style="font-size: 16px;">よって、$D$ は稠密集合であることがわかります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.15</u></div><div style="font-size: 16px;">コンパクト距離位相空間は第2可算公理を満たす。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) コンパクト距離空間であれば、可分であり、</div><div style="font-size: 16px;">可分距離空間ならば第2可算公理を満たします。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">最後に以下を示して終わりました。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理14.16</u></div><div style="font-size: 16px;">次は同値になります。</div><div style="font-size: 16px;">コンパクト距離空間であること</div><div style="font-size: 16px;">$I^{\mathbb N}$ の閉集合であること</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $X$ がコンパクト距離空間であるとします。このとき、</div><div style="font-size: 16px;">第2可算公理を満たす正規空間であるから、ウリゾーンの距離化定理により</div><div style="font-size: 16px;">$I^{\mathbb N}$ に埋め込み可能になります。</div><div style="font-size: 16px;">よって、$I^{\mathbb N}$ のコンパクト集合なので、閉集合なります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">逆に、$I^{\mathbb N}$ の閉集合とすると、$I^{\mathbb N}$ は距離空間であるから</div><div style="font-size: 16px;">その部分集合も距離化可能であり、コンパクト空間の中の閉集合だから</div><div style="font-size: 16px;">コンパクト集合になります。$\Box$</div><div style="font-size: 16px;"><br /></div></div></div></div></div></div></div></div>Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-7449524749709302212022-02-07T00:39:00.004+09:002022-04-19T10:33:55.171+09:00トポロジー入門(第13回)<p> <span style="color: orange; font-family: quot; font-size: 16px;">[場所:オンライン</span><span style="color: orange; font-family: quot; font-size: 16px;">(月曜日3限)]</span></p><div style="font-family: quot; font-size: 16px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><br /></div></div></div></div></div></div><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="margin: 0px;"><div style="font-family: quot; margin: 0px;"><div style="font-size: 16px; margin: 0px;"><a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/21/top.html">トポロジー入門のHP</a></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">今回は、分離公理の後半と、コンパクト空間についての説明を行いました。</div><div style="font-size: 16px;">前回は分離公理としてハウスドルフ空間を行いました。</div><div style="font-size: 16px;">基本的に分離公理とは、2つの交わらない部分集合が開集合を使って</div><div style="font-size: 16px;">分離できるかどうかについての公理でした。</div><div style="font-size: 16px;">分離するとは、$A,B$ を部分集合として、$A\subset U$ のような</div><div style="font-size: 16px;">開集合 $U$ が存在して、$A\subset U$ かつ $U\cap B=\emptyset$</div><div style="font-size: 16px;">となることをいいます。</div><div style="font-size: 16px;">場合によっては、$B$ の方にも同じように開集合 $V$ が取れ、</div><div style="font-size: 16px;">$B\subset V$ となり、 $A\cap V=\emptyset$ となります。</div><div style="font-size: 16px;">この場合、さらに強い分離公理を満たすことになります。</div><div style="font-size: 16px;">また、さらに、$U,V$ に共通部分が内容に取れることもあり、</div><div style="font-size: 16px;">このことを、$A,B$ を<b>開集合で分離する</b>ともいいます。</div><div style="font-size: 16px;"><br /></div><div><span style="font-size: large;">分離公理(正則空間・正規空間)</span></div><div>ハウスドルフ空間 ($T_2$ 公理) の次に行う分離公理は、</div><div>$T_3, T_4$ 公理です。</div><div><br /></div><div><u>定義13.1, 13.3($T_3$-公理、$T_4$-公理) </u></div><div>「$x\in X$ と $x\not\in F$ となる閉集合 $F$ </div><div>に対して、開集合 $U,V$ が存在して、</div><div>$$x\in U,\ F\subset V,\ U\cap V=\emptyset$$</div><div>を満たす」を <b>$T_3$-公理</b>という。</div><div><br /></div><div><div>「$A\cap B=\emptyset$ を満たす閉集合 $A,B \subset X$</div><div>に対して、開集合 $U,V$ が存在して、</div><div>$$A\subset U,\ B\subset V,\ U\cap V=\emptyset$$</div><div>を満たす」を <b>$T_4$-公理</b>という。</div></div><div><br /></div><div><u>定義13.2,13.4</u></div><div>$T_1$ かつ $T_3$ 公理を満たす空間を<b>正則空間</b>といい、</div><div>$T_1$ かつ $T_4$ 公理を満たす空間を<b>正規空間</b>という。</div><div><br /></div><div>ここで、正則空間の言い換えを与えておきます。</div><div><br /></div><div><u>定理13.1</u></div><div>$X$ が $T_3$空間であることと、以下が同値、</div><div>$\forall x\in X$ と、$x\in U$ となる開集合 $U$ に対して、</div><div>$x\in V\subset \overline{V}\subset U$ となる開集合 $V$ が</div><div>存在する。</div><div><br /></div><div>(証明) ($\Rightarrow$)</div><div>$X$ が $T_3$空間とする。</div><div>このとき、$x\in U$ となる開集合 $U$ に対して</div><div>$x,F=U^c$ は1点とそれを含まない閉集合となるから、$T_3$公理から</div><div>$x\in V,F\subset W$ が存在して、$V\cap W=\emptyset$ を満たします。</div><div><br /></div><div>よって、$x\in V\subset W^c\subset U$ を満たします。</div><div>ここで、$W^c$ は閉集合であるから、</div><div>$x\in V\subset \overline{V}\subset W^c\subset U$ となります。</div><div>つまり、条件が成立することになります。</div><div><br /></div><div>($\Leftarrow$)</div><div>下の条件が成り立ったとします。</div><div>このとき、$x\not \in F$ となる閉集合 $F$ に対して、</div><div>$x\in F^c=U$ は開集合であり、条件から、$x\in V\subset \overline{V}\subset U$ を</div><div>満たす開集合 $V$ が存在します。</div><div>このとき、$x\in V$ かつ、$F\subset (\overline{V})^c$ は</div><div>$V\cap (\overline{V})^c=\emptyset$ であるから</div><div>$x,F$ を開集合によって分離していることになります。</div><div>つまり $X$ は $T_3$ 空間であることがわかりました。$\Box$</div><div><br /></div><div>同じように、以下の定理も成り立ちます。</div><div><br /></div><div><u>定理13.2</u></div><div>$T_4$ 空間であることと以下は同値.</div><div>$X$ の互いに交わらない閉集合 $F$ と$F\subset G$ を満たす</div><div>開集合 $G$ に対して、$F\subset V\subset\overline{V}\subset G$</div><div>となる開集合 $V$ が存在する。</div><div><br /></div><div>このことから、以下の関係が成り立つことがすぐにわかります。</div><div><br /></div><div>正規空間 $\Rightarrow$ 正則空間 $\Rightarrow$ ハウスドルフ空間</div><div><br /></div><div>この関係のどの逆も成り立ちません。</div><div><br /></div><div>まず、次の例があります。</div><div><br /></div><div><u>定理13.3</u></div><div>距離位相空間は正規空間。</div><div><br /></div><div>(証明) $A,B$ を交わらない閉集合とします。</div><div>このとき、</div><div>$U=\{x|d(x,A)<d(x,B)\}$</div><div>$V=\{x|d(x,A)>d(x,B)\}$</div><div>とすると、$A\subset U$ かつ $B\subset V$ であり、</div><div>定義から $U\cap V=\emptyset$ となります。</div><div>$\forall x\in A$ とすると、$d(x,A)=0$ であり</div><div>$d(x,B)>0$ となります。</div><div>もし $d(x,B)=0$ なら、$x\in \overline{B}=B$ となり矛盾するからです。</div><div><br /></div><div>また、$U,V$ が開集合であることは、$\varphi(x)=d(x,B)-d(x,A)$ は $\varphi:X\to{\mathbb R}$</div><div>となる連続関数となる。</div><div>よって $U=\varphi^{-1}((0,\infty))$ となるので、$U,V$ は開集合となる。</div><div>よって、$U,V$ は $A,B$ を分離する開集合となります。$\Box$</div><div><br /></div><div>よって、</div><div><br /></div><div>距離空間 $\Rightarrow $ 正規空間</div><div>となります。</div><div>この、逆は成り立ちません。反例は、ゾルゲンフライ直線です。</div><div><br /></div><div>また、正規空間が成り立つ性質についてまとめておきます。</div><div>証明はしません。</div><div><br /></div><div><u>定理13.4(ウリゾーンの補題)</u></div><div>位相空間 $X$ において以下は同値。</div><div>$T_4$ 空間である。</div><div>互いに交わらない閉集合 $F,G$ に対して、</div><div>連続関数 $f:X\to I$ が存在して $f(F)=0$ かつ $f(G)=1$ </div><div>を満たす。</div><div><br /></div><div><u>定理13.5(ウリゾーンの距離化定理)</u></div><div>正規かつ第2可算公理を満たす空間は距離化可能</div><div><br /></div><div><u>定理13.6(ティーチェの拡張定理)</u></div><div>$X$ を正規空間とする。</div><div>このとき、閉集合 $A\subset X$ に対して</div><div>任意の連続関数 $f:A\to X$ に対して連続関数</div><div>$\tilde{f}:X\to [0,1]$ が存在して、$\tilde{f}|_A=f$ となる。</div><div><br /></div><div><u>ゾルゲンフライ平面 ${\mathbb R}^2_l$</u></div><div>を考えましょう。<a href="https://motochans.blogspot.com/2014/06/blog-post_25.html" target="">こちら</a> にてゾルゲンフライ直線、平面についての解説を</div><div>書いたことがあったのでリンクをはりました。</div><div>ゾルゲンフライ直線の2つの直積としてゾルゲンフライ平面を定義します。</div><div><br /></div><div>ティーチェの拡張定理を使うと、</div><div>${\mathbb R}^2_l$ は正規空間ではないことが分かります。</div><div>また、正則空間の2つの直積空間も正則空間</div><div>であるから、</div><div>${\mathbb R}^2_l$ は正則空間となります。</div><div>よって、${\mathbb R}^2_l$ は正則だが、正規な位相空間ということになります。</div><div><br /></div><div>他にも、ハウスドルフだが、正則空間ではない空間も存在しますが</div><div>ここでは紹介しません。</div><div><br /></div><div><span style="font-size: large;"><u>コンパクト空間</u></span></div><div>次にコンパクト空間について解説します。</div><div><br /></div><div>$\mathcal{U}\subset \mathcal{P}(X)$ が<b>被覆</b>であるとは、</div><div>$\forall x\in X$ に対して、$\exists U\in \mathcal{U}$ となる</div><div>ときをいう。</div><div>つまり、$X=\cup \mathcal{U}$ のことと同値。</div><div><br /></div><div>$\mathcal{U}$ が被覆かつ $\mathcal{U}\subset \mathcal{O}$ となるとき、</div><div>$\mathcal{U}$ は<b>開被覆</b>という。</div><div><br /></div><div><div>$\mathcal{U}$ が被覆かつ $\mathcal{U}\subset \mathcal{C}$ となるとき、</div><div>$\mathcal{U}$ は<b>閉被覆</b>という。</div><div><br /></div><div>また、$\mathcal{U}$ が被覆かつ $|\mathcal{U}|<\infty$ を満たすとき、</div><div><b>有限被覆</b>という。</div><div><br /></div></div><div>また、被覆 $\mathcal{U}$ が $\mathcal{V}\subset \mathcal{U}$</div><div>を満たす $\mathcal{V}$ が被覆であるとき、<b>部分被覆</b>という。</div><div><br /></div><div>$\mathcal{V}\subset\mathcal{U}$ が部分被覆かつ有限被覆であるとき、</div><div><b>有限部分被覆</b>という。</div><div><br /></div><div>ここでコンパクト空間の定義をしましょう。</div><div><br /></div><div><u>定理13.5</u></div><div>位相空間 $(X,\mathcal{O})$ に対して、</div><div>$X$ の任意の開被覆 $\mathcal{U}$ に対して、有限部分被覆が存在するとき、</div><div>$(X,\mathcal{O})$ は<b>コンパクト空間</b>という。</div><div><br /></div><div>同様に、 $A\subset X$ が<b>コンパクト集合</b>であることは、</div><div>$A$ が部分空間としてコンパクト空間であることである。</div><div><br /></div><div>言い換えれば、位相空間 $(X,\mathcal{O})$ において、</div><div>$\mathcal{U}\subset \mathcal{O}$ が</div><div>$A\subset \cup\mathcal{U}$ を満たすとき、有限部分集合 $\mathcal{V}\subset \mathcal{U}$ </div><div>が存在して、$A\subset \cup\mathcal{V}$ を満たすことをいいます。</div><div><br /></div><div>コンパクト空間 $X$ に対して、以下の定理が成り立ちます。</div><div><br /></div><div><u>定理13.7</u></div><div>$X$ がコンパクト空間であり、</div><div>$f:X\to Y$ が連続写像であるとき、$f(X)$ もコンパクトである。</div><div><br /></div><div>(証明) $\mathcal{U}$ を $f(X)$ の開被覆とする。</div><div>$f(X)\subset \cup\mathcal{U}$ を満たすとする。</div><div>$\{f^{-1}(U)|U\in \mathcal{U}\}=\mathcal{A}$ は $X$ の開被覆となります。</div><div>$X$ のコンパクトであるから、有限集合 $\mathcal{V}\subset\mathcal{U}$ </div><div>が存在して、$\cup\{f^{-1}(V)|V\in \mathcal{V}\}$ は $X$ の開被覆</div><div>となります。</div><div>よって、$\mathcal{V}$ は $f(X)$ の開被覆となります。</div><div>つまり、$f(X)$ はコンパクトとなります。$\Box$ </div><div><br /></div><div>この定理から、$X$ の任意のコンパクト集合の連続写像による像(つまり連続像)</div><div>もコンパクト集合ということになります。</div><div><br /></div><div>例として、${\mathbb R}$ のコンパクト集合を考えます。</div><div><br /></div><div><u>例13.7</u></div><div>$({\mathbb R},\mathcal{O}_{d_1})$ のコンパクト集合は有界である。</div><div><br /></div><div>(証明) $A$ を有界でない集合とします。</div><div>$\forall n$ に対して、$U_n=(-n,n)$ とします。</div><div>このとき、$\mathcal{U}=\{U_n|n\in {\mathbb N}\}$ は $A$ の開被覆である。</div><div>$\mathcal{U}$ の任意の有限集合 $\mathcal{V}\subset \mathcal{U}$ の和集合</div><div>は、ある開区間$(-M,M)$ に包まれ、$\cup\mathcal{V}=(-M,M)$ となる。$A$ は有界ではないから、$A\not\subset\cup\mathcal{V}$ となります。</div><div>これは $\mathcal{V}$ は $A$ の被覆ではない。</div><div>つまり、$\mathcal{U}$ には有限部分被覆が存在しないことになります。</div><div>よって、$A$ はコンパクトではありません。</div><div>この対偶をとることで、$A$ がコンパクト集合なら有界ではないということが</div><div>成り立ちます。$\Box$</div><div><br /></div><div><u>例13.8</u></div><div>$({\mathbb R},\mathcal{O}_{d_1})$ のコンパクト集合は</div><div>閉集合である。</div><div><br /></div><div>$A\subset {\mathbb R}$ をコンパクト集合とする。</div><div>$A$ が閉集合でないとする。</div><div>$x\in \overline{A}\setminus A$ をとる。</div><div>このとき、$\forall \epsilon >0(Cl(B_{d_1}(x,\epsilon))\cap A\neq\emptyset))$</div><div>とする。このとき、</div><div>$U_\epsilon=[Cl(B_{d_1}(x,\epsilon))]^c$</div><div>とする。</div><div>このとき、$\{x\}=\cap_{\epsilon\in {\mathbb R}_{>0}}Cl(B_{d_1}(x,\epsilon))$</div><div>($X$ がハウスドルフ空間であることと同値の主張になります。<a href="https://motochans.blogspot.com/2020/01/12.html">こちら</a>を見てください。)</div><div>であるから、${\mathbb R}\setminus\{x\}=\cup_{x\in {\mathbb R}_{>0}}U_\epsilon$</div><div>となる。</div><div>よって、$\mathcal{U}=\{U_\epsilon|\epsilon>0\}$</div><div>は $A$ の開被覆となります。</div><div><br /></div><div>$\mathcal{V}=\{U_{\epsilon_i}|i=1,2,\cdots, n\}$ は </div><div>$\mathcal{U}$ の任意の有限部分集合とする。</div><div>$\epsilon=\min\{\epsilon_i|i=1,\cdots, n\}$ とする。このとき、</div><div>$$(\cup\mathcal{V})^c=\cap_{i=1}^nCl(B_{d_1}(x,\epsilon_i))=Cl(B_{d_1}(x,\epsilon))$$</div><div><br /></div><div>よって、$\cup\mathcal{V}=U_\epsilon$ となります。</div><div>$Cl(B_{d_1}(x,\epsilon))\cap A\neq \emptyset$ であり、</div><div>$A\setminus U_\epsilon\neq\emptyset$ つまり、$A\not\subset\cup\mathcal{V}$ であるから、</div><div>$\mathcal{V}$ は $A$ の開被覆にならないので、</div><div>$\mathcal{U}$ は有限部分被覆をもたないことになります。</div><div>ゆえに $A$ はコンパクトではありません。$\Box$</div><div><br /></div><div>${\mathbb R}$ のコンパクト集合はどんなものがあるでしょうか。</div><div><br /></div><div>最後に次の定理を示します。</div><div><u><br /></u></div><div><u>定理<span>13.8</span></u></div><div>閉区間 $[0,1]$ はコンパクト集合</div><div><br /></div><div>(証明) $[0,1]$ がコンパクトでないとする。</div><div>$\mathcal{U}$ が $[0,1]$ の開被覆で、どんな有限部分集合も</div><div>被覆にならないとする。</div><div><br /></div><div>$[0,1/2],[1/2,1]$ のうち、どちらかは、有限部分被覆を持ちません。</div><div>もし、両方とも部分被覆を持つとすると、</div><div>それらを合わせて、$[0,1]$ の有限部分被覆を持つからです。</div><div><br /></div><div>よって、それを $[a_1,b_1]$ とします。</div><div>また、$[a_1,\frac{a_1+b_1}{2}],[\frac{a_1+b_1}{2},b_1]$ のうち、</div><div>どちらかは有限部分被覆を持たないことになります。もし持つとすると、</div><div>それらを合わせて、$[a_1,b_1]$ の有限部分被覆をもちます。</div><div>それを $[a_2,b_2]$ とします。</div><div>このようにして、$[a_1,b_1]\supset [a_2,a_2]\supset [a_3,b_3]\supset \cdots $</div><div>を作っていきます。</div><div>$[a_n,b_n]\supset[a_{n+1},b_{n+1}]$ は、有限部分被覆を持ちません。</div><div><br /></div><div>このとき、$a_n$ は単調増加であり、$b_n$ は単調減少です。</div><div>$0\le a_n\le b_n\le 1$ であるから、数列 $a_n,b_n$ は実数列の条件から</div><div>収束します。</div><div>$\text{diam}([a_n,b_n])=|a_n-b_n|=\frac{1}{2^n}\to 0$ であるから、</div><div>$a_n,b_n$ は同じ実数 $x$ に収束します。</div><div><br /></div><div>このとき、$x\in U\in \mathcal{U}$ となる開集合 $U$ が存在して、</div><div>$x\in B_{d_1}(x,\epsilon)\subset U$ となる $\epsilon$ も存在します。</div><div>また、$1/2^n\to 0$ であるから、$1/2^{n}<\epsilon$ となる自然数 $n$ が</div><div>存在するから、</div><div>$[a_n,b_n]\subset B_{d_1}(x,\epsilon)\subset U$ です。</div><div>しかし、$[a_n,b_n]$ は $\mathcal{U}$ の有限部分被覆 $\{U\}$ が</div><div>存在することになります。これは、$[a_n,b_n]$ が有限部分被覆が存在しないことに</div><div>反します。</div><div><br /></div><div>ゆえに、$\mathcal{U}$ は有限部分被覆を持つということになり、</div><div>$[0,1]$ はコンパクト集合ということになります。$\Box$</div></div><div style="margin: 0px;"><div style="margin: 0px;"><div></div></div></div></div></div></div></div></div>Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-89600152258682202502022-01-27T18:07:00.008+09:002023-01-27T21:26:17.307+09:00トポロジー入門(第12回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所:オンライン</span><span style="color: orange; font-family: inherit;">(月曜日3限)]</span></div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div>
<div style="margin: 0px;">
<div style="font-family: quot; font-size: 16px; margin: 0px;">
<div style="margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/21/top.html">トポロジー入門のHP</a></div></div><div style="margin: 0px;"><div><div><div style="margin: 0px;"><div><div style="font-family: quot; font-size: 16px;">
<br />
<br />今回は、連結性の残りと分離公理についてやりました。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot;"><span style="font-size: large;"><u>連結性</u></span><br /><br />連結に対して、連結成分があったように、</div><div style="font-family: quot; font-size: 16px;">弧状連結に対しても弧状連結成分があります。</div><div style="font-family: quot; font-size: 16px;">位相空間に対して、$a\in X$ に対して</div><div style="font-family: quot; font-size: 16px;">弧 $f:[0,1]=I\to X$ で $f(0)=a$ となるもので、</div><div style="font-family: quot; font-size: 16px;">$f(1)$ を集めたもの $\{f(1)|f:I\to X,f(0)=a\}$ を</div><div style="font-family: quot; font-size: 16px;">$a$ の<b>弧状連結成分</b>といい、$C_{\text{path}}(a)$ とかきます。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">例えば、前回の定理11.8で構成した$I\cup J$ は</div><div style="font-family: quot; font-size: 16px;">$I, J$ がそれぞれが弧状連結成分になります。</div><div style="font-family: quot; font-size: 16px;">ゆえに、弧状連結成分は閉集合になるとは限らず、開集合になるとも</div><div style="font-family: quot; font-size: 16px;">限りません。</div><div style="font-family: quot; font-size: 16px;">また、$a$ の弧状連結成分は、$a$ を含む弧状連結集合の中で最大のもの</div><div style="font-family: quot; font-size: 16px;">となります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">また、次の連結性についても解説しました。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;"><u>定義12.3</u></div><div style="font-family: quot; font-size: 16px;">位相空間$(X,\mathcal{O})$ が $\forall x\in X$ と、</div><div style="font-family: quot; font-size: 16px;">任意の近傍 $\forall V\in \mathcal{N}(x)$ に対して、</div><div style="font-family: quot; font-size: 16px;">ある連結な近傍 $U\in \mathcal{N}(x)$ が存在して、$U\subset V$ を</div><div style="font-family: quot; font-size: 16px;">満たすとき、$(X,\mathcal{O})$ は<b>局所連結</b>であるという。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">局所連結であることは、</div><div style="font-family: quot; font-size: 16px;">基本近傍系として連結なものが取れることと同値になります。</div><div style="font-family: quot; font-size: 16px;">特に、近傍として連結なものが取れる必要があります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">同様に、以下の定義をすることもできます。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;"><u>定義12.4</u></div><div style="font-family: quot; font-size: 16px;"><div>位相空間$(X,\mathcal{O})$ が $\forall x\in X$ と、</div><div>任意の近傍 $\forall V\in \mathcal{N}(x)$ に対して、</div><div>ある弧状連結な近傍 $U\in \mathcal{N}(x)$ が存在して、$U\subset V$ を</div><div>満たすとき、$(X,\mathcal{O})$ は<b>局所弧状連結</b>であるという。</div><div><br /></div><div><br /></div><div>連結ならば弧状連結であり、</div><div>局所連結ならば局所弧状連結になることも同様です。</div><div>また、同様に、各点において弧状連結な近傍が取れます。</div><div>これ以外に関係がないのかと思われるかもしれませんが</div><div>以下の定理があります。</div><div><br /></div></div><div style="font-family: quot; font-size: 16px;"><u>
定理12.1</u><br />
連結かつ局所弧状連結なら弧状連結である。<br />
<br />
証明<br />
$a\in X$ に対して、$C_{\text{path}}(a)$ が開かつ閉集合であることを示せば、</div><div style="font-family: quot; font-size: 16px;">連結性から $C_{\text{path}}(a)=X$ となり $X$ が弧状連結であることがわかります。</div><div style="font-family: quot; font-size: 16px;">
$\forall x\in C_{\text{path}}(a)$ に対して、弧状連結な近傍が取れるので、</div><div style="font-family: quot; font-size: 16px;">それを $V$ とすると、最大性から、</div><div style="font-family: quot; font-size: 16px;">$V\subset C_{\text{path}}(x)$ であることがわかります。</div><div style="font-family: quot; font-size: 16px;">よって、$x\in C_{\text{path}}(a)$ は内点であるから、$C_{\text{path}}(a)$ は</div><div style="font-family: quot; font-size: 16px;">開集合となります。<br /><br /></div><div style="font-family: quot; font-size: 16px;">$x\in\overline{ C_{\text{path}}(a)}$ とします。</div><div style="font-family: quot; font-size: 16px;">任意の近傍 $V\in \mathcal{N}(x)$ に対して、<br /></div><div style="font-family: quot; font-size: 16px;">弧状連結な近傍 $U\in \mathcal{N}(x)$ が存在して、</div><div style="font-family: quot; font-size: 16px;">$U\subset V$ であることがわかります。</div><div style="font-family: quot; font-size: 16px;">ここで、$U\cap C_{\text{path}}(a)\neq \emptyset$</div><div style="font-family: quot; font-size: 16px;">であり、$U\cup C_{\text{path}}(a)$ も弧状連結であるから、</div><div style="font-family: quot; font-size: 16px;">弧状連結成分の最大性から、$U\subset C_{\text{path}}(a)$ が成り立ちます。</div><div style="font-family: quot; font-size: 16px;">特に、$x\in C_{\text{path}}(a)$ が成り立ちます。</div><div style="font-family: quot; font-size: 16px;">よって、$\overline{C_{\text{path}}(a)}\subset C_{\text{path}}(a)$ であることがわかります。</div><div style="font-family: quot; font-size: 16px;">また、$C_{\text{path}}(a)\subset \overline{C_{\text{path}}(a)}$ であることから、</div><div style="font-family: quot; font-size: 16px;">$\overline{C_{\text{path}}(a)}=C_{\text{path}}(a)$ であることがわかります。</div><div style="font-family: quot; font-size: 16px;">よって、$C_{\text{path}}(a)$ が閉集合であることがわかります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">よって、$C_{\text{path}}(a)$ が開かつ閉集合であることがわかりました。$\Box$</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">以上より、連結性を4つ紹介して、その間に成り立つ定理について</div><div style="font-family: quot; font-size: 16px;">まとめましたが、実際、以下のようにそれらが成り立つ位相空間が存在します。</div><div style="font-family: quot; font-size: 16px;">以下にそのような位相空間の例を挙げておきます。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">
<br />
<table border="0" cellpadding="0" cellspacing="0" style="border-collapse: collapse; color: black; text-align: center; width: 385px;"><colgroup><col span="5" style="width: 77pt;" width="77"></col></colgroup><tbody>
<tr height="18" style="height: 18pt;"><td class="xl63" height="18" style="border: 0.5pt solid windowtext; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; height: 18pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap; width: 77pt;" width="77"> </td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: solid solid solid none; border-top-color: windowtext; border-top-width: 0.5pt; border-top: 0.5pt solid windowtext; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap; width: 77pt;" width="77">連結</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: solid solid solid none; border-top-color: windowtext; border-top-width: 0.5pt; border-top: 0.5pt solid windowtext; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap; width: 77pt;" width="77">局所連結</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: solid solid solid none; border-top-color: windowtext; border-top-width: 0.5pt; border-top: 0.5pt solid windowtext; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap; width: 77pt;" width="77">弧状連結</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: solid solid solid none; border-top-color: windowtext; border-top-width: 0.5pt; border-top: 0.5pt solid windowtext; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap; width: 77pt;" width="77">局所弧状連結</td></tr>
<tr height="18" style="height: 18pt;"><td class="xl63" height="18" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-color: windowtext; border-left-width: 0.5pt; border-left: 0.5pt solid windowtext; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; height: 18pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap;">1点 </td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap;">◯</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap;">◯</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap;">◯</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; vertical-align: bottom; white-space: nowrap;">◯</td></tr>
<tr height="18" style="height: 18pt;"><td class="xl63" height="18" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-color: windowtext; border-left-width: 0.5pt; border-left: 0.5pt solid windowtext; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; height: 18pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">(※)</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;"> ◯ </td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">◯</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;"> ◯ </td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">×</td></tr>
<tr height="18" style="height: 18pt;"><td class="xl63" height="18" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-color: windowtext; border-left-width: 0.5pt; border-left: 0.5pt solid windowtext; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; height: 18pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">ワルシャワ円</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">◯</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">×</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">◯</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">×</td></tr>
<tr height="18" style="height: 18pt;"><td class="xl63" height="18" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-color: windowtext; border-left-width: 0.5pt; border-left: 0.5pt solid windowtext; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; height: 18pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">${\mathbb L}’$</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">◯</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">◯</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">×</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">×</td></tr>
<tr height="18" style="height: 18pt;"><td class="xl63" height="18" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-color: windowtext; border-left-width: 0.5pt; border-left: 0.5pt solid windowtext; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; height: 18pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">定理11.8</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">◯</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">×</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">×</td><td class="xl63" style="border-bottom-color: windowtext; border-bottom-width: 0.5pt; border-bottom: 0.5pt solid windowtext; border-left-style: none; border-right-color: windowtext; border-right-width: 0.5pt; border-right: 0.5pt solid windowtext; border-style: none solid solid none; border-top-style: none; font-family: "MS Pゴシック", sans-serif; font-size: 12pt; padding-left: 1px; padding-right: 1px; padding-top: 1px; text-align: center; vertical-align: bottom; white-space: nowrap;">×</td></tr>
</tbody></table><br />
</div><div style="font-family: quot; font-size: 16px;">連結性を満たさない例については、各例に1点を加えれば実現できますので省略しています。</div><div style="font-family: quot; font-size: 16px;">ワルシャワ円(ワルシャワサークル)とは、定理11.8(トポロジストのサインカーブ)の $I,J$ </div><div style="font-family: quot; font-size: 16px;">をつなげて連結にしたものを言います。<a href="https://en.wikipedia.org/wiki/Shape_theory_(mathematics)#Warsaw_Circle">こちら</a>にその絵があります。</div><div style="font-family: quot; font-size: 16px;">(※) の空間は ${\mathbb R}$ 上の補可算位相空間の cone があります。</div><div style="font-family: quot; font-size: 16px;">位相空間 $X$ のcone $C(X)$とは $X\times I$ に $X\times \{1\}$ を1点に</div><div style="font-family: quot; font-size: 16px;">潰して得られる商位相空間です。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">また、補可算位相空間とは、開集合として補集合が高々可算個の点集合のものとする位相空間で、</div><div style="font-family: quot; font-size: 16px;">${\mathbb R}$ 上で考えたものは、連結かつ局所連結だが、弧状連結ではなく局所弧状連結ではない空間になります。そのconeをとることで弧状連結だけが満たされて、</div><div style="font-family: quot; font-size: 16px;">この条件を満たす空間ということになります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">また、${\mathbb L}$ は長い直線と呼ばれ、$[0,1)$ を非可算無限個つなげてできる</div><div style="font-family: quot; font-size: 16px;">位相空間です。可算無限個つなげてできる位相空間は${\mathbb R}$ と同相になります。</div><div style="font-family: quot; font-size: 16px;">それも<a href="https://ja.wikipedia.org/wiki/長い直線">こちら</a>に解説があります。</div><div style="font-family: quot; font-size: 16px;">ここで、 ${\mathbb L}'$ は無限遠点 $\{\infty\}$ を付け加えてできる空間になります。</div><div style="font-family: quot; font-size: 16px;">付け加える(コンパクト化と言います)ことはまだ習っていませんが、$(0,1)$ に</div><div style="font-family: quot; font-size: 16px;">1を加えて $(0,1]$ を作る操作だと考えてください。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot;"><u><span style="font-size: large;">分離公理</span></u></div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">後半はハウスドルフ空間についておこないました。</div><div style="font-family: quot; font-size: 16px;">位相空間の部分集合 $A,B$ が<b>開集合によって分離される</b>とは、</div><div style="font-family: quot; font-size: 16px;">$U,V$ を開集合として、$A\subset U$, $V\subset B$ を満たし、$U\cap V=\emptyset$</div><div style="font-family: quot; font-size: 16px;">を満たすことを言います。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;"><u>定義12.5</u></div><div style="font-family: quot; font-size: 16px;">位相空間 $(X,\mathcal{O})$ の任意の相異なる点 $p,q\in X$ に対して、</div><div style="font-family: quot; font-size: 16px;">開集合 $U,V$ が存在して、$p\in U,q\in V$ かつ $U\cap V=\emptyset$</div><div style="font-family: quot; font-size: 16px;">を満たす時、 $X$ は<b>ハウスドルフ空間</b>という。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">つまり、任意の相異なる2点が開集合によって分離される時にハウスドルフ空間と</div><div style="font-family: quot; font-size: 16px;">言うのです。</div><div style="font-family: quot; font-size: 16px;">公理「任意の相異なる2点が開集合によって分離される」を<b> $T_2$ 公理</b>ともいうので、</div><div style="font-family: quot; font-size: 16px;">ハウスドルフ空間のことを <b>$T_2$ 空間</b>ということもあります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;"><u>例12.4</u></div><div style="font-family: quot; font-size: 16px;">距離空間はハウスドルフ空間になります。</div><div style="font-family: quot; font-size: 16px;">距離空間の相異なる点 $p,q\in X$ に対して、</div><div style="font-family: quot; font-size: 16px;">$2\delta=d(p,q)$ とすることで、$p,q$ には、$B_d(p,\delta)$ と $B_d(q,\delta)$</div><div style="font-family: quot; font-size: 16px;">が $p,q$ を分離することがわかります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;"><u>例12.5</u></div><div style="font-family: quot; font-size: 16px;">$n$ 点集合が離散位相を持つならそれはハウスドルフ空間であることもすぐわかります。</div><div style="font-family: quot; font-size: 16px;">逆に、有限点集合がハウスドルフなら離散位相空間になります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">よって、有限点集合上の離散位相ではないものは皆 $T_2$ 公理を満たさないことになります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">また、ハウスドルフ空間は位相的性質であることは簡単にわかります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;"><u>定理12.3 </u></div><div style="font-family: quot; font-size: 16px;">ハウスドルフ性は位相的性質である。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;">また、ハウスドルフ空間の同値な言い換えがsいくつかあります。</div><div style="font-family: quot; font-size: 16px;"><br /></div><div style="font-family: quot; font-size: 16px;"><u>定理12.3, 12.4</u></div><div style="font-family: quot; font-size: 16px;">ハウスドルフ空間であることは以下のそれぞれと同値</div><div><ul style="text-align: left;"><li><span style="font-family: quot;">$\Delta=\{(x,x)|x\in X\}$ が閉集合である。</span></li><li><span style="font-family: quot;">$\{x\}=\cap \{\overline{W}|W\in \mathcal{N}(x)\}$ である。</span></li></ul><div><br /></div><div><span style="font-family: quot;">(証明)ハウスドルフ空間であれば、$(p,q)\in \Delta^c$ は、$p\neq q$ であるから、</span></div><div><span style="font-family: quot;">開集合 $U,V$ が存在して、$p\in U,q\in V$ を満たし、$U\cap V=\emptyset$ を満たす。</span></div><div><span style="font-family: quot;">よって、$U\times V\subset \Delta^c$ であるから、$(p,q)\in \Delta^c$ は</span></div><div><span style="font-family: quot;">内点。つまり$\Delta^c$ は閉集合。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">逆に$\Delta^c$が閉集合であれば、$p\neq q\in X$ に対して、</span></div><div><span style="font-family: quot;">$(p,q)\in W\subset \Delta^c$ となる開集合 $W$ が存在して、</span></div><div><span style="font-family: quot;">直積位相から、$(p,q)\subset U\times V\subset W$ が成り立ちます。</span></div><div><span style="font-family: quot;">$U\times U\cap \Delta^c=\emptyset$ であるから、</span></div><div><span style="font-family: quot;">$U\cap V=\emptyset$ となり、$X$ はハウスドルフ空間であることがわかります。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">$X$ がハウスドルフ空間であるとします。</span></div><div><span style="font-family: quot;">このとき、$x\in X$ に対して $x\neq y$ となる $y$ を取ります。</span></div><div><span style="font-family: quot;">この時、開集合$U,V$ が存在して、$x\in U,y\in V$ とし、</span></div><div><span style="font-family: quot;">$U\cap V=\emptyset$ となります。$x\in U\subset V^c$ であり、</span></div></div><div><span style="font-family: quot;">$V^c$ は閉集合であるから、</span></div><div><span style="font-family: quot;">$x\in U\subset \overline{U}\subset V^c$ であるから、</span></div><div><span style="font-family: quot;">$y\not\in \overline{U}$ であるから特に、</span></div><div><span style="font-family: quot;">$y\not\in \{\overline{W}|W\in \mathcal{N}(x)\}$</span></div><div><span style="font-family: quot;">となる。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">逆</span><span style="font-family: quot;">に、$\forall x\in X$ に対して、$\{x\}=\cap\{\overline{W}|W\in \mathcal{N}(x)\}$</span></div><div><span style="font-family: quot;">が成り立つとすると、$y\neq y$ とすると、</span></div><div><span style="font-family: quot;">$W\in \mathcal{N}(x)$ が存在して、$y\not\in \overline{W}$ となります。</span></div><div><span style="font-family: quot;">$V=(\overline{W})^c$ また、$x\in U\subset W$ となる開集合 $U$ が存在するので、</span></div><div><span style="font-family: quot;">$x\in U$ かつ $y\in V$ かつ $U\cap V=\emptyset$ となります。</span></div><div><span style="font-family: quot;">よって$X$ がハウスドルフ空間となります。$\Box$</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">また、次の定理を示しました。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;"><u>定理12.5</u></span></div><div><span style="font-family: quot;">$X$ を位相空間、$Y$ ハウスドルフ空間とする。</span></div><div><span style="font-family: quot;">$F:G:X\to Y$ を連続写像とする。今、$D\subset X$ を稠密集合とする。</span></div><div><span style="font-family: quot;">$F|_{D}=G|_{D}$ であるとすると、$F=G$ が成り立つ。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">関数 $F,G$ が等しいというのは、$\forall x\in X$ に対して、</span></div><div><span style="font-family: quot;">$F(x)=G(x)$ となることを意味します。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">(証明) </span></div><div><span style="font-family: quot;">$F,G:X\to Y$ に対して、ハウスドルフ空間 $Y$ に対して、$F\neq G$</span></div><div><span style="font-family: quot;">であるとします。</span></div><div><span style="font-family: quot;">この時、$F(x)\neq G(x)$ となる $x\in X$ が存在します。</span></div><div><span style="font-family: quot;">ハウスドルフ性から、$Y$ の開集合 $U,V$ が存在して、$F(x)\in U$, $G(x)\in V$</span></div><div><span style="font-family: quot;">かつ $U\cap V=\emptyset$ となります。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">$x\in F^{-1}(U)\cap G^{-1}(V)$ であり、$F,G$ が連続であることから、</span></div><div><span style="font-family: quot;">$F^{-1}(U)\cap G^{-1}(V)$ は $X$ の開集合となります。</span></div><div><br /></div><div><span style="font-family: quot;">よって、$F^{-1}(U)\cap G^{-1}(V)\cap D\neq \emptyset$ ですから、</span></div><div><span style="font-family: quot;">$d\in D$ であって、$d\in F^{-1}(U)\cap G^{-1}(V)$ を満たします。</span></div><div><span style="font-family: quot;">よって、$F(d)\in U$ かつ $G(d)\in V$ となります。$U\cap V=\emptyset$ であることから</span></div><div><span style="font-family: quot;">少なくとも、$F(d)\neq G(d)$ であるから $F|_D\neq G|_D$ となります。$\Box$</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">この定理から、$X$ 上の ${\mathbb R}$ に値をもつ連続関数の集合 $C(X)$</span></div><div><span style="font-family: quot;">の濃度を決定することができます。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">$X={\mathbb R}$ とすると、可算集合 ${\mathbb Q}$ からの連続関数</span></div><div><span style="font-family: quot;">によって一意的に決定されるから、</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">$$|C({\mathbb R})|\le |{\mathbb R}^{\mathbb Q}|= (2^{\aleph_0})^{\aleph_0}\le 2^{\aleph_0\times \aleph_0}\le 2^{\aleph_0}$$</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">となります。一方、$C({\mathbb R})$ には定数関数が ${\mathbb R}$ の分だけ存在するので、</span></div><div><span style="font-family: quot;">$|C({\mathbb R})|\ge |{\mathbb R}|=2^{\aleph_0}$ となり、</span></div><div><span style="font-family: quot;">$|C({\mathbb R})|=2^{\aleph_0}=|{\mathbb R}|$ が成り立ちます。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">つまり、$C({\mathbb R})$ は 連続体濃度存在することがわかります。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">又、ハウスドルフ空間に対して以下の性質が成り立ちます。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;"><u>定理12.6</u></span></div><div><span style="font-family: quot;">ハウスドルフ空間の任意の部分空間もハウスドルフ空間である。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;"><u>定理12.7</u></span></div><div><span style="font-family: quot;">全ての因子空間がハウスドルフ空間であるような任意の直積位相空間も</span></div><div><span style="font-family: quot;">ハウスドルフ空間である。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">これらの証明はそれほど難しくないので、証明はここでは省略します。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">また、$T_2$ 公理を弱くした次の $T_1$ 公理もあります。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;"><u>定義12.6</u></span></div><div><span style="font-family: quot;">位相空間 $(X,\mathcal{O})$ が相異なる $x,y$ に対して、</span></div><div><span style="font-family: quot;">$\exists U,V\in \mathcal{O}((x\in U,y\not\in U)\wedge(x\not\in V,y\in V))$</span></div><div><span style="font-family: quot;">を<b>$T_1$ 公理</b>と言う。$T_1$ 公理を満たす空間を</span></div><div><span style="font-family: quot;"><b>$T_1$ 空間</b>という。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">$T_1$ 公理は、</span><span style="font-family: quot;">$\exists U,V\in \mathcal{O}(x\in U,y\not\in U))$</span></div><div><span style="font-family: quot;">というだけで同じことです。</span></div><div><span style="font-family: quot;">というのも、$x,y$ を $y,x$ にするだけで、もう一つの条件も満たすからです。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">ここで $T_1$ 空間には次の性質があります。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;"><u>定理12.8</u></span></div><div><span style="font-family: quot;">$(X,\mathcal{O})$ が $T_1$ 空間であることと、 $X$ の各点が閉集合であることは</span></div><div><span style="font-family: quot;">同値である。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">(証明) $X$ が$T_1$ 空間であるとします。</span></div><div><span style="font-family: quot;">$\forall x\in X$ に対して、$y\in\{x\}^c$ をとります。</span></div><div><span style="font-family: quot;">この時、$\exists U\in\mathcal{O}$ が存在して、$y\in U$ かつ$x\not\in U$ </span></div><div><span style="font-family: quot;">つまり、$U\subset \{x\}^c$ となる。つまり、$y$ は $\{x\}^c$ の内点。</span></div><div><span style="font-family: quot;">よって、$\{x\}^c$ は開集合であるから、$x$ は閉集合となります。</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;">$\forall x\in X$ に対して $\{x\}$ が閉集合であるとします。</span></div><div><span style="font-family: quot;">このとき、$x,y\in X $を $x\neq y$ とします。</span></div><div><span style="font-family: quot;">$U=\{y\}^c$ と $V=\{x\}^c$ とおきます。</span></div><div><span style="font-family: quot;">この時、$U,V$ は開集合であり、$x\in U,y\not\in U$ かつ $y\in V,x\not\in V$</span></div><div><span style="font-family: quot;">となります。つまり、$X$ は $T_1$ 空間となります。 $\Box$</span></div><div><span style="font-family: quot;"><br /></span></div><div><span style="font-family: quot;"><u>例12.8</u></span></div><div><span style="font-family: quot;">$T_1$ 空間であってハウスドルフ空間ではないものが存在します。</span></div><div><span style="font-family: quot;">例えば、無限集合上の補有限位相 $(X,\mathcal{O}_{\text{cf}})$ を取りますと、</span></div><div><span style="font-family: quot;">$(X,\mathcal{O}_{\text{cf}})$ は $T_1$ 空間にはなりますが、</span></div><div><span style="font-family: quot;">ハウスドルフ空間にはなりません。</span></div><div><span style="font-family: quot;">実際、空ではない任意の開集合が交わりを持つことがわかります。</span></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-29282372349292694052022-01-05T16:31:00.092+09:002022-04-19T10:34:48.110+09:00トポロジー入門(第11回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所オンライン(月曜日3限)]</span></div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div>
<div style="margin: 0px;">
<div style="font-size: 16px; margin: 0px;">
<div style="margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/21/top.html">トポロジー入門のHP</a></div>
</div>
<div style="margin: 0px;">
<div>
<div>
<div style="margin: 0px;">
<div>
<div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">今回は、連結性と弧状連結性についてやりました。</div><div style="font-size: 16px;"><br /></div><div><span style="font-size: large;"><b><u>連結性</u></b></span></div><div style="font-size: 16px;">位相空間が連結であることは前回定義しました。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義10.3</u></div><div style="font-size: 16px;">位相空間 $X$ が空でも $X$ でもない開集合 $U,V$ </div><div style="font-size: 16px;">を用いて、$X=U\sqcup V$ と表せないとき、 $X$ は</div><div style="font-size: 16px;"><b>連結</b>といい、$X$ が連結ではないとき</div><div style="font-size: 16px;">$X$ は<b>不連結</b>という。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この定義を次のように言い換えることができます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理10.6</u></div><div style="font-size: 16px;">$X$ が連結であることは、$X$ の開かつ閉集合は $X$ もしくは </div><div style="font-size: 16px;">$\emptyset$ のみであることと同値である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">となる。</div><div style="font-size: 16px;">連結集合について定義しておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義11.1</u></div><div style="font-size: 16px;">$(X,\mathcal{O})$ を位相空間とする。</div><div style="font-size: 16px;">$A\subset X$ が連結集合であるとは、相対位相</div><div style="font-size: 16px;">$(A,\mathcal{O}_A)$ が連結空間であることとして</div><div style="font-size: 16px;">定義する。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">つまり、このことは言い換えれば、</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">任意の開集合 $U,V\in \mathcal{O}$ に対して、</div><div style="font-size: 16px;">$A\subset U\cup V$ かつ</div><div style="font-size: 16px;">$A\cap U\cap V=\emptyset$ </div><div style="font-size: 16px;">なら、$A\subset U$ もしくは $A\subset V$</div><div style="font-size: 16px;">ということになります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">これは、</div><div style="font-size: 16px;">$(A\cap U)\cup(A\cap V)=A$ であることは、$A\subset U\cup V$ と同値</div><div style="font-size: 16px;">$(A\cap U)\cap (A\cap V)=\emptyset$ であることは $A\cap U\cap V=\emptyset$ と同値</div><div style="font-size: 16px;">$A\cap U=A$ であることは $A\subset U$ と同値</div><div style="font-size: 16px;">$A\cap V=A$であることは $A\subset V$ と同値</div><div style="font-size: 16px;">であることからわかります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">最終的に、$A\subset U$ もしくは $A\subset V$ が成り立つのですが、</div><div style="font-size: 16px;">このうちどちらかしか成り立ちません。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">いくつかの例の前に以下を示しておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理11.1</u></div><div style="font-size: 16px;">$({\mathbb R},\mathcal{O}_{d_1})$ は連結である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) ${\mathbb R}$ が連結でないとします。</div><div style="font-size: 16px;">${\mathbb R}=U\sqcup V$ となる空ではない開集合 $U,V$ が</div><div style="font-size: 16px;">存在することになります。</div><div style="font-size: 16px;">$a\in U$ かつ $b\in V$ をとり、</div><div style="font-size: 16px;">$a<b$ としておきます。もしそうでなかったら、$U,V$ の役目を入れ替えれば</div><div style="font-size: 16px;">実現出来ます。</div><div style="font-size: 16px;">$c=\sup\{x\in U|x\le b\}$</div><div style="font-size: 16px;">とおきましょう。</div><div style="font-size: 16px;">このとき、$\epsilon >0$ が存在して、$B_{d_1}(c,2\epsilon )\subset U$ となります。</div><div style="font-size: 16px;">しかし、$c+\epsilon\in U$ であり、$c+\epsilon<b$ であることから、</div><div style="font-size: 16px;">$c=\sup\{x\in U|x\le b\}$ であることに反します。</div><div style="font-size: 16px;">よって、$c\in V$ ということになります。</div><div style="font-size: 16px;">同様に、$\delta>0$ が存在して、$B_{d_1}(c,\delta)\subset V$ となり</div><div style="font-size: 16px;">$(c-\delta,c]\subset V$ が成り立ちます。</div><div style="font-size: 16px;">これは、$c$ が$\{x\in U|x\le b\}$ の上限であることに反します。</div><div style="font-size: 16px;">もし上限であるなら、$\forall \epsilon>0$ に対して、$c-\epsilon <x\le c$</div><div style="font-size: 16px;">となる$x\in U$ が存在するからです。</div><div style="font-size: 16px;">よって $c$ は $U,V$ のどちらも含まれないので ${\mathbb R}=U\cup V$</div><div style="font-size: 16px;">であることに反します。</div><div style="font-size: 16px;">よって、${\mathbb R}$ が連結になるということになります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><span>ここで次の定理を示しておきましょう。</span><br /></div><div style="font-size: 16px;"><span><br /></span></div><div style="font-size: 16px;"><u>定理11.2</u></div><div style="font-size: 16px;">$X,Y$ を位相空間とする。$f:X\to Y$ を全射連続写像とする。</div><div style="font-size: 16px;">$X$ が連結なら $Y$ も連結である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $U\subset Y$ を開かつ閉集合とします。</div><div style="font-size: 16px;">このとき、$f^{-1}(U)$ も開かつ閉集合です。</div><div style="font-size: 16px;">$X$ は連結なので </div><div style="font-size: 16px;">$f^{-1}(U)=\emptyset$ か $f^{-1}(U)=X$ となります。</div><div style="font-size: 16px;">$Y$ が全射であることから、$U=\emptyset$ もしくは $Y$ となります。</div><div style="font-size: 16px;">これは $Y$ が連結であることを意味します。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">このことから、$f$ が全射でなくても、$X$ が連結なら</div><div style="font-size: 16px;">$f(X)$ も連結であることが分かります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、$A\subset X$ が連結集合であるなら、$f(A)$ は連結</div><div style="font-size: 16px;">ということになります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">つまり、</div><div style="font-size: 16px;">連結集合の連続写像による像(<b>連続像</b>といいます)は連結ということになります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">このことから、連結性は位相的性質になることも分かります。</div><div style="font-size: 16px;">なぜなら</div><div style="font-size: 16px;">$f:X\to Y$ が同相であるとします。</div><div style="font-size: 16px;">このとき、$X$ が連結とすると、定理11.2から $Y$ も連結になります。</div><div style="font-size: 16px;">よって、連結性は位相的性質になるからです。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">次の定義をしましょう。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義11.2</u></div><div style="font-size: 16px;">$a\in X$ に対して、$a$ を含む連結集合の内最大のものを</div><div style="font-size: 16px;">$a$ の<b>連結成分</b>といい、$C_X(a)$ と書く。</div><div style="font-size: 16px;">また簡単に $C(a)$ と書くこともある。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$a\sim x$ は同じ連結成分に属するとして $X$ 上に同値関係を定めることができます。</div><div style="font-size: 16px;">そうすると、</div><div style="font-size: 16px;">$X$ のこの同値関係の同値類によって、分解</div><div style="font-size: 16px;">$$X=\sqcup_{\lambda\in \Lambda}C_X(a_\lambda)$$</div><div style="font-size: 16px;">を与えることができます。</div><div style="font-size: 16px;">これを<b>連結成分分解</b>といいます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$X$ が連結であることは、$X=C_X(a)$</div><div style="font-size: 16px;">のように$X$ の任意の元がただ1つの連結成分に属することと同値になります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで次の定理を示しておきます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>命題11.2</u></div><div style="font-size: 16px;">$A\subset X$ が開かつ閉集合であれば、</div><div style="font-size: 16px;">$A$ は $X$ の連結成分のいくつかの和集合となる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $A$ が開かつ閉集合とします。このとき、</div><div style="font-size: 16px;">$\forall a\in A$ に対して、$A\cap C_X(a)\subset C_X(a)$</div><div style="font-size: 16px;">より、$A\cap C_X(a)$ は $C_X(a)$ の開かつ閉集合になります。</div><div style="font-size: 16px;">$C_X(a)$ は連結であり、$a\in A\cap C_X(a)$ は空ではないから</div><div style="font-size: 16px;">$A\cap C_X(a)=C_X(a)$ となります。</div><div style="font-size: 16px;">よって、$C_X(a)\subset A$ ととなり、よって、$\forall a\in A$ に対して、$C_X(a)\subset A$ であるから</div><div style="font-size: 16px;">$\cup_{a\in A}C_X(a)\subset A$ また、$A\subset \cup_{a\in A}C_X(a)$</div><div style="font-size: 16px;">であるから</div><div style="font-size: 16px;">$$A=\cup_{a\in A}C_X(a)$$</div><div style="font-size: 16px;">となり、$A$ はいくつかの連結成分の和集合となります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この証明の途中で用いたことを復習しておきます。</div><div style="font-size: 16px;"><br /></div><div><span style="font-size: medium;"><u>相対位相の閉集合について</u></span></div><div style="font-size: 16px;">$A\subset X$ での相対位相において、</div><div style="font-size: 16px;">$A$ における開集合は $X$ の開集合 $U$ を用いて $A\cap U$ と書き表されます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $A$ における閉集合 $F$ は、$A-F=A\cap F^c$ より $A$ における開集合だから、</div><div style="font-size: 16px;">$A\cap F^c=A\cap U$ となる $X$ の開集合 $U$ が存在します。</div><div style="font-size: 16px;">よって、この補集合を取ると、</div><div style="font-size: 16px;">$A^c\cup F=A^c\cup U^c$</div><div style="font-size: 16px;">$A^c\cap F=\emptyset$ であるから、$A^c$ の部分を取ると、</div><div style="font-size: 16px;">$F=(A^c\cup U^c)\cap A=(A^c\cap A)\cup (U^c\cap A)=A\cap U^c$</div><div style="font-size: 16px;">よって、$A$ 上の閉集合は、$X$ のある閉集合 $G$ を用いて、</div><div style="font-size: 16px;">$A\cap G$ とかけることがわかります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">次の定理を示しましょう。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理11.4</u></div><div style="font-size: 16px;">連結集合 $A$ に対して、$A\subset B\subset \overline{A}$ </div><div style="font-size: 16px;">を満たす任意の集合 $B$ は連結である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $B\subset U\cup V$ を満たす $X$ の開集合</div><div style="font-size: 16px;">が $B\cap U\cap V=\emptyset$ を満たすとします。</div><div style="font-size: 16px;">このとき、</div><div style="font-size: 16px;">$B\cap U\neq \emptyset$ を満たすと仮定します。</div><div style="font-size: 16px;">このとき、 $B\subset \overline{A}$ であるから</div><div style="font-size: 16px;">$\overline{A}\cap U\neq \emptyset$ です。</div><div style="font-size: 16px;">$x\in\overline{A}\cap U$ をとります。</div><div style="font-size: 16px;">このとき、$U$ は開集合だから $U\in \mathcal{N}(x)$ となります。</div><div style="font-size: 16px;">よって、$A\cap U\neq \emptyset$ を満たします。</div><div style="font-size: 16px;">$A$ は連結だから、</div><div style="font-size: 16px;">$A\subset U\cup V$ かつ $A\cap U\cap V=\emptyset$ より</div><div style="font-size: 16px;">$A\subset U$ もしくは $A\subset V$ です。</div><div style="font-size: 16px;">しかし、$A\cap U\neq \emptyset$ であるから $A\subset U$ です。</div><div style="font-size: 16px;">よって、$A\cap V=\emptyset$ であるから $\overline{A}\cap V=\emptyset$</div><div style="font-size: 16px;">である。つまり、$B\cap V=\emptyset$ です。</div><div style="font-size: 16px;">このことから、$B\subset U$ が分かります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、$B\cap V\neq \emptyset$ である場合からも、</div><div style="font-size: 16px;">同じように $B\subset V$ が証明することができます。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この定理から次が成り立ちます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理11.5</u></div><div style="font-size: 16px;">任意の $a\in X$ に対して、連結成分 $C(a)$ は閉集合である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $C(a)\subset \overline{C(a)}$ が成り立つ。</div><div style="font-size: 16px;">また定理11.4から $\overline{C(a)}$ は連結になります。</div><div style="font-size: 16px;">連結成分の最大性により、$\overline{C(a)}\subset C(a)$ </div><div style="font-size: 16px;">が成り立ちます。この包含関係から、</div><div style="font-size: 16px;">$C(a)=\overline{C(a)}$ が成り立つ、つまり $C(a)$ が閉集合になります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">例</div><div style="font-size: 16px;">$(a,b)\subset {\mathbb R}$ は ${\mathbb R}$ と同相であるから、</div><div style="font-size: 16px;">連結は位相的性質なので、$(a,b)$ も連結性になります。</div><div style="font-size: 16px;">この閉包 $[a,b]$ も同相になります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>命題11.3</u></div><div style="font-size: 16px;">$\forall r\in {\mathbb Q}$ に対して、$r$ の連結成分 $C(r)$ について</div><div style="font-size: 16px;">$C(r)=\{r\}$ が成り立ちます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明) $\forall r,s\in {\mathbb Q}$ に対して、</div><div style="font-size: 16px;">$r<s$ に対して、$r<q<s$ となる無理数 $q$ が存在して、</div><div style="font-size: 16px;">${\mathbb Q}\cap (-\infty,q)={\mathbb Q}\cap (-\infty,q]$</div><div style="font-size: 16px;">は開かつ閉集合であるから、${\mathbb Q}\cap (-\infty,q]$ は</div><div style="font-size: 16px;">連結成分の和集合つまり、$C(r)\subset {\mathbb Q}\cap (-\infty,q]$</div><div style="font-size: 16px;">であり、</div><div style="font-size: 16px;">同様に、$C(s)\subset {\mathbb Q}\cap [q,\infty)$</div><div style="font-size: 16px;">であるから、$C(r)\cap C(s)=\emptyset$ であるから、</div><div style="font-size: 16px;">$\forall r\in{\mathbb Q}$ の連結成分は $r$ のみとなります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">このように、$\forall a\in X$ に対して、$C_X(a)=\{a\}$</div><div style="font-size: 16px;">であるとき、$X$ は<b>完全不連結</b>であるという。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">よって、$({\mathbb Q},\mathcal{O}_{d_1})$ は完全不連結であり、</div><div style="font-size: 16px;">ゾルゲンフライ直線 $({\mathbb R},\mathcal{O}_{l})$ も完全不連結となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、連結性の最後に中間値の定理を示しました。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理11.6</u></div><div style="font-size: 16px;">位相空間 $X$ と連続関数 $f:X\to {\mathbb R}$ をとる。</div><div style="font-size: 16px;">$X$ の連結な部分集合 $A$ に対して、</div><div style="font-size: 16px;">$a,b\in A$ が $f(a)<f(b)$ を満たすとする。</div><div style="font-size: 16px;">このとき、$f(a)<\forall <span> c </span><f(b)$ となる $c$ に対して、</div><div style="font-size: 16px;">$f(x)=c$ となる $x\in A$ が存在する。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明)$A$ は連結なので、$f(A)\subset {\mathbb R}$ も連結であり、</div><div style="font-size: 16px;">$f(a),f(b)\in f(A)$ が成り立ちます。</div><div style="font-size: 16px;">よって、このとき、$f(a)<c<f(b)$ が $f(x)=c$ となる $x$ が存在しないとする。</div><div style="font-size: 16px;">このとき、$B=f(A)\cap (-\infty, c)=f(A)\cap (-\infty ,c]$ より、</div><div style="font-size: 16px;">$B$ は空でも全体でもない $f(A)$ の開かつ閉集合であるから矛盾する。</div><div style="font-size: 16px;">よって、$f(x)=c$ となる $x\in A$ が存在する。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><br /></div><div><span style="font-size: large;"><b><u>弧状連結性</u></b></span></div><div style="font-size: 16px;">弧状連結の定理をしましょう。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定義11.4</u></div><div style="font-size: 16px;">位相空間 $X$ が<b>弧状連結</b>であるとは以下を満たすことを意味します。</div><div style="font-size: 16px;">$\forall a,b\in X$ に対して、ある連続写像 $f:I\to X$ が</div><div style="font-size: 16px;">存在して、$f(0)=a$ かつ $f(1)=b$ を満たす。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで、$I$ は閉区間 $[0,1]$ です。</div><div style="font-size: 16px;">すぐにわかるのは次の定理です。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理11.7</u></div><div style="font-size: 16px;">弧状連結なら連結である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明)$X$ が弧状連結であるとします。</div><div style="font-size: 16px;">このとき、$\forall a,b \in X$ に対して、連続写像 $f:I\to X$ が存在して</div><div style="font-size: 16px;">$f(0)=a$ かつ $f(1)=b$ を満たし、つまり、$a,b\in f(I)\subset X$ を満たします。</div><div style="font-size: 16px;">$I$ が連結であるから、$f(I)$ も連結になり、$f(I)\subset C(b)$ であるから</div><div style="font-size: 16px;">$a$ は $b$ の連結成分に属します。</div><div style="font-size: 16px;">つまり、$a\in C(b)$ であり、つまり $X\subset C(b)$ より</div><div style="font-size: 16px;">$X=C(b)$ となります。つまり $X$ は連結となります。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">この定理の逆は一般には成り立ちません。</div><div style="font-size: 16px;">つまり、連結だが、弧状連結であるものが存在します。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><u>定理11.8</u></div><div style="font-size: 16px;">$J=\left\{\left(x,\sin\frac{1}{x}\right)|x\in {\mathbb R}_{>0}\right\}$</div><div style="font-size: 16px;">とし、</div><div style="font-size: 16px;">$I=\{(0,t)|-1\le t\le 1\}$</div><div style="font-size: 16px;">とする。このとき $X=I\cup J$ は連結であるが、弧状連結ではない。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">$J$ は下の図のようなグラフになっており、</div><div style="font-size: 16px;">$I$ はこの $y$ -軸の $[-1,1]$ の区間です。</div><div style="font-size: 16px;"><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEhP9bSsxSRkSInQA4-5GtlgyUz-XUn_2NMfktv8GmpeSIiGu9bGde9_MGkzib51rut171g6h0fq8akw8ePqAyuTU0fingK-jRlL-Spb9_VahWp-4Hh8BKkpxfSefKQeYp190CymInpP7n0Dg94bHpBSsP3YwgfFBRlZqeAhO9DvF04xnHd-zpBfIRwKlQ=s1229" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="752" data-original-width="1229" height="196" src="https://blogger.googleusercontent.com/img/a/AVvXsEhP9bSsxSRkSInQA4-5GtlgyUz-XUn_2NMfktv8GmpeSIiGu9bGde9_MGkzib51rut171g6h0fq8akw8ePqAyuTU0fingK-jRlL-Spb9_VahWp-4Hh8BKkpxfSefKQeYp190CymInpP7n0Dg94bHpBSsP3YwgfFBRlZqeAhO9DvF04xnHd-zpBfIRwKlQ=s320" width="320" /></a></div><br /><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明)連結性について</div><div style="font-size: 16px;">$J$ は ${\mathbb R}_{>0}$ 上の連続関数 $\sin \frac{1}{x}$ のグラフであり</div><div style="font-size: 16px;">連続像になります。よって、定理11.2から $J$ は連結。</div><div style="font-size: 16px;">また、$\forall (0,t)\in I$ に対して、数列 $\left(\frac{1}{2n\pi+\text{arcsin}t},t\right)$</div><div style="font-size: 16px;">は、$J$ 上の点列であり、$n\to \infty$ において $(0,t)$ に収束するので、</div><div style="font-size: 16px;">$(0,t)\in \overline{J}$ となります。</div><div style="font-size: 16px;">これは、$I\cup J\subset \overline{J}$ であることを示しており、</div><div style="font-size: 16px;">$$I\subset I\cup J\subset \overline{J}$$</div><div style="font-size: 16px;">から、定理11.4 から $X=I\cup J$ も連結ということになります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">非弧状連結について</div><div style="font-size: 16px;">$O\in X$ を ${\mathbb R}^2$ の原点とし $A=(\frac{1}{\pi},0)$ がある連続写像</div><div style="font-size: 16px;">$f:I\to X$ によって $f(0)=O$ かつ $f(1)=A$ とならないことを示します。</div><div style="font-size: 16px;">もし結べるとすると、連続写像 $f:[0,1]\to X$ が存在して、</div><div style="font-size: 16px;">$f(0)=O$かつ $f(1)=A$ を満たします。</div><div style="font-size: 16px;">そうすると、今、$f^{-1}(O)=0$ と仮定しておきます。</div><div style="font-size: 16px;">このとき、</div><div style="font-size: 16px;">$B_{d_2}(O,\frac{1}{2})$をとり、</div><div style="font-size: 16px;">$0\in f^{-1}(B_{d_2}(O,\frac{1}{2}))$ の近傍$[0,\delta)$ が存在して、</div><div style="font-size: 16px;">$[0,\delta)\subset f^{-1}(B_{d_2}(O,\frac{1}{2}))$を満たします。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで、$B$ は下の円の内側になります。</div><div style="font-size: 16px;"><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEgVHZeo6gmWW6TYQ1BHUP1FFbya_kEg9oIwqFn2aOE36rXDGsC6VLVrkmeF4s55DRr8f3nJH99R2lxKRwvxPc-pjMv7UR_teIoPBDuVg1W8EsW4O5SmIBJ0JIzaC49_BT780gvf6qHL4znTN0Zg69zDy5BpYKVpGsnqP-pw-jQaanN8F6ayuKjb22CfgQ=s1475" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="838" data-original-width="1475" height="182" src="https://blogger.googleusercontent.com/img/a/AVvXsEgVHZeo6gmWW6TYQ1BHUP1FFbya_kEg9oIwqFn2aOE36rXDGsC6VLVrkmeF4s55DRr8f3nJH99R2lxKRwvxPc-pjMv7UR_teIoPBDuVg1W8EsW4O5SmIBJ0JIzaC49_BT780gvf6qHL4znTN0Zg69zDy5BpYKVpGsnqP-pw-jQaanN8F6ayuKjb22CfgQ=s320" width="320" /></a></div><br /><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">そうすると $f^{-1}(O)=0$ と仮定したことから $\text{pr}_1(f([0,\delta)))=[0,a)$ となる</div><div style="font-size: 16px;">$a>0$ が存在します。</div><div style="font-size: 16px;">よって、$\frac{2}{(4n+1)\pi}<a$ を満たす$n$ が存在し、</div><div style="font-size: 16px;">$(\frac{2}{(4n+1)\pi},1)\in f([0,\delta))\subset B$ を満たします。</div><div style="font-size: 16px;">しかし、$(\frac{2}{(4n+1)\pi},1)\not\in B$ であるので、</div><div style="font-size: 16px;">矛盾します。</div><div style="font-size: 16px;">よって、$X$ は弧状連結ではないということになります。$\Box$ </div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">途中で、$f^{-1}(O)=0$ と仮定しましたが、</div><div style="font-size: 16px;">これは簡単に証明できます。$\{t\in I|f(t)=O\}$ となるものの</div><div style="font-size: 16px;">$\sup$ を $t_0$ を取り、$[t_0,1]$ を $[0,1]$ に線形に引き伸ばしたものを</div><div style="font-size: 16px;">$f$ に合成したものを再び $f$ とおけば良いです。</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-39336076331939650672021-12-24T02:08:00.002+09:002022-04-19T10:35:28.220+09:00トポロジー入門(第10回) <div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所オンライン(月曜日3限)]</span></div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px; margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/21/top.html">トポロジー入門のHP</a></div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="font-size: 16px; margin: 0px;">今回は商位相空間と連結性について解説しました。</div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="margin: 0px;"><b><u><span style="font-size: large;">商位相</span></u></b></div><div style="margin: 0px;"><b><u><span style="font-size: large;"><br /></span></u></b></div><div style="margin: 0px;"><span>まず、次の定義をします。</span></div><div style="margin: 0px;"><span><br /></span></div><div style="margin: 0px;">定義10.1</div><div style="margin: 0px;">$(X,\mathcal{O}_X)$、$(Y,\mathcal{O}_Y)$ を位相空間とし、</div><div style="margin: 0px;">$f:X\to Y$ を全射とする。</div><div style="margin: 0px;">$$f^{-1}(U)\in \mathcal{O}_X\Leftrightarrow U\in \mathcal{O}_X$$</div><div style="margin: 0px;">が成り立つとき、$f$ は<b>商写像</b>という。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">すぐわかることは、商写像なら連続ということです。</div><div style="margin: 0px;">上の⇐が $f$ の連続性を意味するからです。</div><div style="margin: 0px;">しかし、商写像は全射連続だけではありません。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">その条件をまずは見ていきます。</div><div style="margin: 0px;">商写像の例として、連続で開な全射があります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><u>定理10.1</u></div><div style="margin: 0px;">$f:(X,\mathcal{O}_X)\to (Y,\mathcal{O}_Y)$ が連続な全射な開写像なら、$f$ は商写像。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">(証明) 連続性は仮定されているから、写像であることの$\Rightarrow$ </div><div style="margin: 0px;">を示せばよいです。</div><div style="margin: 0px;">$U\in \mathcal{P}(Y)$ に対して、$f^{-1}(U)\in \mathcal{O}_X$</div><div style="margin: 0px;">なら、$f$ が開写像であることから $f(f^{-1}(U))=U\in\mathcal{O}_Y$</div><div style="margin: 0px;">分かります。</div><div style="margin: 0px;">よって、写像の条件の$\Rightarrow$ が分かります。</div><div style="margin: 0px;">このことから $f$ は商写像であることがわかります。<span> $\Box$</span></div><div style="margin: 0px;"><span><br /></span></div><div style="margin: 0px;"><span>商写像の特徴付けをここで与えます。</span></div><div style="margin: 0px;"><span><br /></span></div><div style="margin: 0px;"><span><u>定理10.3</u></span></div><div style="margin: 0px;">全射 $f:(X,\mathcal{O}_X)\to (Y,\mathcal{O}_Y)$</div><div style="margin: 0px;">が商写像であることとは、</div><div style="margin: 0px;">$\mathcal{O}_Y$ が $f$ を連続にする最大の位相であることの</div><div style="margin: 0px;">必要十分条件である。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">$\mathcal{O}_Y$ が $f$ を連続にする最大の位相であるとは、</div><div style="margin: 0px;">$\mathcal{O}$ が $f:(X,\mathcal{O}_X)\to (Y,\mathcal{O})$ を連続とする</div><div style="margin: 0px;">任意の位相なら $\mathcal{O}\subset\mathcal{O}_Y$ であることを意味します。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">(定理10.3の証明) (十分性) $f$ が商写像であるとします。</div><div style="margin: 0px;">このとき、$\mathcal{O}$ を $f:(X,\mathcal{O}_X)\to (Y,\mathcal{O})$ を連続にする任意の位相なら</div><div style="margin: 0px;">$\forall U\in \mathcal{O}$ に対して、$f^{-1}(U)\in \mathcal{O}_X$</div><div style="margin: 0px;">であることから、商写像の定義から $U\in\mathcal{O}_Y$ が分かります。</div><div style="margin: 0px;">ゆえに、$\mathcal{O}\subset\mathcal{O}_Y$ となります。このことから</div><div style="margin: 0px;">$\mathcal{O}_Y$ は $f$ を連続にする最大の位相であることが分かります。</div><div style="margin: 0px;">(必要性) $\mathcal{O}_Y$ が $f$ を連続にする最大の位相とする。</div><div style="margin: 0px;">まず $f$ が連続であることから、商写像の条件の$\Leftarrow$ が成り立つ。</div><div style="margin: 0px;">商写像の条件の $\Rightarrow$ を示す。</div><div style="margin: 0px;">任意の $f^{-1}(U)\in \mathcal{O}_X$ に対して、</div><div style="margin: 0px;">$\mathcal{O}=\{\emptyset,U,Y\}$ は $Y$ 上の位相であり、$f:(X,\mathcal{O}_X)\to (Y,\mathcal{O})$ は連続である。</div><div style="margin: 0px;">ここで、$\mathcal{O}_Y$ は $f$ を連続にする最大の位相であったから $\mathcal{O}\subset \mathcal{O}_Y$ が分かる。<br /></div><div style="margin: 0px;">特に、$U\in\mathcal{O}_Y$ である。</div><div style="margin: 0px;">よって、これは、$f:(X,\mathcal{O}_X)\to (Y,\mathcal{O}_Y)$ が商写像</div><div style="margin: 0px;">であることを意味します。$\Box$</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><span>ここで、 次の定義をしましょう。</span></div><div style="margin: 0px;"><span>位相空間 $(X,\mathcal{O})$ と全射 $f:X\to Y$ が存在したとき、</span></div><div style="margin: 0px;"><span>$Y$ 上の $f$ の最小にする位相を $\mathcal{O}_f$ と書くことにします。</span></div><div style="margin: 0px;"><span>このとき、直前の定理から $\mathcal{O}_f$ は </span></div><div style="margin: 0px;"><span>$(X,\mathcal{O})\to (Y,\mathcal{O}_f)$ が</span></div><div style="margin: 0px;">商写像となるような $Y$ 上の位相ということになります。</div><div style="margin: 0px;">そのような位相の一意性もわかることになります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">このとき、$\mathcal{O}_f$は、</div><div style="margin: 0px;">$$\{U\subset Y|f^{-1}(U)\in\mathcal{O}_X\}$$</div><div style="margin: 0px;">としても定義されます。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">このようにして、全射 $f:X\to Y$ を通して、$X$ 上の位相から</div><div style="margin: 0px;">$f$ の連続性を通して $Y$ 上の位相を"標準的に"作る唯一の方法があることになります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">この方法を通して、商集合 $X/\sim$ を位相空間にすることができます。</div><div style="margin: 0px;"><b>商集合</b>とは、集合 $X$ 上に導入された同値関係$\sim$ </div><div style="margin: 0px;">によって作られる同値類全体の集合のことです。</div><div style="margin: 0px;">$\sim$ が<b>同値関係</b>であるとは、$\forall a,b,c\in X$ に対して、</div><div style="margin: 0px;">(i) $a\sim a$</div><div style="margin: 0px;">(ii) $a\sim b\rightarrow b\sim a$</div><div style="margin: 0px;">(iii) $a\sim b$ かつ $b\sim c$ ならば $a\sim c$</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">を満たす2項関係のことです。</div><div style="margin: 0px;">このとき、$C(a)=\{x\in X|a\sim x\}$ として<b>同値類</b>を表します。</div><div style="margin: 0px;">この同値類全体の集合 $\{C(a)|a\in X\}$ を<b>商集合</b>といい、</div><div style="margin: 0px;">$X/\sim$ と書くのでした。</div><div style="margin: 0px;">写像 $p:X\to X/\sim$ を</div><div style="margin: 0px;">$a\mapsto C(a)$ として定めたとき、</div><div style="margin: 0px;">$p$ を<b>自然な射影</b>といいます。</div><div style="margin: 0px;">記号の使い方として、$C(a)$ のことを $[a]$ とかぎかっこを</div><div style="margin: 0px;">使って書くことがあります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">今、$X$ 上に位相 $\mathcal{O}$ が入っていたとき、</div><div style="margin: 0px;">自然な射影 $p:X\to X/\sim$ を通して、$X/\sim$ に</div><div style="margin: 0px;">位相を入れることができます。</div><div style="margin: 0px;">この時できる $(X/\sim,\mathcal{O}_p)$ を</div><div style="margin: 0px;">$X/\sim$ 上の<b>商位相空間(</b>$\mathcal{O}_p$ を<b>商位相</b><b>)</b>といいます。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">つまり、$p$ が商写像となるような位相空間を $X/\sim$ </div><div style="margin: 0px;">上に導入したことになります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><u>例10.4</u></div><div style="margin: 0px;">$({\mathbb R}^2,\mathcal{O}_{d_2})$ </div><div style="margin: 0px;">に、同値関係として、$(x_1,y_1)\sim(x_2,y_2)\leftrightarrow x_1=x_2$</div><div style="margin: 0px;">として導入する。自然な射影を $p$ とします。</div><div style="margin: 0px;">$(x,y)$ を含む同値類を $[x,y]$ と表すことにします。</div><div style="margin: 0px;">このとき、構成される商位相空間 $({\mathbb R}^2/\sim,\mathcal{O}_{d_2,p})$</div><div style="margin: 0px;">は $({\mathbb R},\mathcal{O}_{d_1})$ と同相になります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">実際、$\varphi([x,y])=f(x)$ として定義すると</div><div style="margin: 0px;">$\varphi:{\mathbb R}^2/\sim\to {\mathbb R}$ は同相になります。</div><div style="margin: 0px;">まず写像になることは、$[x,y]=[x',y']$ なら、$\varphi([x,y])=x=x'=\varphi([x',y'])$</div><div style="margin: 0px;">であるからです。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">単射性は、$\varphi([x,y])=\varphi([x',y'])$ なら、$x=x'$ が成り立ち、</div><div style="margin: 0px;">同値関係の定義から $(x,y)\sim (x',y')$ が成り立つので、$[x,y]=[x',y']$</div><div style="margin: 0px;">が成り立ちます。</div><div style="margin: 0px;">全射性は、$\forall x\in {\mathbb R}$ に対して $z\in {\mathbb R}$ を</div><div style="margin: 0px;">任意に選ぶと、$\varphi([x,z])=x$ が得られることからわかります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">同相であることは次の定理の証明に一般化して証明します。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><u>定理10.4</u></div><div style="margin: 0px;">$f:(X,\mathcal{O}_X)\to (Y,\mathcal{O}_Y)$ を商写像とする。</div><div style="margin: 0px;">このとき、$f(x)=f(x')\leftrightarrow x\sim x'$ として </div><div style="margin: 0px;">$X$ に同値関係 $\sim$ を導入する。その自然な射影を $p$ とする。</div><div style="margin: 0px;">このとき、商位相空間 $(X/\sim,\mathcal{O}_p)$ は </div><div style="margin: 0px;">$(Y,\mathcal{O}_Y)$ と同相である。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">(証明) $\varphi:X/\sim\to Y$ を $\varphi([x])=f(x)$ として定義すると</div><div style="margin: 0px;">$\varphi$ が写像であることや、全単射であることは上記の例10.4と</div><div style="margin: 0px;">同様の証明をすることでわかります。</div><div style="margin: 0px;">また、$\varphi([x])=\varphi(p(x))=\varphi\circ p(x)=f(x)$</div><div style="margin: 0px;">であることから、$\varphi\circ p=f$ であることもわかります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">$\varphi$ が同相写像であることを示そう。</div><div style="margin: 0px;">(連続性) $U\in \mathcal{O}_Y$ とすると、$p^{-1}(\varphi^{-1}(U))=(\varphi\circ p)^{-1}(U)=f^{-1}(U)$</div><div style="margin: 0px;">より、$f$ の連続性から $p^{-1}(\varphi^{-1}(U))\in \mathcal{O}_X$ であることが</div><div style="margin: 0px;">わかり、$p$ が商写像であることから$\varphi^{-1}(U)\in \mathcal{O}_{X,p}$</div><div style="margin: 0px;">であることが分かります。これは、$\varphi$ が連続であることを意味します。</div><div style="margin: 0px;">(開写像性) $V\in \mathcal{O}_{X,p}$ に対して、</div><div style="margin: 0px;">$p^{-1}(V)=p^{-1}(\varphi^{-1}(\varphi(V)))=(\varphi\circ p)^{-1}(\varphi(V))=f^{-1}(\varphi(V))$</div><div style="margin: 0px;">であり、$p$ が連続であることと $f$ が商写像であることから、$\varphi(V)\in \mathcal{O}_{Y}$</div><div style="margin: 0px;">であることがわかります。</div><div style="margin: 0px;">これは$\varphi$ が開写像であることを意味します。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">全単射な連続な開写像は同相写像であるから、$f$ は同相写像となります。$\Box$</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">これらのことから、例10.4の写像</div><div style="margin: 0px;">$({\mathbb R}^2/\sim,\mathcal{O}_{d_2,p})\to ({\mathbb R},\mathcal{O}_{d_1})$</div><div style="margin: 0px;">は同相写像であることがわかります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">ここで次の定理を用意します。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><u>定理10.5</u></div><div style="margin: 0px;">$f:X\to Y$ を連続写像とする。</div><div style="margin: 0px;">このとき、$\tilde{f}:X\to f(X)$ を $\tilde{f}(x)=f(x)$ </div><div style="margin: 0px;">とするとき、$\tilde{f}$ も連続である。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">この定理は、連続写像の終域を像に縮めた写像も連続であることを</div><div style="margin: 0px;">主張しています。ここで、$f(X)$ の位相は $Y$ からくる相対位相と</div><div style="margin: 0px;">考えます。証明は易しいのでここでは省略します。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><u>例10.5</u></div><div style="margin: 0px;">${\mathbb S}^1=\{(x,y)\in {\mathbb R}^2|x^2+y^2=1\}$ とおきます。</div><div style="margin: 0px;">$f:{\mathbb R}\to {\mathbb S}^1$ を$f(\theta)=(\cos\theta,\sin\theta)$ とします。</div><div style="margin: 0px;">${\mathbb S}^1\subset{\mathbb R}^2$ の相対位相は、</div><div style="margin: 0px;">${\mathbb R}$ 上の同値関係 $\theta\sim \theta'\leftrightarrow \theta_1-\theta_2\in 2\pi{\mathbb Z}$</div><div style="margin: 0px;">によって得られる商位相空間と同相となります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">(証明)</div><div style="margin: 0px;">$\tilde{f}:{\mathbb R}\to {\mathbb R}^2$ を$\tilde{f}(\theta)=f(\theta)$</div><div style="margin: 0px;">として得られるものとします。</div><div style="margin: 0px;">$\tilde{f}$ は連続となります。$\mathcal{O}_{d_2}$ は $\mathcal{O}_{d_1}$</div><div style="margin: 0px;">の2つの直積位相になるので、</div><div style="margin: 0px;">$\tilde{f}:{\mathbb R}\to {\mathbb R}^2$ が連続であるためには、</div><div style="margin: 0px;">$\text{pr}_1\circ \tilde{f}(\theta)=\cos\theta$</div><div style="margin: 0px;"><div style="margin: 0px;">$\text{pr}_2\circ \tilde{f}(\theta)=\sin\theta$</div><div style="margin: 0px;">が連続であることが必要十分ですが、$\cos\theta$ や $\sin\theta$ は</div><div style="margin: 0px;">連続関数なので $\tilde{f}$ は連続写像であることが分かります。</div><div style="margin: 0px;">上の定理10.5から $f$ は連続となることがわかります。</div><div style="margin: 0px;">$U\in \mathcal{O}_{d_1}$ とすると、$U=\cup_{\lambda\in \Lambda}(a_\lambda,b_\lambda)$</div><div style="margin: 0px;">のように開区間の和集合になります。</div><div style="margin: 0px;">$f(U)=\cup_{\lambda\in \Lambda}f((a_\lambda,b_\lambda))=\cup_{\lambda\in \Lambda}(f(a_{\lambda}),f(b_{\lambda}))$</div><div style="margin: 0px;">$\{(\cos\theta,\sin\theta)|a<\theta<b\}$ が ${\mathbb S}^1$</div><div style="margin: 0px;">の相対位相の開基となるので、$f(U)\in \mathcal{O}_{d_2,{\mathbb S}^1}$ となります。</div><div style="margin: 0px;">よって、$f$ は全射連続開写像であるから$f$ は商写像になります。</div><div style="margin: 0px;">よって、$f$ によって得られる同値関係は、</div><div style="margin: 0px;">$f(\theta)=f(\theta')\leftrightarrow (\cos\theta,\sin\theta)=(\cos\theta',\sin\theta')\leftrightarrow \theta_1-\theta_2\in 2\pi{\mathbb Z}$</div><div style="margin: 0px;">が成り立ちます。</div><div style="margin: 0px;">ゆえに、${\mathbb R}$ に $\theta_1\sim\theta_2\leftrightarrow \theta_1-\theta_2\in 2\pi{\mathbb Z}$ であるように</div><div style="margin: 0px;">いれた同値関係による商位相空間 $({\mathbb R}/\sim,\mathcal{O}_p)$</div><div style="margin: 0px;">は相対位相空間 $({\mathbb S}^1,\mathcal{O}_{d_2,{\mathbb S}^1})$ </div><div style="margin: 0px;">に$\varphi$ を介して同相になります。</div><div style="margin: 0px;">つまり、</div><div style="margin: 0px;">$$\varphi:({\mathbb R}/\sim,\mathcal{O}_p)\to ({\mathbb S}^1,\mathcal{O}_{d_2,{\mathbb S}^1})$$</div><div style="margin: 0px;">は同相になります。$\Box$</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><b><u><span style="font-size: large;">連結性</span></u></b></div><div style="margin: 0px;">次に連結性に入りましたが、定義のみで終わりました。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><u>定義10.3</u></div><div style="margin: 0px;">位相空間 $(X,\mathcal{O})$ が空でも $X$ でもない</div><div style="margin: 0px;">開集合 $U,V$ を用いて $X= U\sqcup V$ と表せないとき</div><div style="margin: 0px;">$X$ は<b>連結</b>という。</div><div style="margin: 0px;">$X$ が連結ではないとき、$X$ は<b>非連結</b>という。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">ここで、$\sqcup$ は交わりのない和集合を意味します。</div><div style="margin: 0px;">連結性は、「表せないとき」ということを条件としており、</div><div style="margin: 0px;">理解がしずらいです。</div><div style="margin: 0px;">ですので少し次のように言い換えましょう。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;"><u>定理10.6</u></div><div style="margin: 0px;">$X$ が連結であるとは、$X$ の開かつ閉集合は $X$ もしくは $\emptyset$ </div><div style="margin: 0px;">のみであることと同値である。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">(証明) $X$ が不連結であるとすると、空でも、$X$ でもない 開集合</div><div style="margin: 0px;">$U,V\subset X$が存在して</div><div style="margin: 0px;">$X=U\cup V$ が成り立ちます。このとき、$U^c=V$ は開集合であるから</div><div style="margin: 0px;">$U$ は開かつ閉集合となります。</div><div style="margin: 0px;">よって、空でも $X$ でもない開かつ閉集合が存在します。</div><div style="margin: 0px;">空集合と $X$ は開かつ閉集合であるから、</div><div style="margin: 0px;">この否定は、空もしくは $X$ ではない開かつ閉集合が存在しないとき連結性を</div><div style="margin: 0px;">意味することになります。$\Box$</div></div></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-22813265952631636432021-12-14T01:06:00.002+09:002022-04-19T10:35:39.539+09:00トポロジー入門(第9回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所:</span><span style="color: orange; font-family: inherit;">オンライン(月曜日3限)]</span></div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px; margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/21/topclass.html">トポロジー入門のHP</a></div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="font-size: 16px; margin: 0px;">今回は、点列の収束性と、直積位相についてやりました。</div><div style="font-size: 16px; margin: 0px;">その前に、誘導位相について残っていた部分を片付けました</div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="margin: 0px;"><span style="font-size: large;"><b><u>誘導位相・相対位相</u></b></span></div><div style="font-size: 16px; margin: 0px;">定理9.1</div><div style="font-size: 16px; margin: 0px;">位相空間$(X,\mathcal{O})$ の開基を</div><div style="font-size: 16px; margin: 0px;">$\mathcal{B}$ とするとき、$A\subset X$ の相対位相<br />$(A,\mathcal{O}_A)$ の開基は、<br />$$\mathcal{B}_A=\{A\cap B|B\in \mathcal{B}\}$$<br />である。</div><div style="font-size: 16px; margin: 0px;"><br />(証明) $\mathcal{O}_A$ の開集合 $A\cap U$ をとります。<br />$p\in A\cap U$ に対して、$p\in B\subset U$<br />となる $B\in \mathcal{B}$ が存在します。<br />よって<br />$$p\in A\cap B\subset A\cap U$$<br />となり、$A\cap B\in \mathcal{B}_A$ となるので<br />$\mathcal{B}_A$ は $\mathcal{O}_A$ の開基となります。$\Box$</div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="font-size: 16px; margin: 0px;">先週は誘導位相ついてやったのですが その時は写像が<span lang="EN-US">1</span>つだけの場合にやりました。<br />しかし今回は複数の写像について誘導位相を定義しました。<br /></div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="font-size: 16px; margin: 0px;">定義9.1(誘導される位相2)</div><div style="font-size: 16px; margin: 0px;">$(Y_\lambda,\mathcal{O}_\lambda)$ 位相空間の族</div><div style="font-size: 16px; margin: 0px;">写像 $f_\lambda:X\to Y_\lambda$ ($\lambda\in \Lambda)$</div><div style="font-size: 16px; margin: 0px;">$\mathcal{F}=\{f_\lambda|\lambda\in \Lambda\}$ とする。</div><div style="font-size: 16px; margin: 0px;">$\langle \mathcal{F}\rangle$ を $\{f_\lambda^{-1}(U_\lambda)|U_\lambda\in \mathcal{O}_\lambda,\lambda\in \Lambda\}$</div><div style="font-size: 16px; margin: 0px;">を準開基とする開集合系。つまり、</div><div style="font-size: 16px; margin: 0px;">$$\langle \mathcal{F}\rangle:=\langle \{f_\lambda^{-1}(U_\lambda)|U_\lambda\in\mathcal{O}_\lambda,\lambda\in \Lambda\}\rangle$$</div><div style="font-size: 16px; margin: 0px;">とする。これを $\mathcal{F}$ によって<b>誘導される位相</b>(<b>誘導位相</b>)</div><div style="font-size: 16px; margin: 0px;">という。</div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="font-size: 16px; margin: 0px;">このとき、$\langle \mathcal{F}\rangle$ は、$\mathcal{F}$ </div><div style="font-size: 16px; margin: 0px;">の任意の写像を連続にするような $X$ 上の最小の位相ということになります。</div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="margin: 0px;"><u><b><span style="font-size: large;">点列の収束</span></b></u></div><div style="margin: 0px;"><u><span style="font-size: medium;"><br /></span></u></div><div style="font-size: 16px; margin: 0px;">について説明をしました。</div><div style="font-size: 16px; margin: 0px;"><br /></div><div style="margin: 0px; text-align: left;"><div style="font-size: 16px; margin: 0px; text-align: left;">そもそも点列というものは、$x_1,x_2,\cdots $ であって、それらは、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$(x_n)\in X^{\mathbb N}$ と考えられます。<br /><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">点列が収束することについての定義をしておきます。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;"><b>定義9.3<br /></b>点列が収束するというのは、$x\in X$ が存在して、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$$\forall U\in \mathcal{N}(x)\exists N>0\forall n>N(x_n\in U)$$</div><div style="font-size: 16px; margin: 0px; text-align: left;">を満たすことをいう。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">距離空間の場合には、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$$\forall U\in \mathcal{N}_d(x)\exists N>0\forall n>N(x_n\in U)$$</div><div style="font-size: 16px; margin: 0px; text-align: left;">となりますが、これは、</div><div style="font-size: 16px; margin: 0px;">$$\forall \epsilon>0\exists N>0\forall n>N(x_n\in B_d(x,\epsilon))$$</div><div style="font-size: 16px; margin: 0px; text-align: left;"><div style="margin: 0px;">と同値になります。</div><div style="margin: 0px;"><br /></div><div style="margin: 0px;">というのも、</div></div><div style="font-size: 16px; margin: 0px; text-align: left;">$\Rightarrow$ は、$\forall B_d(x,\epsilon)\in \mathcal{N}(x)$ </div><div style="font-size: 16px; margin: 0px; text-align: left;">に対して成り立つことから、直ちにわかり、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$\Leftarrow$ は、$\forall U\in \mathcal{N}(x)$ に対して、</div><div style="font-size: 16px;">$\exists \epsilon >0(B_d(x,\epsilon )\subset U)$</div><div style="font-size: 16px; margin: 0px; text-align: left;">が成り立つので、条件から、$\exists N>0\forall n>N$ に対して、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$x_n\in B_d(x,\epsilon )\subset U$ がなりたちます。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">また、閉集合であることから、点列の性質として</div><div style="font-size: 16px; margin: 0px; text-align: left;">つぎのようなことが言えます。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">定理9.2</div><div style="font-size: 16px; margin: 0px; text-align: left;">$(X,\mathcal{O})$ を位相空間とするとき、$F\subset X$ </div><div style="font-size: 16px; margin: 0px; text-align: left;">が閉集合なら、任意の $n$ に対して $x_n\in F$ を満たす点列 $(x_n)$ に対して、</div><div style="font-size: 16px; margin: 0px; text-align: left;">ですので、$x\in X$ に収束するなら、$x\in F$</div><div style="font-size: 16px; margin: 0px; text-align: left;">である。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">つまり、閉集合 $F$ 上の任意の点列で、$X$ に収束するものがあれば、</div><div style="font-size: 16px; margin: 0px; text-align: left;">その収束先も $F$ の中に入るということです。</div><div style="font-size: 16px; margin: 0px; text-align: left;">このような性質を点列閉という言い方をします。</div><div style="font-size: 16px; margin: 0px; text-align: left;">上の定理は、部分集合は閉ならば、点列閉となります。</div><div style="font-size: 16px; margin: 0px; text-align: left;">逆に、点列閉ならば、閉であることは一般には成り立ちません。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">(定理9.2の証明) 条件を満たす点列の収束先が $x\in F^c$ を満たすとします</div><div style="font-size: 16px; margin: 0px; text-align: left;">このとき、$F^c$ は開集合だから、$\exists U\in \mathcal{N}(x)$ </div><div style="font-size: 16px; margin: 0px; text-align: left;">であって、$U\subset F^c$ を満たします。</div><div style="font-size: 16px; margin: 0px; text-align: left;">また、収束点列であることから、$\exists N>0\forall n>N(x_n\in U)$ </div><div style="font-size: 16px; margin: 0px; text-align: left;">となりますがそのような $x_n\in F^c$ であり、$x_n\in F$ であることに</div><div style="font-size: 16px; margin: 0px; text-align: left;">反します。よって、$x\in F$ となります。(証明終了)</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="margin: 0px; text-align: left;"><span style="font-size: large;"><b><u>直積位相</u></b></span></div><div style="margin: 0px; text-align: left;">直積位相について定義しました。</div><div style="margin: 0px; text-align: left;">位相空間 $(X_i,\mathcal{O}_i)$ $i=1,2,\cdots,n$ に対して、直積集合</div><div style="margin: 0px; text-align: left;">$X_1\times X_2\times\cdots\times X_n$</div><div style="margin: 0px; text-align: left;">の上に、直積位相を次のように定めます。</div><div style="margin: 0px; text-align: left;"><br /></div><div style="margin: 0px; text-align: left;">定義9.3</div><div style="margin: 0px; text-align: left;">$U_i\in\mathcal{O}_i$ に対して、</div><div style="margin: 0px; text-align: left;">$U_1\times U_2\times \cdots \times U_n$ </div><div style="margin: 0px; text-align: left;">を開基とする位相を</div><div style="margin: 0px; text-align: left;">$X_1\times X_2\times \cdots\times X_n$ </div><div style="margin: 0px; text-align: left;">に入れたものを直積位相といい、</div><div style="margin: 0px; text-align: left;">$$\mathcal{O}_1\times \mathcal{O}_2\times \cdots \times \mathcal{O}_n$$</div><div style="font-size: 16px; margin: 0px; text-align: left;">とかく。これを<b>直積位相</b>(<b>直積位相空間</b>)という。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">上の $\mathcal{O}_1\times \mathcal{O}_2\times \cdots \times \mathcal{O}_n$ の部分に使われている記号 $\times $ は集合の直積とは意味合いが異なるので注意を</div><div style="font-size: 16px; margin: 0px; text-align: left;">してください。</div><div style="font-size: 16px; margin: 0px; text-align: left;">もし、任意の $i$ に対して $\mathcal{O}_i=\mathcal{O}$ であるなら、この直積は $\mathcal{O}^n$</div><div style="font-size: 16px; margin: 0px; text-align: left;">として書くことにします。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">標準射影 $\text{pr}_i:X_1\times \cdots\times X_n\to X_i$</div><div style="font-size: 16px; margin: 0px; text-align: left;">を $\text{pr}_i(x_1,x_2,\cdots, x_n)=x_i$</div><div style="font-size: 16px; margin: 0px; text-align: left;">とします。この写像を標準射影と言います。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">このとき、直積位相は、標準射影 $\text{pr}_i$ の全てが連続になるための</div><div style="font-size: 16px; margin: 0px; text-align: left;">最弱な位相になります。</div><div style="font-size: 16px; margin: 0px; text-align: left;">つまり、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$$\mathcal{O}_1\times \mathcal{O}_2\times \cdots\times \mathcal{O}_n=\langle \text{pr}_1,\text{pr}_2,\cdots,\text{pr}_n\rangle$$</div><div style="font-size: 16px; margin: 0px; text-align: left;">となります。</div><div style="font-size: 16px; margin: 0px; text-align: left;">左辺の開基は $U_1\times \cdots \times U_n$ の形ですが、これは、</div><div style="font-size: 16px; margin: 0px; text-align: left;">右辺において、$\text{pr}_1^{-1}(U_1)\cap\cdots\cap \text{pr}_n^{-1}(U_n)$ </div><div style="font-size: 16px; margin: 0px; text-align: left;">と表されるので、同じ開基を用いていることから両者は同じ位相ということになります。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">例9.8</div><div style="font-size: 16px; margin: 0px; text-align: left;">$({\mathbb R}^2,\mathcal{O}_{d_1}^2)$, $({\mathbb R}^2,\mathcal{O}_{d_2})$ は同値である。</div><div style="font-size: 16px; margin: 0px; text-align: left;">$\mathcal{O}_{d_1}^2=\mathcal{O}_{d_1}\times \mathcal{O}_{d_1}$ は2つのユークリッド距離位相の直積であり、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$\mathcal{O}_{d_2}$ は2次元のユークリッド位相です。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">前者は、長方形からなる開集合を開基とする位相であって</div><div style="font-size: 16px; margin: 0px; text-align: left;">後者は、円からなる開集合を開基とする位相です。</div><div style="font-size: 16px; margin: 0px; text-align: left;">これらは、位相として同値になります。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">これは、$x=(x_1,x_2)$ として、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$$B_{d_2}(x,\epsilon)\subset B_{d_1}(x_1,\epsilon)\times B_{d_1}(x_2,\epsilon)$$</div><div style="font-size: 16px; margin: 0px; text-align: left;">であること、また、</div><div style="font-size: 16px; margin: 0px; text-align: left;"><div style="margin: 0px;">$$B_{d_1}\left(x_1,\frac{\epsilon}{\sqrt{2}}\right)\times B_{d_1}\left(x_2,\frac{\epsilon}{\sqrt{2}}\right)\subset B_{d_2}(x,\epsilon)$$</div><div>が成り立つことから直ちに導かれます。</div></div><div style="font-size: 16px; margin: 0px; text-align: left;">ここに証明を書くのは面倒なので、</div><div style="font-size: 16px; margin: 0px; text-align: left;">詳しくは拙著「例題形式で探求する集合・位相」</div><div style="font-size: 16px; margin: 0px; text-align: left;">の方を読んでください。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">このことから、$\mathcal{O}_M$ を ${\mathbb R}^2$ 上のマンハッタン距離</div><div style="font-size: 16px; margin: 0px; text-align: left;">による距離位相空間とすると、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$\mathcal{O}_{d_2}=\mathcal{O}_M=\mathcal{O}_{d_1}^2$</div><div style="font-size: 16px; margin: 0px; text-align: left;">が成り立ちます。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">また、よく使う定理としては、以下があります。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">定理9.6</div><div style="font-size: 16px; margin: 0px; text-align: left;">直積位相空間 $(X_1\times \cdots\times X_n,\mathcal{O}_1\times \cdots\times \mathcal{O}_n)$</div><div style="font-size: 16px; margin: 0px; text-align: left;">と位相空間 $(Y,\mathcal{O})$ に対して、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$f:Y\to X_1\times \cdots\times X_n$ が連続であるための必要十分条件</div><div style="font-size: 16px; margin: 0px; text-align: left;">は、任意の $i$ に対して、$\text{pr}_i\circ f$ が連続であることである。。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">必要性は、$\text{pr}_i$ と $f$ が連続であることから、</div><div style="font-size: 16px; margin: 0px; text-align: left;">その合成写像も連続であることから成り立ちます。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">十分性は、直積位相空間の任意の開基 $U_1\times \cdots\times U_n$ に</div><div style="font-size: 16px; margin: 0px; text-align: left;">対して、その逆像は、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$$f^{-1}(U_1\times \cdots \times U_n)=f^{-1}(\text{pr}_1^{-1}(U_1)\cap \cdots\cap \text{pr}_n^{-1}(U_n))$$</div><div style="font-size: 16px; margin: 0px; text-align: left;">$$=f^{-1}(\text{pr}_1^{-1}(U_1))\cap \cdots\cap f^{-1}(\text{pr}_n^{-1}(U_n))$$</div><div style="font-size: 16px; margin: 0px; text-align: left;">$$=(\text{pr}_1\circ f)^{-1}(U_1)\cap \cdots\cap (\text{pr}_n\circ f)^{-1}(U_n)$$</div><div style="font-size: 16px; margin: 0px; text-align: left;">よって、この各成分は開集合であり、位相の条件(II)からこれらは、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$\mathcal{O}$ の開集合になります。</div><div style="font-size: 16px; margin: 0px; text-align: left;">よって十分性が成り立ちます。 $\Box$</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">最後に無限直積の直積位相を導入しました。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">$(X_\lambda,\mathcal{O}_\lambda)$ $\lambda\in\Lambda$</div><div style="font-size: 16px; margin: 0px; text-align: left;">を位相空間の族とします。</div><div style="font-size: 16px; margin: 0px; text-align: left;">$\prod_{\lambda\in \Lambda}X_\lambda$ 上に</div><div style="font-size: 16px; margin: 0px; text-align: left;">直積位相を誘導位相</div><div style="font-size: 16px; margin: 0px; text-align: left;">$\langle \{\text{pr}_\lambda|\lambda\in \Lambda\}\rangle$</div><div style="font-size: 16px; margin: 0px; text-align: left;">として定義します。</div><div style="font-size: 16px; margin: 0px; text-align: left;">このとき、直積集合 $\prod_{\lambda\in \Lambda}X_\lambda$</div><div style="font-size: 16px; margin: 0px; text-align: left;">上に位相を導入することができました。</div><div style="font-size: 16px; margin: 0px; text-align: left;">このような位相空間は開基として、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$\{\lambda_1,\cdots,\lambda_n\}\subset \Lambda$ に対して、</div><div style="font-size: 16px; margin: 0px; text-align: left;">$\text{pr}_{\lambda_1}^{-1}(U_{\lambda_1})\cap \cdots\cap \text{pr}_{\lambda_n}^{-1}(U_{\lambda_n})$</div><div style="font-size: 16px; margin: 0px; text-align: left;">となるものを集めたものになります。</div><div style="font-size: 16px; margin: 0px; text-align: left;">ここで、$U_{\lambda_i}\in \mathcal{O}_{\lambda_i}$ が成り立つ任意の開集合になります。</div><div style="font-size: 16px; margin: 0px; text-align: left;"><br /></div><div style="font-size: 16px; margin: 0px; text-align: left;">つまり、直積位相の開集合において、$(x_\lambda)\in \prod_{\lambda\in \Lambda}X_\lambda$</div><div style="font-size: 16px; margin: 0px; text-align: left;">の各成分 $x_\lambda\in X_\lambda$ </div><div style="font-size: 16px; margin: 0px; text-align: left;">は、有限個の$\lambda_{i}\in \Lambda$ 以外の $j$ は全て $X_j$ </div><div style="font-size: 16px; margin: 0px; text-align: left;">とるような開集合となります。 </div></div></div>
</div>
</div>
</div></div></div></div></div>Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-45696562808931253972021-12-08T12:18:00.005+09:002022-04-19T10:35:48.753+09:00トポロジー入門(第8回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所:オンライン(月曜日3限)]</span></div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px; margin: 0px;">
<div style="margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/21/top.html">トポロジー入門のHP</a></div></div><div style="margin: 0px;"><div style="font-size: 16px;"><div style="font-size: 16px;"><div style="margin: 0px;"><div style="font-size: 16px;">
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">一昨年(2019年度)トポロジー入門の講義の内容をまとめていましたが、</div><div style="font-size: 16px;">7回目を残したまま更新しませんでした。</div><div style="font-size: 16px;">2021年度と去年は2019年度とほぼ同じ速度で進んでいます。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">後半の8回以降のトポロジー入門のブログを再開します。</div><div style="font-size: 16px;">また、勝手な都合でまた休止するかもしれません。</div>
<div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><div>今回は、開基についての解説の続きと</div><div>生成される位相と相対位相について説明をしました。</div></div><div style="font-size: 16px;"><br /></div>
<div style="font-size: 16px;">
<b>定理8.1</b></div>
<div style="font-size: 16px;"><div>$(𝑋,\mathcal{O}_X),(Y,\mathcal{O}_Y)$ を位相空間とする。</div><div>$\mathcal{B}\subset O_Y$ を開基とする。</div><div>$f:X\to Y$ が連続であるためには、</div><div>$\forall V\in \mathcal{B}\to f^{-1}(V)\in \mathcal{O}_X$ であること</div><div>が必要十分である。</div></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明)</div><div style="font-size: 16px;"><div>𝑓が連続であれば、</div><div>$U\in \mathcal{O}_Y$ であるの、</div><div>とくに $\mathcal{B}$ に対して正しい。</div><div>逆に、 $\forall 𝑉\in \mathcal{B}\to f^{-1}(V)\in \mathcal{O}_X$ が成り立 つとする。</div><div>$\forall \in \mathcal{O}_Y$ に対して、 $\mathcal{B}'\subset \mathcal{B}$ が存在して、 $U=\cup_{B\in \mathcal{B}'}B$ と書ける。</div><div>$f^{-1}(U)=f^{-1}(\cup_{B\in \mathcal{B}'}B)=\cup_{B\in \mathcal{B}'}f^{-1}(B)$ よって $f$ が連続である。$\Box$</div></div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">講義では言いませんでしたが、次の定理も示しておきます。</div><div style="font-size: 16px;">$\mathcal{B}\subset \mathcal{P}(X)$ がある</div><div style="font-size: 16px;">位相の開基となるためには以下の条件が必要です。</div><div style="font-size: 16px;"><br />
<b>定理8.$1\frac{1}{2}$</b><br />
$X$を集合とし、$\mathcal{B}\subset \mathcal{P}(X)$ とするとき、<br />
$\mathcal{B}$ が $X$ 上のある位相の開基となるための条件は<br />
以下の2つを満たすことであり、その位相は$\mathcal{O}_B=\{\cup \mathcal{B}'|\mathcal{B}'\subset \mathcal{B}\}$ となる。<br />
(i) $\cup\mathcal{B}=X$<br />
(ii) $\forall B_1\in B_2\in \mathcal{B}$ と $\forall p\in B_1\cap B_2$ に対して、<br />
$\exists B\in \mathcal{B}(B\subset B_1\cap B_2)$ を満たす。<br />
<br />
(証明)<br />
$\mathcal{O}$が$\mathcal{B}$ を開基とする位相とします。<br />
このとき、$\mathcal{O}$は$\mathcal{O}_\mathcal{B}=\{\cup\mathcal{B}'|\mathcal{B}'\subset \mathcal{B}\}$ となります。<br />
(i),(ii)は、位相の条件(I)の$X\in \mathcal{O}$ から、<br />
$\forall p\in X$ に対して、$V_p\in \mathcal{B}$ が存在して、$p\in V_p$<br />
となります。<br />
よって 、そのような$V_p$ に対して $X=\cup_{p\in X}V_p\subset \cup \mathcal{B}$<br />
となります。<br />
よって、(i)が成り立ちます。<br />
つぎに、(ii)をみます。<br />
$\forall B_1,B_2\in \mathcal{B}\subset\mathcal{O}$ に対して、<br />
$B_1\cap B_2\in\mathcal{O}$ であるので、$p\in B\in \mathcal{B}$<br />
が存在して、$B\subset B_1\cap B_2$ となります。よって(ii)が成り立ちます。<br />
<br />
つぎに、(i)と(ii)が成り立つと仮定します。<br />
このとき、$\mathcal{O}_\mathcal{B}=\{\cup\mathcal{B}'|\mathcal{B}'\subset \mathcal{B}\}$<br />
が開集合系となることを示します。<br />
開集合系の(I)が成り立つことを示します。<br />
$\emptyset$は、$\mathcal{B}'=\emptyset$ とすればよく成り立ちます。<br />
(i)は$X\in \mathcal{O}$を意味します。<br />
(II)が成り立つことを示します。<br />
$U,V\in \mathcal{O}_\mathcal{B}$ とすると、<br />
$U=\cup\mathcal{B}_U$かつ $V=\cup\mathcal{B}_V$ となります。<br />
ただし、$\mathcal{B}_U\subset\mathcal{B}$ かつ $\mathcal{B}_V\subset \mathcal{B}$<br />
です。<br />
このとき、$U\cap V=\cup_{B_U\in\mathcal{B}_U,B_V\in \mathcal{B}_V}B_U\cap B_V$<br />
が成り立ちます。<br />
よって、$\forall p\in B_U\cap B_V$に対して、$p\in B\in \mathcal{B}$<br />
となる$B$が存在します。<br />
よって、$\mathcal{B}_{U,V}\subset \mathcal{B}$に対して、<br />
$\cup \mathcal{B}_{U,V}=B_U\cap B_V$ となります。<br />
よって、$\cup \{\mathcal{B}_{U,V}|U\in \mathcal{B}_U,V\in \mathcal{B}_V\}\subset \mathcal{B}$<br />
であり、この和集合は、 $U\cap V$ となります。<br />
つまり、位相の条件の(II)が成り立つことになります。<br />
<br />
(III)が成り立つことをしめします。<br />
また、$\mathcal{U}=\{U_\lambda\in \mathcal{O}_{\mathcal{B}}|\lambda\in \Lambda\}$<br />
とすると、$U_\lambda\in \mathcal{U}$ に対して、ある$\mathcal{B}_\lambda\subset \mathcal{B}$<br />
が存在して、$U_\lambda=\cup \mathcal{B}_\lambda$となります。<br />
$\cup\mathcal{U}=\cup_{\lambda\in \Lambda}(\cup \mathcal{B}_\lambda)$ですから、<br />
$\mathcal{B}_\mathcal{U}=\cup \{\mathcal{B}_\lambda|\lambda\in\Lambda\}$とすると、<br />
$\cup\mathcal{U}=\cup \mathcal{B}_\mathcal{U}$ が成り立つ。<br />
よって、$\mathcal{O}_\mathcal{B}$ は開集合系であり、<br />
構成から、$\mathcal{B}$ は $\mathcal{O}$ の開基となります。<br />
<br /></div><div style="font-size: 16px;">また、以下の定理を示しました。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><b>定理8.2.</b></div><div style="font-size: 16px;">$({\mathbb R},\mathcal{O}_l)$ は可分であるが、第2可算公理を満たさない。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明の概略)</div><div style="font-size: 16px;">可分であることは、下限位相においても ${\mathbb Q}$ </div><div style="font-size: 16px;">が稠密部分集合であることから成り立ちます。</div><div style="font-size: 16px;">また、逆に、任意に開基 $\mathcal{B}$ を取った時に、</div><div style="font-size: 16px;">任意の半開区間 $[p,p+1)$ において、$p\in V_p\subset [p,p+1)$ </div><div style="font-size: 16px;">となる $V_p\in \mathcal{B}$ が存在します。</div><div style="font-size: 16px;">ここで、$\{V_p|p\in {\mathbb R}\}\subset \mathcal{B}$</div><div style="font-size: 16px;">であり、この右辺は非可算集合であるので、第2可算公理を満たしません。$\Box$</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">このことから、下限位相は距離化可能でないことがわかります。</div><div style="font-size: 16px;">これは距離化可能であれば、可分であることと第2可算公理を満たすことが</div><div style="font-size: 16px;">同値となるからです。</div><div style="font-size: 16px;"><br /></div>
<span style="font-size: large;"><u>生成される位相</u></span><br />
<div style="font-size: 16px;">
を定義します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定義8.1 </b></div>
<div style="font-size: 16px;">
$X$ を集合とする。$\mathcal{S}\subset\mathcal{P}(X)$ とする、</div>
<div style="font-size: 16px;">
$\mathcal{S}$ の全ての元を開集合として含む $X$ 上の最弱の位相を</div>
<div style="font-size: 16px;">
$\langle \mathcal{S}\rangle$ とかいて、$\mathcal{S}$ によって<b>生成される位相</b>という。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
そのような位相はただ一つ存在し、</div>
<div style="font-size: 16px;">
$\langle \mathcal{S}\rangle=\underset{\mathcal{S}\subset \mathcal{O}:\text{位相}}{\cap}\mathcal{O}$</div>
<div style="font-size: 16px;">
となります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">生成される位相の例として、$\mathcal{O}$ を開集合系とすると、$\mathcal{O}=\langle \mathcal{O}\rangle$</div>
<div style="font-size: 16px;">
となります。</div>
<div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">例8.1 $\mathcal{S}=\{\{0,1\},\{1,2\}\}$ としたとき、</div><div style="font-size: 16px;">$\langle \mathcal{S}\rangle=\{\emptyset,\{1\},\{0,1\},\{1,2\},\{0,1,2\}\}$</div><div style="font-size: 16px;">となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">また、次のように定義します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定義8.2</b> 位相空間 $(X,\mathcal{O})$ に対して $\langle \mathcal{S}\rangle=\mathcal{O}$</div>
<div style="font-size: 16px;">
とするとき、$\mathcal{S}$ は $\mathcal{O}$ の<b>準開基</b>と言う。</div>
<div style="font-size: 16px;"><br /></div>
<div style="font-size: 16px;">ここで、次を示します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定理8.3.</b></div>
<div style="font-size: 16px;">
$\langle \mathcal{S}\rangle$ は、$\mathcal{B}=\{U_1\cap\cdots\cap U_n|n\in{\mathbb N}_0,0\le \forall i\le n(U_i\in S)\}$</div>
<div style="font-size: 16px;">
としたとき、$\langle \mathcal{S}\rangle$ は $\mathcal{B}$ を開基とする位相になる。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">ここで${\mathbb N}_0={\mathbb N}\cup \{0\}$ です。</div>
<div style="font-size: 16px;">
これを証明をします。</div><div style="font-size: 16px;"><br /></div>
<div style="font-size: 16px;">
(証明)まず、$\mathcal{B}$ が位相の開基となることを証明します。<br />
つまり、定理8.$1\frac{1}{2}$ の(i),(ii)が成り立つことを示します。</div>
<div style="font-size: 16px;">
(i) $X\in \mathcal{B}$ であるから、あきらかに $X=\cup\mathcal{B}$です。<br />
(ii) $U_1,\cdots, U_n,V_1,\cdots ,V_m\in \mathcal{B}$ として、<br />
$B_1=U_1\cap \cdots \cap U_n\in \mathcal{B}$ かつ $B_2=V_1\cap \cdots \cap V_m\in \mathcal{B}$<br />
$B_1\cap B_2=U_1\cap \cdots \cap U_n\cap V_1\cap \cdots \cap V_m\in \mathcal{B}$ ですから、<br />
(ii) は明らかに成り立ちます。<br />
よって、$\mathcal{B}$ は $X$ 上のある位相の開基となります。</div>
<div style="font-size: 16px;">
その位相を$\mathcal{O}_{\mathcal{S}}$ とすると、生成される位相の定義から</div><div style="font-size: 16px;">$\mathcal{S}\subset \mathcal{O}_{\mathcal{S}}$ であり、その最小性から、$\langle \mathcal{S}\rangle \subset \mathcal{O}_{\mathcal{S}}$ となります。</div>
<div style="font-size: 16px;">また、任意に $U\in \mathcal{O}_{\mathcal{S}}$ をとると、</div><div style="font-size: 16px;">$\mathcal{B}'\subset \mathcal{B}$ が存在して、$U=\cup \mathcal{B}'$</div><div style="font-size: 16px;">となり、任意の $U_1\cap U_2\cap \cdots\cap U_n\in \mathcal{B}'$</div><div style="font-size: 16px;">に対して、開集合系の条件(II)から $U_1\cap U_2\cap \cdots\cap U_n\in \langle \mathcal{S}\rangle$</div><div style="font-size: 16px;">であり、開集合系の条件(III)から $\cup \mathcal{B}'\in \langle \mathcal{S}\rangle$</div><div style="font-size: 16px;">が成り立ちます。</div><div style="font-size: 16px;">つまり $\mathcal{O}_{\mathcal{S}}\subset \langle \mathcal{S}\rangle$ であることから</div><div style="font-size: 16px;">$\langle \mathcal{S}\rangle =\mathcal{O}_{\mathcal{S}} $となります。$\Box$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;"><b><br /></b></div><div style="font-size: 16px;"><div>例8.4</div><div>$\mathcal{O}_{d_1}$ の時、$\mathcal{S}=\{(-\infty,a)|a\in {\mathbb R}\}\cap\{(b,\infty)|b\in{\mathbb R}\}$</div><div>とすると、$\mathcal{S}$ は$\mathcal{O}_{d_1}$ の準開基となります。</div><div>というのも、$(-\infty,a),(b,\infty)$ が $\mathcal{S}$ に入るので、</div><div>そのような位相は、必ず $(b,a)$ も入るので、任意の開区間を含む位相ということになります。</div><div>$\mathcal{B}=\{(a,b)|a,b\in {\mathbb R}\}$ としておくと、このような位相は</div><div>$\mathcal{B}$ を開基とする位相空間ということになります。</div><div>そのような位相は、$\mathcal{O}_{d_1}$ ということになります。</div><div><br /></div><div>$\mathcal{S}$ の有限共通部分をとると、</div><div>$n=0$ の場合には ${\mathbb R}$</div><div>が成り立ち、</div><div>$n=1$ の場合には、$\mathcal{S}$ となります。</div><div>また、$n=2$ の場合、</div><div>$\{(-\infty,a)|a\in {\mathbb R}\}$ と $\{(b,\infty)|b\in{\mathbb R}\}$</div><div>のそれぞれから選んで共通部分を取ると、有限開区間 $(a,b)$ もしくは空集合が作られます。</div><div>同じ側から選ぶと、$(-\infty,a)$ の形の開集合か、$(b,\infty)$ の形の開集合のどちらかです。結果的に $\mathcal{S}$ の形に戻ります。</div><div><br /></div><div>3こ選んできたときも同様に考えると、$\mathcal{S}$ の形の開集合か、空集合か、</div><div>$(a,b)$ と $(-\infty,c)$ もしくは $(d,\infty)$ の共通部分となります。</div><div>それらは開区間か空集合なのでそれらを合わせても $\mathcal{B}\cup \{\emptyset\}$ から</div><div>外に出ることはありません。</div><div><br /></div></div><div style="font-size: 16px;">
<b>定義8.3</b></div>
<div style="font-size: 16px;">
$(Y,\mathcal{O})$ <span style="color: black;">を位</span><span style="color: black;">相空間とし、</span></div>
<div style="font-size: 16px;">
<span style="color: black;">$f:X\to Y$ を連続写像とします。</span></div>
<div style="font-size: 16px;">
<span style="color: black;">このとき、</span></div>
<div style="font-size: 16px;">
<span style="color: black;">$S=\{f^{-1}(U)|U\in \mathcal{O}\}$</span></div>
<div style="font-size: 16px;">
<span style="color: black;">$\langle S\rangle$ を </span>$\langle f\rangle$と書き、<b>$f$ によって</b></div>
<div style="font-size: 16px;">
<span style="color: black;"><b>誘導された位相(誘導位相)</b>ということにす</span>る。</div>
<div style="font-size: 16px;"><br /></div>
<div style="font-size: 16px;">
命題</div>
<div style="font-size: 16px;">
$f:X\to Y$を写像とする。</div>
<div style="font-size: 16px;">
$\langle f\rangle =\{f^{-1}(U)|U\in \mathcal{O}_{Y}\}$</div>
<div style="font-size: 16px;">
である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明)</div><div style="font-size: 16px;">$\mathcal{B}=\{f^{-1}(U_1)\cap \cdots \cap f^{-1}(U_n)|U_i\in \mathcal{O}_Y\}$ とすると、</div><div style="font-size: 16px;">$\langle f\rangle$ は $\mathcal{B}$ を開基とする位相となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;"><div>$f^{-1}(U_1)\cap \cdots \cap f^{-1}(U_n)=f^{-1}(U_1\cap \cdots\cap U_n)$</div><div>ですから、$\mathcal{B}=\{f^{-1}(U)|U\in \mathcal{O}_Y\}$ となります。</div><div>また、$\cup_{U\in \mathcal{B}'\subset\mathcal{B}}f^{-1}(U)=f^{-1}(\cup \mathcal{B}')$</div><div>が成り立つので、</div></div>
<div style="font-size: 16px;">
$\langle f\rangle=\langle \{f^{-1}(U)|U\in \mathcal{O}_Y\}\rangle=\mathcal{O}_\mathcal{B}=\mathcal{B}$</div>
<div style="font-size: 16px;">がなりたちます。 $\Box$</div>
<div style="font-size: 16px;">
<br /></div><div style="font-size: 16px;">誘導位相で重要な性質は以下です。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">定理8.3</div><div style="font-size: 16px;">$Y$ を位相空間とする。</div><div style="font-size: 16px;">$f:X\to Y$ を写像とすると、誘導位相 $\langle f\rangle$ は</div><div style="font-size: 16px;">$f$ を連続にする $X$ の最小の位相である。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">(証明略)</div><div style="font-size: 16px;"><br /></div>
<div style="font-size: 16px;">
次に、相対位相の定義をします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
定義8.4</div>
<div style="font-size: 16px;">
$(X,\mathcal{O})$ を位相空間とし、$A\subset X$ とする。</div>
<div style="font-size: 16px;">
このとき、$\mathcal{O}_A=\{A\cap U|U\in \mathcal{O}\}$ として</div>
<div style="font-size: 16px;">
定義すると、$(A,\mathcal{O}_A)$ は、$A$ の位相空間を与えており、</div>
<div style="font-size: 16px;">
それを<b>部分位相空間</b>という。また、$\mathcal{O}_A$ を<b>相対位相</b>ともいう。</div>
<div style="font-size: 16px;">
<br /></div><div style="font-size: 16px;">つまり、開集合を $A$ に制限して得られるような開集合全体を</div><div style="font-size: 16px;">$A$ の相対位相ということになります。</div><div style="font-size: 16px;"><br /></div>
<div style="font-size: 16px;">
定理8.5</div>
<div style="font-size: 16px;">
$A\subset X$ を部分集合とする。</div>
<div style="font-size: 16px;">
$i:A\hookrightarrow X$ を包含写像とする。</div>
<div style="font-size: 16px;">
このとき、$(A,\langle i\rangle)$ は相対位相 $(A,\mathcal{O}_A)$ と一致する。</div><div style="font-size: 16px;"><br /></div>
<div style="font-size: 16px;">
(証明)$i^{-1}(U))=A\cap U$ であることを示されればよいです。</div>
<div style="font-size: 16px;">
$\forall a\in i^{-1}(U)$ とすると、$i(a)=a\in U$ であるから、$a\in A\cap U$ であり、</div>
<div style="font-size: 16px;">
$a\in A\cap U$ とすると、$i(a)=a\in U$ であるから、$a\in i^{-1}(U)$ </div>
<div style="font-size: 16px;">
ですから、$i^{-1}(U)=A\cap U$ が成り立ちます。$\Box$</div>
<div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">例8.5</div><div style="font-size: 16px;">${\mathbb R}$ の通常の距離位相空間において、$A=(0,1)$ </div><div style="font-size: 16px;">上の相対位相は ${\mathbb R}$ の開集合 $U$ を用いて</div><div style="font-size: 16px;">$(0,1)\cap U$ となる開集合ですが、</div><div style="font-size: 16px;">$(0,1)$ は ${\mathbb R}$ の開集合なので、$(0,1)\cap U$ </div><div style="font-size: 16px;">も$A$での開集合となります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">例8.6</div><div style="font-size: 16px;"><div>${\mathbb R}$ の通常の距離位相空間において、$A=[0,1]$ </div><div>上の相対位相は ${\mathbb R}$ の開集合 $U$ を用いて</div><div>$[0,1]\cap U$ となる開集合ですが、</div><div>$[0,1]$ は ${\mathbb R}$ の開集合ではないので、$[0,1]\cap U$</div><div>は $A$ での開集合ですが、一般には ${\mathbb R}$ での開集合に</div><div>なりません。</div><div>例えば、$U=(-1/2,1/2)$ としたときには、</div><div>$A\cap U=[0,1/2)$ となって、これは ${\mathbb R}$ での開集合では</div><div>ありませんが、$A$ の開集合となります。</div><div><br /></div><div>例8.7</div><div>$A={\mathbb Z}\subset {\mathbb R}$ を選んでやると、</div></div><div style="font-size: 16px;">$A$ 上で制限してできる開集合は、$A$ 上の離散位相空間になります。</div><div style="font-size: 16px;"><br /></div><div style="font-size: 16px;">例8.8</div><div style="font-size: 16px;">$A=\{1/n|n\in {\mathbb N}\}$ を選んでやると、</div><div style="font-size: 16px;">やはり$A$ 上の離散位相空間が得られますが、</div><div style="font-size: 16px;">$\overline{A}=A\cup\{0\}$ となり、</div><div style="font-size: 16px;">$\overline{A}$ 上の相対位相は、$\overline{A}$ 上の</div><div style="font-size: 16px;">離散位相とは異なる位相空間になります。</div><div style="font-size: 16px;">離散位相空間には、集積点は含まれませんが、$\overline{A}$</div><div style="font-size: 16px;">上の位相は集積点 $0$ が含まれます。</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-18187006507583915002020-07-20T14:26:00.002+09:002020-07-20T14:59:43.303+09:00積分と極限の順序交換定理(アルゼラの定理)<div dir="ltr" style="text-align: left;" trbidi="on">
積分と極限の交換定理について書きます。<br />
積分と極限の交換とは<br />
$$\lim_{n\to\infty}\int_a^bf_n(x)dx=\int_a^b\lim_{n\to\infty}f_n(x)dx$$<br />
が成り立つことを言います。<br />
<br />
微積分で最初に習う積分と極限の交換のための(十分)条件は、<br />
おそらく、その関数列が一様収束する場合だと思いますが、<br />
関数列が一様収束しなくても積分と極限の交換が成り立つ場合がありますので<br />
下で紹介し、証明します。<br />
<br />
一様収束する関数列の積分と極限の順序交換については、<br />
このブログの2016年の微積分II演習のページ(<a href="https://motochans.blogspot.com/2016/02/ii15.html">リンク</a>)<br />
に書いていますのでそちらを参照してください。<br />
一様収束の定義もそちらを参照してください。<br />
<br />
このページでは基本、大学1年生向け微積分を扱いますので<br />
ルベーグ積分は仮定せずに、リーマン積分の範疇だけで話を進めます。<br />
ルベーグ積分を学んだあとであれば、これらの積分と極限の交換定理は<br />
ルベーグの収束定理から直ちに導かれます。<br />
<br />
具体的には下の定理を証明します。<br />
この証明は、小平邦彦著の「解析入門」(岩波基礎数学)の第5章にあり、<br />
「解析入門」によると、Hausdorffの1927年の論文(Beweis eines Satzes von Arzelà)<br />
が元だそうです。(私は論文を読んだわけではありません。)<br />
また、この定理は、解析学のアスコリ=アルゼラの定理(<a href="https://ja.wikipedia.org/wiki/%E3%82%A2%E3%82%B9%E3%82%B3%E3%83%AA%EF%BC%9D%E3%82%A2%E3%83%AB%E3%83%84%E3%82%A7%E3%83%A9%E3%81%AE%E5%AE%9A%E7%90%86">wikipedia</a>)としても一般化されます。<br />
<b></b><br />
<u><span style="font-size: large;">アルゼラの定理</span></u><br />
<b>定理(積分と極限の順序交換定理(アルゼラの定理))</b><br />
<b>$f_n$ を区間 $I=[a,b]$ で定義された連続関数列とし、ある実数 $M>0$ が存在して、任意の $n\in {\mathbb N}$ と $x\in I$ に対して、$|f_n(x)|\le M$ が成り立つとする。この関数列は $I$ 上で $f(x)$ に各点収束し、$(a,b)$ で連続であるとする。このとき以下の積分と極限の順序交換が成り立つ。</b><br />
<b>$$\lim_{n\to \infty}\int_a^bf_n(x)dx=\int_a^b\lim_{n\to \infty}f_n(x)dx$$</b><br />
<b></b><br />
「解析入門」の本文では、$f(x)$ の連続性は、$(a,b)$ ではなく、$[a,b]$ となっていますが<br />
下の証明にあるように、$(a,b)$ でだけの連続であれば成り立ちます。<br />
<br />
リーマン積分可能であれば、連続である必要もないのですが、<br />
リーマン可積分の条件などを証明途中で使うのは面倒なので連続関数の積分のみを用いて証明します。もし、不連続点をさらに多く含むような場合を考えたい場合は、上に書いたようにルベーグ積分を勉強して理解することで全て解決されます。<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<br /></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
定理の中に記述されている<br />
「ある実数 $M>0$ が存在して、任意の $n\in {\mathbb N}$ と $x\in I$に対して、</div>
$|f_n(x)|\le M$ が成り立つ」ことを<b>一様有界</b>といいます。<br />
では証明していきます。<br />
(「解析入門」とは、文字など適宜変えた部分がありますのでご了承ください。)<br />
<br />
<u>また、証明は少々ですが長いので読む覚悟のある人以外は証明の内容は理解せず、</u><br />
<u>定理の内容だけ理解して、下の方にある例にまでスキップして使い方だけ</u><br />
<u>学んでもよいかと思います。</u><br />
<u></u><br />
<b>(証明)</b><br />
$|\int_a^bf_n(x)dx-\int_a^bf(x)dx|\le \int_a^b|f_n(x)-f(x)|dx$<br />
から、$g_n(x)=|f_n(x)-f(x)|$ とおいたとき、<br />
$$\lim_{n\to\infty}\int_a^bg_n(x)dx=0\cdots (\ast)$$<br />
であることを証明すればよいことになります。<br />
<br />
$f_n(x)$は一様有界なので、ある$M$が存在して、$|f_n(x)|<M$ が成り立ちます。<br />
また、各点の極限をとることで、$|f(x)|\le M$が成り立つので、<br />
$|g_n(x)|\le |f_n(x)|+|f(x)|\le 2M$ が成り立ちます。<br />
$f_n(x)$ は各点収束するので、任意の $x$ に対して $g_n(x)\to 0$ となります。<br />
<br />
各点$a\le x\le b$ に対して、数列 $g_n(x),g_{n+1}(x),\cdots $ の上限を<br />
$s_n(x)$ とします。つまり、<br />
$$s_n(x)=\underset{m\ge n}{\sup}g_m(x)$$<br />
であり、$a\le x\le b$ に対して<br />
$$2M\ge s_1(x)\ge s_2(x)\cdots s_n(x)\ge \cdots$$<br />
したがって、$g_n(x)\to 0$であるから、<br />
$$\lim_{n\to \infty}s_n(x)=\limsup_{n\to \infty}g_n(x)=0\cdots(\dagger)$$<br />
が成り立ちます。<br />
<br />
ここで、$s_n(x)$ が連続である場合には下のDiniの定理により、<br />
$s_n(x)\to 0$は、$a\le x\le b$ で $0$を与える関数に<br />
一様収束し、$\lim_{n\to\infty}\int_a^bs_n(x)dx=0$ となります。<br />
<br />
そこで、$s_n(x)$ が連続とは限らない場合を考えます。<br />
ここで、<br />
$$S_n=\underset{h\le s_n}{\sup}\int_a^bh(x)dx$$<br />
とします。<br />
この上限は、任意の$x\in I$ に対して、$h(x)\le s_n(x)$ を満たすような<br />
連続関数全体に対して上限を取ることを意味します。<br />
そのような $h(x)$ に対して、<br />
$\int_a^bh(x)dx\le \int_a^b2Mdx=2M(b-a)$ であるから、<br />
$$2M(b-a)\ge S_1\ge S_2\ge \cdots \cdots \ge S_n\ge \cdots$$<br />
であり、$g_n(x)\le s_n(x)$ であるから、<br />
$$0\le \int_a^bg_n(x)dx\le S_n$$<br />
<br />
よって、$S_n\to 0$であることを示せば $(\ast)$ が成り立つことになります。<br />
任意の正数 $\epsilon>0$ に対して、$\epsilon_n=\epsilon/2^n$と定義します。<br />
このとき、<br />
$$\int_a^bh_n(x)dx> S_n-\epsilon_n\cdots (\ast\ast)$$<br />
となる連続関数 $h_n(x)$ で<br />
$h_n\le s_n$ となるものが存在します。<br />
<br />
ここで、$k_n(x)=\min\{h_1(x),\cdots ,h_n(x)\}$<br />
とおきます。このとき、下の補題を用いれば、$k_n(x)$ も連続関数になります。<br />
<br />
そして、明らかに、$k_1(x)\ge k_2(x)\ge \cdots \ge k_n(x)\ge \cdots$<br />
かつ<br />
$k_n(x)\le h_n(x)\le s_n(x)$ が成り立ち、これから、<br />
$$\int_a^bk_n(x)dx>S_n-\epsilon_1-\epsilon_2-\cdots-\epsilon_n$$<br />
が成り立ちます。<br />
<br />
実際、<br />
$\mu_n(x)=\max\{k_{n-1}(x),h_n(x)\}$ と定義します。<br />
このとき、$\mu_n(x)$ は $I$上で連続であり、<br />
($\max$や$\min$が連続性を保持すること、つまり、$f_1(x)$ と $f_2(x)$ が<br />
連続なら、$\max\{f_1(x),f_2(x)\}$ や $\min\{f_1(x),f_2(x)\}$ が連続であること<br />
は、このページの下の方にある補題からわかります)<br />
<br />
$\min\{\xi,\eta\}+\max\{\xi,\eta\}=\xi+\eta$ であるから、<br />
$k_n(x)+\mu_n(x)=k_{n-1}(x)+h_n(x)$<br />
であり、<br />
$$\int_a^bk_n(x)dx=\int_a^bk_{n-1}(x)dx+\int_a^bh_n(x)dx-\int_a^b\mu_n(x)dx$$<br />
ここで、<br />
$k_{n-1}(x)\le s_{n-1}(x)$<br />
かつ<br />
$h_n(x)\le s_n(x)\le s_{n-1}(x)$<br />
であるから、<br />
$\mu_n(x)\le s_{n-1}(x)$<br />
であり、したがって、<br />
$\int_a^b\mu_n(x)dx\le S_{n-1}$ となります。<br />
よって、この不等式と不等式 $(\ast\ast)$ を用いれば、<br />
$$\int_a^bk_n(x)dx>\int_a^bk_{n-1}(x)dx+S_n-\epsilon_n-S_{n-1}$$<br />
が成り立ちます。<br />
$\int_a^bk_1(x)dx=\int_a^bh_1(x)dx>S_1-\epsilon_1$であるから、<br />
この不等式を繰り返すことで、<br />
$$\int_a^bk_n(x)dx>(S_n-\epsilon_n-S_{n-1})+(S_{n-1}-\epsilon_{n-1}-S_{n-2})+\cdots+(S_{2}-\epsilon_2-S_{1})+\int_a^bk_1(x)dx>S_n-\epsilon_1-\epsilon_2-\cdots-\epsilon_n$$<br />
となります。<br />
<br />
よって、$\epsilon_1+\epsilon_2+\cdots+\epsilon_n<\epsilon$であるから、<br />
$\int_a^bk_n(x)dx>S_n-\epsilon$ が成り立ちます。<br />
<br />
ここで、$k_n(x)$ は連続関数列であり、<br />
$$k_1(x)\ge k_2(x)\ge k_2(x)\ge \cdots \ge k_n(x)\ge\cdots $$<br />
であり、$0\le k_n(x)\le s_n(x)$ であるから、<br />
$(\dagger)$ において各 $a\le x\le b$ に対して、$k_n(x)\to 0$ となります。<br />
つまり、$k_n(x)$ は単調非増加な連続関数であり、恒等的に $0$ となる関数に<br />
各点収束するから、Diniの定理から、$k_n(x)$ は 一様収束し、<br />
$\lim_{n\to \infty}\int_a^bk_n(x)dx=0$ となります。<br />
よって、$\lim_{n\to\infty}S_n\le \epsilon$ が成り立つので、<br />
$$\lim_{n\to\infty}S_n=0$$<br />
が成り立ちます。$\Box$<br />
<br />
<u><span style="font-size: large;">Diniの定理</span></u><br />
<b>定理(Diniの定理)</b><br />
<b>閉区間 $[a,b]$ において、連続な関数 $f_n(x)$ を項とする単調非増加な</b><br />
<b>関数列とする。つまり、$a\le x\le b$ において、</b><br />
<b>$$f_1(x)\ge f_2(x)\ge \cdots \ge f_n(x)\ge\cdots $$</b><br />
<b>となる関数列とし、連続関数 $f(x)$ に収束するとする。</b><br />
<b>このとき、$f_n(x)$ は $[a,b]$ で $f(x)$ に一様収束する。</b><br />
<b></b><br />
(証明)<br />
$g_n(x)=|f_n(x)-f(x)|$ とするとき、$g_n(x)$ が一様に $0$ に収束することを<br />
示せばよいことになります。<br />
仮定から、$g_n(x)\le g_{n-1}(x)$ より、$g_n(x)$ は単調非増加で<br />
$0$ に収束します。 <br />
この関数列が一様収束ではないとします。<br />
このとき、ある正の実数 $\epsilon_0$ に対して、どんな自然数 $n$ に対しても<br />
$m>n$ と$ c_n\in [a,b]$ が存在して、$g_m(c_n)>\epsilon_0$ が存在します。<br />
ここで、$g_n(c_n)\ge g_m(c_n)>\epsilon_0$ が成り立つことがわかります。<br />
つまり、$g_n(c_n)>\epsilon_0$ が成り立ちます。<br />
ここで、ボルツァーノ-ワイエルシュトラスの定理により、<br />
$c_n$ には $[a,b]$ において収束する部分列が存在します。<br />
<br />
そのような部分列は自然数の単調増加列 $n_1<n_2<n_3<\cdots$ を用いて<br />
$c_{n_1},c_{n_2},\cdots, c_{n_j}\cdots $ であるとし、その極限を $c\in[a,b]$ とすると、<br />
$g_n$ は連続であるから、$g_n(c_{n_j})\to g_n(c)$ となります。<br />
ここで、$n_j\ge n$ なる任意の$j$ に対して $g_n(c_{n_j})\ge g_{n_j}(c_{n_j})\ge \epsilon_0$<br />
が成り立ちます。$g_n$ の連続性から<br />
$g_n(c)\ge \epsilon_0>0$ が成り立ち、これは、$g_n(x)$ が $0$ に<br />
収束することに矛盾します。(特に$g_n(c)$ は $0$ に収束しない。)<br />
<br />
ゆえに、$g_n(x)$ は、一様に $0$ に収束します。$\Box$<br />
<br />
Diniの定理についての話は、<br />
2015年度の微積分II演習のページ(<a href="https://motochans.blogspot.com/2015/01/ii13_27.html">リンク</a>)<br />
にも書いていますのでご参照ください。<br />
<br />
<u><span style="font-size: large;">補題</span></u><br />
アルゼラの定理の中で出てきている等式をひとつ示しておきます。<br />
<b></b><br />
<b>補題</b><br />
<b>実数 $\xi,\eta$ に対して、</b><br />
<b>$\min\{\xi,\eta\}=\frac{1}{2}(\xi+\eta-|\xi-\eta|)$である。</b><br />
<b></b><br />
(証明)<br />
もし$\xi\ge \eta$ なら、<br />
$\frac{1}{2}(\xi+\eta-(\xi-\eta))=\eta$ となり<br />
もし$\xi\le \eta$ なら、<br />
$\frac{1}{2}(\xi+\eta-(-\xi+\eta))=\xi$ となり<br />
この等式が成り立つ。$\Box$<br />
<br />
同様に、<br />
$\max\{\xi,\eta\}=\frac{1}{2}(\xi+\eta+|\xi-\eta|)$ となります。<br />
<br />
とくに、$a(x)$ と $b(x)$ が連続関数であるとすると、<br />
$\min\{a(x),b(x)\}$ や $\min\{a(x),b(x)\}$ も連続関数である。$\Box$<br />
<br />
<br />
<span style="font-size: large;"><u>例</u></span><br />
一様収束しないが、一様有界で極限と積分の交換が成り立つ例をいくつか与えます。<br />
(といっても挙げた例はアルゼラの定理を使わなくても交換が成り立つことがわかる<br />
ものばかりでした。何か他に良い例があると思いますが、私は思いつきませんでした。)<br />
<br />
例1<br />
$[0,1]$ 閉区間において、<br />
$$f_n(x)=\begin{cases}2nx&0\le x\le \frac{1}{2n}\\-2nx+2&\frac{1}{2n}\le x\le\frac{1}{n}\\0&\frac{1}{n}\le x\le 1\end{cases}$$<br />
として定義しますと、この関数列は、恒等的に $0$ となる関数に<br />
一様収束しませんが、各点収束します。<br />
また、一様有界な関数列で、$[0,1]$ で連続な関数に収束しますので、<br />
積分と極限の順序交換が成り立ち、<br />
$$\lim_{n\to\infty}\int_0^1f_n(x)dx=0$$<br />
が成り立ちます。<br />
<br />
例2<br />
$[0,1]$ による関数として<br />
$f_n(x)=x^n$ とします。<br />
このとき、$f_n(x)$ は一様収束しませんが、一様有界で<br />
$(0,1)$ で連続な関数に収束するので、<br />
$$\lim_{n\to\infty}\int_0^1x^ndx=0$$<br />
となります。<br />
<br />
閉区間に不連続点が有限個あったととき、その関数の積分は<br />
その点を除いて積分してやることで計算したものになります。<br />
<br />
なので、例2の関数の各点収束先は、<br />
$$f(x)=\begin{cases}0&1\le x<1\\1&x=1\end{cases}$$<br />
ですが、この積分は、$f(x)\equiv 0$ の積分と等しくなります。<br />
<br />
ただ、この例だと、具体的に積分が計算できてしまい、<br />
$\int_0^1x^ndx=\frac{1}{n+1}$ なので、上のアルゼラの定理を<br />
使うまでもないということになります。<br />
積分が計算できるという点では例1もそうですが。<br />
<br />
例3<br />
$[0,\frac{\pi}{2}]$において、<br />
$f_n(x)=\sin^nx$ とすると、<br />
この場合も、一様収束せず、一様有界で、<br />
$[0,\frac{\pi}{2})$ で0で、$x=\frac{\pi}{2}$ のときに1<br />
となる関数に各点収束します。<br />
<br />
よって、この場合も<br />
$\lim_{n\to\infty}\int_0^{\frac{\pi}{2}}\sin^nxdx=0$<br />
となります。<br />
この積分は、部分積分を繰り返すことで<br />
$$\int_0^{\frac{\pi}{2}}\sin^nxdx=\begin{cases}\frac{(n-1)!!}{n!!}&n\text{奇数}\\\frac{(n-1)!!}{n!!}\frac{\pi}{2}&n\text{偶数}\end{cases}$$<br />
と計算できます。<br />
つまり、アルゼラの定理を使えば、偶奇のどちらにしても<br />
$\frac{(n-1)!!}{n!!}\to 0$がわかるということになります。</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-64473708629647113372020-06-09T01:10:00.003+09:002022-06-22T00:19:08.671+09:00数学リテラシー1(第10回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(水曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="text-align: left;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<br />
<br />
<div style="tab-stops: 271.6pt;">
今回の授業では、置換を用いて 行列式を定義しました。<span lang="EN-US"></span></div>
<br />
<div style="tab-stops: 271.6pt;">
<br /></div>
<div style="tab-stops: 271.6pt;">
<span style="font-size: large;"><u>符号</u></span></div>
<div style="tab-stops: 271.6pt;">
まず符号 </div>
<div style="tab-stops: 271.6pt;">
$$\text{sgn}:S_n\to \{1,-1\}$$</div>
<div style="tab-stops: 271.6pt;">
の定義をします。<span lang="EN-US"> </span></div>
<br />
<div style="tab-stops: 271.6pt;">
置換 $\sigma$ を <span lang="EN-US">$m$
</span>個の互換 $\tau_1\cdots, \tau_m$ の積で</div>
<div style="tab-stops: 271.6pt;">
$$\sigma=\tau_1\cdots \tau_m$$</div>
<div style="tab-stops: 271.6pt;">
と書けたとするとき、</div>
<br />
<div style="tab-stops: 271.6pt;">
<span lang="EN-US">$$\text{sgn}(\sigma)=(-1)^m$$</span></div>
<br />
<div style="tab-stops: 271.6pt;">
と定義します。</div>
<div style="tab-stops: 271.6pt;">
この数 1 or -1 を置換 $\sigma$ の<b>符号</b>といいます。<span lang="EN-US"></span></div>
<br />
<div style="tab-stops: 271.6pt;">
この定義は、ある置換の互換の積が偶数である場合は、$1$</div>
<div style="tab-stops: 271.6pt;">
または、奇数である場合は <span lang="EN-US">$-1$ </span>ということと同じです。<span lang="EN-US"></span></div>
<br />
<div style="tab-stops: 271.6pt;">
符号をきちんと定義するためには置換を互換の積で書いた時に、それが偶数個の</div>
<div style="tab-stops: 271.6pt;">
互換の積であるか、奇数個の互換の積であるかはその表示の仕方によらないことを<br />
示す必要があります。</div>
<br />
<br />
<div style="tab-stops: 271.6pt;">
例えば<span lang="EN-US">1</span>から<span lang="EN-US">3</span>までの数を並べ替える方法は6個になります そのうちで 以下の置換</div>
<div style="tab-stops: 271.6pt;">
$e$, $(1,2,3)$, $(1,3,2)$ は、必ず偶数個の互換の積に書くことができます。</div>
<div style="tab-stops: 271.6pt;">
<br /></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiLjEicQre5qFWqnfauOSxEsaLOrCke6LrkUWSx9upDSCRv5XyKeZ7_vGs3ju_8-hRRqrX8Cs60rgYFZws2vLBeBCbQJRl263vEAZ3C5iisKwfPukj4s8EIHu1xw_KLBc6NmmnOSQ1Or5Xu/s1600/fugo1.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="846" data-original-width="1517" height="178" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiLjEicQre5qFWqnfauOSxEsaLOrCke6LrkUWSx9upDSCRv5XyKeZ7_vGs3ju_8-hRRqrX8Cs60rgYFZws2vLBeBCbQJRl263vEAZ3C5iisKwfPukj4s8EIHu1xw_KLBc6NmmnOSQ1Or5Xu/s320/fugo1.png" width="320" /></a></div>
<div style="tab-stops: 271.6pt;">
置換をあみだくじ、もしくは、上のような数字を結ぶ線を描いたものと</div>
<div style="tab-stops: 271.6pt;">
考えることもできます。つまり、線には、1から3までの数が割り当てられている<br />
ことになります。</div>
<div style="tab-stops: 271.6pt;">
<br /></div>
<div style="tab-stops: 271.6pt;">
これらの置換の交わった点の個数が互換の個数ということになります。</div>
<div style="tab-stops: 271.6pt;">
<span lang="EN-US">たとえば、恒等置換の場合を考えます。</span></div>
<div style="tab-stops: 271.6pt;">
<span lang="EN-US">1の線分を描きます。その上に2の線を描いたときに、必ず、偶数個で交わります。</span></div>
<div style="tab-stops: 271.6pt;">
<span lang="EN-US">というのも、1,2は、上と下とで順番同じなので、必ず、1に交わったら、もう1回まじわらないと、この順番にたどり着くことができません。</span></div>
<div style="tab-stops: 271.6pt;">
<span lang="EN-US">また、3の線をかいたときも同じで、3の線が1の線と交わったら、もう一回</span></div>
<div style="tab-stops: 271.6pt;">
<span lang="EN-US">交わらないと、この順番でたどり着きません。</span></div>
<div style="tab-stops: 271.6pt;">
<span lang="EN-US">3の線が2の線と交わったときも同様です。</span></div>
<div style="tab-stops: 271.6pt;">
<span lang="EN-US">よって、このとき、交わった点の数は、必ず偶数個になります。</span></div>
<div style="tab-stops: 271.6pt;">
<span lang="EN-US">$(1,2,3)$ や $(1,3,2)$ の場合の置換も同じことを表します。</span><br />
(実は、このようにあみだを使って説明をしてもしても置換の全ての<br />
互換の積がいつでも偶数ということの証明にはならないのですが.....<br />
というのも、このように、あみだによって書かれる互換の積の分解は、<br />
<span lang="EN-US">隣り合った数の互換の積によってかけているだけです。)</span></div>
<div style="tab-stops: 271.6pt;">
<br /></div>
<br />
<div style="tab-stops: 271.6pt;">
おなじように、次の3つの 置換は 奇数個の互換の積によってかけます。<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhaOREjIYZGMdBMv9oGYAd1Tg_dYHCh9ZEThAC7B1SaOmJvh0GwPYOndbyfvPy1j2VC07Aj4cnyph05MbKVU1q_mLSoiZvFytCXqGOL57-07XgFlTjInIhxUqvCLqj2azBHF0pMvMpPB5zv/s1600/fugo2.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="820" data-original-width="1546" height="169" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhaOREjIYZGMdBMv9oGYAd1Tg_dYHCh9ZEThAC7B1SaOmJvh0GwPYOndbyfvPy1j2VC07Aj4cnyph05MbKVU1q_mLSoiZvFytCXqGOL57-07XgFlTjInIhxUqvCLqj2azBHF0pMvMpPB5zv/s320/fugo2.png" width="320" /></a></div>
</div>
<div style="tab-stops: 271.6pt;">
<span lang="EN-US"><br /></span></div>
<div style="tab-stops: 271.6pt;">
<br /></div>
<div style="tab-stops: 271.6pt;">
<u><span style="font-size: large;">行列式の定義</span></u><span lang="EN-US"></span></div>
<u></u><span style="font-size: large;"></span><br />
<div style="tab-stops: 271.6pt;">
では、符号を使って行列式を定義しましょう。</div>
<div style="tab-stops: 271.6pt;">
$(n,n)$ 行列 $A=(a_{ij})$ の行列式を以下のように定義します<br />
<br />
$$\det(A)=\sum_{\sigma\in S_n}\text{sgn}(\sigma)a_{1\sigma(1)}a_{2\sigma(2)}\cdots a_{n\sigma(n)}$$<br />
<br />
ここで、Σの記号の下に$\sigma\in S_n$ が書かれているのは、<br />
$S_n$ の元に従って足し合わせるということです。<br />
ですので、特に、$n!$ 個の和であるということになります。<br />
<br />
例えば<span lang="EN-US">$(2,2)$ </span>行列の場合に行列式の定義に従って計算します。<br />
<br />
$S_2=\{e,(1,2)\}$ですから、<br />
以下のようになります。<span lang="EN-US"></span></div>
$$\det(A)=a_{11}a_{22}-a_{12}a_{21}$$<br />
ということになります。<br />
<div style="tab-stops: 271.6pt;">
$A=\begin{pmatrix}a&b\\c&d\end{pmatrix}$ の場合、$\det(A)=ad-bc$<br />
と定義したわけですから確からこの定義と合っています。<span lang="EN-US"></span></div>
<br />
<div style="tab-stops: 271.6pt;">
次に、<span lang="EN-US">$(3,3)$ </span>行列の行列式を計算をします。<span lang="EN-US"></span></div>
$$A=\begin{pmatrix}a_{11}&a_{12}&a_{13}\\a_{21}&a_{22}&a_{23}\\a_{31}&a_{32}&a_{33}\end{pmatrix}$$<br />
として、6個の和を書き下してみると、以下のようになります。<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
この項の順番は、上に登場した置換の順番に沿ってかきました。<br />
<br />
$$\det(A)=a_{11}a_{22}a_{33}+a_{12}a_{23}a_{31}+a_{13}a_{21}a_{32}-a_{12}a_{21}a_{33}-a_{11}a_{23}a_{32}-a_{13}a_{22}a_{31}$$<br />
<br />
<div style="tab-stops: 271.6pt;">
この公式を覚えるのに役に立つのが、<b>サラス</b>の公式です。<br />
$(3,3)$ 行列に、下のような6本の線をかいてやります。<br />
そのとき、その線にぶつかった成分をかけて符号をつけて足し上げることによって<br />
行列式が計算できます。<br />
ただし、黒いほうを正の符号とし、赤いほうを負の符号とするのです。<br />
<span lang="EN-US"></span></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi9hP-sCQRAIxVYUIDOrhD0q5l89_8GPv1J0vBqYHFCu1EsUvmiJONuirHrLgnwW5FSOJJ90wCnn-fpDDMiJKKn9ggZv5JsPCXhQ4xFogjCklENbqWa3GhC4QRzZ5uU_ukmlW2XSoSDvIoa/s1600/sarasu.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="732" data-original-width="909" height="257" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi9hP-sCQRAIxVYUIDOrhD0q5l89_8GPv1J0vBqYHFCu1EsUvmiJONuirHrLgnwW5FSOJJ90wCnn-fpDDMiJKKn9ggZv5JsPCXhQ4xFogjCklENbqWa3GhC4QRzZ5uU_ukmlW2XSoSDvIoa/s320/sarasu.png" width="320" /></a></div>
<br />
<br />
<div style="tab-stops: 271.6pt;">
<u><span style="font-size: large;">転置行列と行列式</span></u></div>
<u></u><span style="font-size: large;"></span><br />
<div style="tab-stops: 271.6pt;">
転置行列の行列式は元の行列の行列式と同じになります。<span lang="EN-US"></span></div>
$\sigma$に対して、転置行列の行列式は、<br />
<div style="tab-stops: 271.6pt;">
<br />
$$\det(^tA)=\sum_{\sigma\in S_n}\text{sgn}(\sigma)a_{\sigma(1)1}\cdots a_{\sigma(n)n}$$<br />
ですが、$\sigma$に対して、$\sigma^{-1}$ はその符号を変えず、<br />
これは、積の順番を入れ替えれば、<br />
$$=\sum_{\sigma\in S_n}\text{sng}(\sigma)a_{1\sigma^{-1}(1)}\cdots a_{n\sigma^{-1}(n)}$$<br />
となりますが、$\sigma\mapsto \sigma^{-1}$ は一対一ですから、<br />
$\sigma^{-1}$ を再び $\sigma$ に置きなおすことで、<br />
$$=\sum_{\sigma\in S_n}\text{sgn}(\sigma)a_{1\sigma(1)}a_{2\sigma(2)}\cdots a_{n\sigma(n)}$$<br />
となるので、これは$\det(A)$ となります。<br />
よって、転置行列の行列式はもとの行列の行列式と同じになります。</div>
<br />
<div style="tab-stops: 271.6pt;">
<u><span style="font-size: large;">行列式の交代性</span></u><span lang="EN-US"></span></div>
<u></u><span style="font-size: large;"></span><br />
<div style="tab-stops: 271.6pt;">
2つの行を交換して得られる行列と、2つの列を交換して得られる行列の行列式は</div>
<div style="tab-stops: 271.6pt;">
ちょうど元の行列式の <span lang="EN-US">$-1$ </span>倍になります。<span lang="EN-US"></span></div>
<br />
<div style="tab-stops: 271.6pt;">
例えば、$i$ 列と $j$ 列を入れ替えた行列を$A'$ とする。$i<j$<br />
であると仮定します。<br />
$$\det(A')=\sum_{\sigma\in S_n}\text{sgn}(\sigma)a_{1\sigma(1)}\cdots a_{j\sigma(i)}\cdots a_{i\sigma(j)}\cdots a_{n\sigma(n)}$$<br />
ここで、$\tau=(i,j)$とすると、<br />
$$=\sum_{\sigma\in S_n}\text{sgn}(\sigma\tau)\text{sgn}(\tau)a_{1\sigma\tau(1)}\cdots a_{i\sigma\tau(i)}\cdots a_{j\sigma\tau(j)}\cdots a_{n\sigma\tau(n)}$$<br />
$$=\sum_{\sigma\in S_n}\text{sng}(\tau)\text{sgn}(\sigma)a_{1\sigma(1)}\cdots a_{i\sigma(i)}\cdots a_{j\sigma(j)}\cdots a_{n\sigma(n)}$$<br />
$$=-\sum_{\sigma\in S_n}\text{sng}(\sigma)a_{1\sigma(1)}\cdots a_{n\sigma(n)}=-\det(A)$$<br />
<br />
2つの行を交換して得られる行列式が$-1$になることも、<br />
上の転置をとることによって行列式が変わらないことから<br />
従います。<br />
<br /></div>
この交代性から、<br />
$A$の第 $i$ 列 ${\bf a}_i$ と第 $j$ 列 ${\bf a}_j$ が、${\bf a}_i={\bf a}_j$ であるとするとき、その第 $i$ 列と $j$ を交換した行列を $A'$ とすると、$A$ と $A'$ は同じ行列であるから、<br />
$\det(A)=-\det(A)$ である。<br />
よって、$\det(A)=0$ となります。<br />
<br />
<br />
<br />
<div style="tab-stops: 271.6pt;">
<u><span style="font-size: large;">多重線形性</span></u><br />
$n$ 個の $n$ 次元ユークリッド空間の元のペアから実数への写像</div>
<div style="tab-stops: 271.6pt;">
$$f:{\mathbb R}^n\times \cdots \times {\mathbb R}^n\to {\mathbb R}$$<br />
が、任意の$i$ 個目のベクトルに対して、<br />
$$f({\bf v_1},\cdots,{\bf v}_i+{\bf v}_i',\cdots, {\bf v}_n)=f({\bf v}_1,\cdots, {\bf v}_i,\cdots, {\bf v}_n)+f({\bf v}_1,\cdots, {\bf v}_i',\cdots, {\bf v}_n)$$<br />
かつ、実数 $c$ に対して<br />
<br />
$$f({\bf v}_1,\cdots, c{\bf v}_i,\cdots {\bf v}_n)=cf({\bf v}_1,\cdots, {\bf v}_i,\cdots, {\bf v}_n)$$<br />
を満たすとき、$f$ は<b>多重線形性を満たす</b>といいます。<br />
<br />
$\det$ は多重線形写像であることを示します。<br />
$A''$ の第 $i$ 列が $\begin{pmatrix}a_{1i}+a_{1i}'\\a_{2i}+a_{2i}\\\vdots \\a_{ni}+a_{ni}'\end{pmatrix}$<br />
であり、他の第 $(k,j)$ 成分は、$a_{kj}$ であるとします。<br />
また、$A=(a_{ij})$とし、$A'$ の第$i$列は、<br />
$\begin{pmatrix}a_{1i}'\\a_{2i}'\\\vdots \\a_{ni}'\end{pmatrix}$<br />
であるとします。<br />
このとき、<br />
<br />
$$\det(A'')=\sum_{\sigma\in S_n}\text{sgn}(\sigma)a_{1\sigma(1)}\cdots (a_{i\sigma(i)}+a_{i\sigma(i)}')\cdots a_{n\sigma(n)}$$<br />
$$=\sum_{\sigma\in S_n}\text{sgn}(\sigma)a_{1\sigma(1)}\cdots a_{i\sigma(i)}\cdots a_{n\sigma(n)}$$<br />
$$+\sum_{\sigma\in S_n}\text{sng}(\sigma)a_{1\sigma(1)}\cdots a_{i\sigma(i)}'\cdots a_{n\sigma(n)}$$<br />
$$=\det(A)+\det(A')$$<br />
となり、実数倍の方も同様にやることにより、$\det$ の多重線形性が満たされました。<br />
<br />
$A({\bf x})$ を第 $i$ 列が ${\bf x}$ となる行列であるとすると、<br />
$A$ の第 $k$ 列を ${\bf a}_k$ であるとする。<br />
このとき、多重線形性を用いると、<br />
$$\det(A({\bf a}_i+c{\bf a}_j))=\det(A({\bf a}_i))+\det(A(c{\bf a}_j))$$<br />
$$=\det(A({\bf a}_i)+c\det(A({\bf a}_j))=\det(A({\bf a}_i))=\det(A)$$</div>
<div style="tab-stops: 271.6pt;">
<br /></div>
<div style="tab-stops: 271.6pt;">
<u>まとめ</u></div>
<u></u><br />
<div style="tab-stops: 271.6pt;">
行列式を写像 <span lang="EN-US">$\det:{\mathbb
R}^n\times \cdots \times {\mathbb R}^n\to {\mathbb R}$</span>と思うと、交代的な多重線形性を持つ写像。<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: "Times New Roman"; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
行列 $A$ の第 $j$ 列を他の第 $i$ 列に定数倍をして足してやったものは、</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: "Times New Roman"; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
元の行列式を変えない。<br />
<br />
また、単位行列 $E$ に対して、<br />
$\det(E)=1$</div>
であることも定義から従います。<br />
<br />
<br /></div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike></div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-76584862899055481962020-06-02T09:44:00.003+09:002022-04-19T10:13:16.600+09:00数学リテラシー1(第9回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(火曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="text-align: left;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
今回と次回で、行列式を定義します。<br />
最終的に、次回において行列式を定義しますが、今回は</div>
行列式を定義するために用いる置換について解説しました。<br />
行列式は、$(2,2)$ 行列のときに $\begin{pmatrix}a&b\\c&d\end{pmatrix}$<br />
のときに $ad-bc$ として定義していますが、<br />
ここでは、一般の $(n,n)$ 行列のときの行列式を定義します。<br />
<br />
<u><span style="font-size: large;">$n$ 文字の集合の間の全単射の集合と置換</span></u><br />
まず、<br />
$\Omega$ を $n$ 個の数字からなる集合 $\{1,2,\cdots, n\}$ とし、<br />
$S_n=\{\sigma:\Omega\to \Omega|\sigma:\text{全単射}\}$<br />
とおきます。<br />
つまり、$\sigma\in S_n$ は $\Omega$ 上の写像で、全単射なものということです。<br />
<br />
ここで全単射というのは、全射かつ単射な写像であったこと思い出してください。<br />
つまり 写像で、逆写像が存在するようなもののことでした。<span lang="EN-US"></span><br />
<br />
この集合は<span lang="EN-US">$n$ </span>個の数字をちょうど入れ替えており、<br />
つまり<span lang="EN-US">$n$ </span>個の数字を並べ替える方法全体と一致しています。<br />
<br />
この並び替え、つまり、置換が、$S_n$ の元ということです。<br />
つまり、<span lang="EN-US">$S_n$ </span>は <span lang="EN-US">$n$ </span>個の文字の並び替えの全体の集合といってもよいわけです。<br />
<br />
よって $S_n$ に含まれる要素は $n<span lang="EN-US">$ 個の数字の</span>順列の数と同じということで、<br />
$S_n$ には、$n!$ 個の元が</div>
<div style="color: black; font-family: "times new roman"; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
含まれることになります。</div>
よって、$S_n$ の元のことをこれから単に<b>置換</b>ということにします。<br />
<span style="font-size: large;"></span><br />
<u><span style="font-size: large;">置換の書き方</span></u><br />
置換は、$(2,n)$ 行列を用いて、<br />
<br />
$$\begin{pmatrix}1&2&\cdots&n\\\sigma(1)&\sigma(2)&\cdots&\sigma(n)\end{pmatrix}$$<br />
として書き表すことが一般的です。<br />
つまり、置換を写像として考えたとき、縦ベクトルとして、<br />
$i$ とその行先(像)の関係を書いたものということになります。<br />
ですので、
<br />
この行列のある縦ベクトルを他の縦ベクトルと入れ替えてできる<br />
$(2,n)$ 行列が表す置換も同じ置換を表すことになります。<br />
<br />
例えば、$n=3$ の場合であれば、$S_3$ の元は<br />
$$\begin{pmatrix}1&2&3\\1&2&3\end{pmatrix},\begin{pmatrix}1&2&3\\2&1&3\end{pmatrix},\begin{pmatrix}1&2&3\\1&3&2\end{pmatrix}$$<br />
$$\begin{pmatrix}1&2&3\\3&2&1\end{pmatrix},\begin{pmatrix}1&2&3\\2&3&1\end{pmatrix},\begin{pmatrix}1&2&3\\3&1&2\end{pmatrix}$$
</div>
<br />
の3!=6通りとなります。<br />
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
また、<span lang="EN-US">$n$ </span>個の数字を並び替える集合とは、次のように<br />
<span lang="EN-US">$n$ </span>個の数字をあみだくじによって入れ替えを行ったものと考えることもできます。<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhSPLVsxz5yxEbCwKHBNP4mBem_ESIc866RWIyoFWqGjspM6F8jFWcqL2bc9knkkfjvH5_DneG1NVs1p90UQbtbyYS8uqSrXj8uW8Jyk3zTHD1LYcCWRxyASPp56LsUiuC51N4bFBht-OkH/s1600/amida.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="924" data-original-width="1058" height="348" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhSPLVsxz5yxEbCwKHBNP4mBem_ESIc866RWIyoFWqGjspM6F8jFWcqL2bc9knkkfjvH5_DneG1NVs1p90UQbtbyYS8uqSrXj8uW8Jyk3zTHD1LYcCWRxyASPp56LsUiuC51N4bFBht-OkH/s400/amida.png" width="400" /></a></div>
これは、上の $S_3$ の置換をあみだくじで書いたものです。<br />
<br />
最も左上のあみだくじに対応するものを
<b>恒等置換</b>といいます。恒等置換というのは何も置き換えない置換のことをいいます。
写像でいえば、恒等写像に対応します。
<span lang="EN-US"></span><br />
<br />
ただし、置換に対して、あみだくじで書く方法はさまざまあります。<br />
これは考えればすぐにわかることですが、たとえば、<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiYgxEJ5Y5e_p2HAZIihDPMuW_Qkm5MH3ZvAdJDQwIY95IQQBzm3v8eOkUuBABzBziq9DIB6gSbiy44FsqJn8mXm4QE-GzD8Ust5fBajBu8plK3K6zzomgGvKViVZYEsW_2UenmLubugESx/s1600/jimei.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="220" data-original-width="210" height="200" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiYgxEJ5Y5e_p2HAZIihDPMuW_Qkm5MH3ZvAdJDQwIY95IQQBzm3v8eOkUuBABzBziq9DIB6gSbiy44FsqJn8mXm4QE-GzD8Ust5fBajBu8plK3K6zzomgGvKViVZYEsW_2UenmLubugESx/s200/jimei.png" width="190" /></a></div>
<div class="separator" style="clear: both; text-align: center;">
</div>
のようなあみだを考えれば、このあみだが恒等置換を表すことがわかるでしょう。<br />
<span style="font-size: large;"></span><br />
<span style="font-size: large;"><u>置換の積</u></span><br />
置換には、積を定義することができます。
<br />
あみだくじで言えばあみだくじを縦に並べたもの、<br />
写像の言葉で言えば写像の合成のことです。<br />
例えば、<br />
$$\sigma=\begin{pmatrix}1&2&3\\2&1&3\end{pmatrix}$$<br />
と<br />
$$\tau=\begin{pmatrix}1&2&3\\3&2&1\end{pmatrix}$$
の積を考えると、<br />
$$\begin{pmatrix}1&2&3\\2&3&1\end{pmatrix}$$<br />
のようになります。できた置換は、$\tau\sigma$ <br />
<div class="separator" style="clear: both; text-align: center;">
</div>
と書きます。<br />
この順番は、写像の合成の意味が込められており、この順番に<br />
なっています。<br />
<br />
このことをあみだの言葉で言えば、<br />
上に $\tau$、下に $\sigma$ を並べてやることで、<br />
<div class="separator" style="clear: both; text-align: center;">
</div>
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiwjJJJhq1k4L29GDC-WSDN082vKbIZ9u4IJJRbSl_YoKIdATsKO5WU5KsAnoFK78up-AkO5VQRJeJL4exhtuvI2uvGkmkYDE5yTh1Ac9FQh8Qr7V0IF31vSNDJtr0IX4h5b-48McQxjmZy/s1600/amida1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="625" data-original-width="1058" height="189" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiwjJJJhq1k4L29GDC-WSDN082vKbIZ9u4IJJRbSl_YoKIdATsKO5WU5KsAnoFK78up-AkO5VQRJeJL4exhtuvI2uvGkmkYDE5yTh1Ac9FQh8Qr7V0IF31vSNDJtr0IX4h5b-48McQxjmZy/s320/amida1.png" width="320" /></a></div>
のようになります。このように絵を使ってこの計算をすることができます。<br />
なんとなく、$\tau$ を下に置くのかと勘違いするかもしれませんが、<br />
逆ですので注意してください。<br />
<br />
<u><span style="font-size: large;">逆置換</span></u><br />
<b>逆置換</b>というのは写像の逆写像を取ったものつまり あみだくじの鏡像をとったものとして定義できます。<br />
<br />
鏡像というのは水平な線を鏡として得られるようなあみだくじのことです。<br />
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
例えば、<br />
$$\begin{pmatrix}1&2&3\\3&1&2\end{pmatrix}$$<br />
<br />
の場合ですと、この置換の上と下を入れ替えて、<br />
$$\begin{pmatrix}3&1&2\\1&2&3\end{pmatrix}$$<br />
となり、また、上の行を1,2,3の順にあるように、縦ベクトルを<br />
入れ替えると、<br />
$$\begin{pmatrix}1&2&3\\2&3&1\end{pmatrix}$$<br />
のようになります。<br />
<br />
置換の性質をここで おさらいしておきます。<br />
<ol style="text-align: left;">
<li>2つの置換の積を取ることができる。</li>
<li>任意の置換に対して逆置換が必ず存在すること。</li>
<li>恒等置換が存在すること。</li>
</ol>
ということになります。 <span lang="EN-US"></span><br />
実は、この性質は、数学における<a href="https://ja.wikipedia.org/wiki/群論">群論</a>の初歩をやっていることに相当します。<br />
群論というのは、大学では3年生くらいで習う、何かの変換全体の集合を<br />
考える理論です。<br />
置換全体は群の例の一つですが、<br />
<br />
群論においては置換全体の集合は $n$ 次対称群と呼ばれます。<br />
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
数学全般はもちろんのこと、物理や化学でも群論の考えかたを応用した<br />
理論は多いです。<br />
ちなみに、群論を初めて応用したのは、<a href="https://ja.wikipedia.org/wiki/エヴァリスト・ガロア">エヴァリスト・ガロア</a>(19世紀の数学者)<br />
で、群論を用いて5次以上の方程式の解に、<br />
その係数の四則演算べき乗根で解けないものが存在することを証明しました。<br />
<br />
<u><span style="font-size: large;">巡回置換と互換</span></u><span lang="EN-US"></span><br />
<u></u><span style="font-size: large;"></span><br />
次に置換 $\sigma\in S_n$ を<br />
$\{i_1,\cdots, i_r\}\subset \Omega$ に対して、<br />
$$i_1\mapsto i_2,\ i_2\mapsto i_3\cdots, i_{r}\mapsto i_1$$<br />
と写し、それ以外では数を変えないものを<b>巡回置換</b>といいます。<br />
<br />
そのような巡回置換を $(i_1,i_2,\cdots, i_r)$ のように書きます。<br />
この巡回の長さ $r$ を巡回置換の<b>長さ</b>といいます。<span lang="EN-US"></span><br />
<br />
特に長さが<span lang="EN-US">2</span>の巡回置換のことを<b>互換</b>といいます。 <span lang="EN-US"></span><br />
任意の置換 $\sigma\in S_n$ に対して、任意の $i\in \Omega$ に対して、その行先を<br />
順次たどっていけば、必ずもとの $i$ に戻ってくる巡回置換をなしています。<br />
<br />
よって、すべての置換は 互いに交わらない巡回置換の積に書くことができます。<span lang="EN-US"></span><br />
<br />
また巡回置換は いくつかの互換を使って<span lang="EN-US"></span><br />
$$(i_1,\cdots, i_r)=(i_1,i_{r})\cdots (i_1,i_3)(i_1,i_2)$$<br />
のように書くことができます。<br />
他の書き方として、<br />
$$(i_1,\cdots, i_r)=(i_1,i_{2})(i_2,i_3)\cdots (i_{r-1},i_r)$$<br />
もあります。<br />
<br />
これらのことから、全ての置換はいくつかの互換の積によって書き表すことができる<br />
ことになります。<br />
<br />
これは、あみだくじを考えれば明らかで、任意の置換はあみだくじですから、<span lang="EN-US"></span><br />
<br />
置換をあみだくじによって描いたとき、線を少々ずらすことで、交わりを全て二重点だけにしておくことができます。<span lang="EN-US"></span><br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
各交差点において、数字が互換のようにして<br />
入れ替わり、その合成によって全ての置換を互換の</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
積に書くことができるということになります。</div>
(このとき、全ての線は上から下に、途中で止まらずに流れていくものとします。)<span lang="EN-US"></span><br />
<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjwydgA4frQiiN8_5n3U6FBgRqLNGO7iKaraqJdumksEIEh-7lpt4qpRZj1OzSxoCL0-bFD3TKyl9OhLcHSaUIogRDu5xxxBLzeXqtRVJoQxvUpgYtqdt6neqctGzYb-c1gmcnYBbX7fmO8/s1600/amidadukuri.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="275" data-original-width="828" height="106" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjwydgA4frQiiN8_5n3U6FBgRqLNGO7iKaraqJdumksEIEh-7lpt4qpRZj1OzSxoCL0-bFD3TKyl9OhLcHSaUIogRDu5xxxBLzeXqtRVJoQxvUpgYtqdt6neqctGzYb-c1gmcnYBbX7fmO8/s320/amidadukuri.png" width="320" /></a></div>
<br />
この横線の数が表示した互換の数に相当します。<br />
<br />
例として教科書やスライドにあった次の置換を用いて考えます<span lang="EN-US"></span><br />
<br />
$$\begin{pmatrix}1&2&3&4&5&6&7\\4&5&6&7&2&3&1\end{pmatrix}$$<br />
個々の数字をたどっていくことで、<br />
この置換は長さが3の巡回置換 $(1,4,7)$ と長さが2の2つの巡回置換 $(2,5)$ と<br />
$(,6)$ の積に書かれることがわかるでしょう。<br />
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-29296839201706653862020-05-28T00:57:00.003+09:002022-04-19T10:13:46.931+09:00数学リテラシー1(第8回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(土曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="text-align: left;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
第6回では、直交行列によって対角化できる行列は<br />
対称行列であることを証明しましたが、<br />
その逆が成り立つことを今回は示します。<br />
<br />
つまり、<br />
<br />
定理<br />
実対称行列は固有値は実数であり、<br />
さらに、ある直交行列によって対角化できる。<br />
<br />
です。<br />
今回も、$(2,2)$ 行列しか扱いません。<br />
<br />
まず、最初の主張は、$\begin{pmatrix}a&b\\b&c\end{pmatrix}$<br />
の固有多項式を求めると、<br />
$t^2-(a+d)t+ad-b^2$ となり、<br />
この判別式を求めると、<br />
$D=(a+d)^2-4(ad-b^2)=(a-d)^2+4b^2$ となり<br />
この値は正の数または0です。<br />
よって、固有多項式は、2つの実数を解に持つということになります。<br />
(重解を含む。)<br />
もし、重解をもつとすると、$a=d$ かつ $b=0$ であるから、<br />
そのような行列は、スカラー行列 $aE$ ということになります。<br />
ここで、$a$ はある実数で、$aE$ は行列 $\begin{pmatrix}a&0\\0&a\end{pmatrix}$<br />
のことです。<br />
<br />
実対称行列の実固有値は $\lambda_1,\lambda_2$ の2つ存在します。<br />
ただし、$\lambda_1=\lambda_2$ の場合も存在します。<br />
そのうちの一つの固有値を $\lambda_1$ とします。<br />
その固有ベクトルは存在し、${\bf v}_1$ とします。<br />
ここで、${\bf v}_1$ は長さが1であるとしておきます。<br />
<br />
このとき、$A{\bf v}_1=\lambda_1{\bf v}_1$ となります。<br />
また、${\bf v}_1$ に直交するベクトルを ${\bf v}_2$ とします。<br />
${\bf v}_2$ も長さが1のベクトルであるとしておきます。<br />
<br />
このとき、${\bf v}_1,{\bf v}_2$ は、正規直交基底とよばれ、それを<br />
並べてできる $(2,2)$ 行列 $R=({\bf v}_1{\bf v}_2)$ は<br />
直交行列になります。つまり、$^tRR=E$ を満たします。<br />
<br />
(ここで注意として、固有ベクトルを取るときには、<br />
長さを1にしたり、直交ベクトルを取る必要はありません。<br />
この話では、直交行列によって行列を対角化するために<br />
このような操作をしています。)<br />
<br />
このとき、<br />
<br />
$A{\bf v}_2=p{\bf v}_1+q{\bf v}_2$ のように、${\bf v}_1$ と ${\bf v}_2$ の<br />
一次結合で表されます。ここで、$p,q$ は、何か実数です。<br />
<br />
そうすると、$A({\bf v}_1{\bf v}_2)=({\bf v}_1{\bf v}_2)\begin{pmatrix}\lambda_1&p\\0&q\end{pmatrix}$<br />
のようにあらわされます。<br />
<br />
よって、<br />
<br />
<br />
このとき、$R=({\bf v}_1{\bf v}_2)$ とすると、<br />
$AR=R\begin{pmatrix}\lambda_1&p\\0&q\end{pmatrix}=RS$<br />
<br />
とします。<br />
このとき、$A=RSR^{-1}$ となり、<br />
$A$ が対称行列であることから、全体に転置行列を取ることによって、<br />
$A={}^tA={}^t(RSR^{-1})=R{}^tSR^{-1}$<br />
となり、$AR=R{}^tS=RS$<br />
ですから、$R$ を左からかけて、$S={}^tS$ となります。<br />
つまり、$S$ は対称行列にならなければならないから、$p=0$ ということです。<br />
つまり、$A{\bf v}_2=q{\bf v}_2$ であることから、<br />
${\bf v}_2$ は固有ベクトルであり、$q=\lambda_1$ であるなら<br />
${\bf v}_2$ は$\lambda_1$ の固有ベクトルであり、<br />
$q\neq \lambda_2$ であるなら、${\bf v}_2$ は相異なる固有値を<br />
もち、${\bf v}_2$ はその固有ベクトルということになります。<br />
<br />
ゆえに、実対称行列 $A$ は、直交行列 $R$ を用いて、<br />
$R^{-1}AR$ を対角行列にすることができます。<br />
<br />
例1<br />
$A=\begin{pmatrix}1&-3\\-3&1\end{pmatrix}$ とすると、<br />
$A$ は実対称行列であることから、直交行列によって対角化されます。<br />
<br />
固有値を計算すると、$-2,4$ であり、<br />
それぞれのこゆうベクトルを求めます。<br />
$-2E-A=\begin{pmatrix}-3&3\\3&-3\end{pmatrix}$<br />
ですから、連立一次方程式 $(-2E-A){\bf x}={\bf 0}$ の解として、<br />
$\begin{pmatrix}1\\1\end{pmatrix}$<br />
を選べます。<br />
また、<br />
$4E-A=\begin{pmatrix}3&3\\3&3\end{pmatrix}$<br />
ですから、連立一次方程式 $(4E-A){\bf x}={\bf 0}$ の解として<br />
$\begin{pmatrix}1\\-1\end{pmatrix}$<br />
を選べます。<br />
このとき、$\tilde{{\bf v}}_1=\begin{pmatrix}1\\1\end{pmatrix}$ とし、<br />
$\tilde{\bf v}_2=\begin{pmatrix}1\\-1\end{pmatrix}$ とすると、<br />
それらは直交していますね。<br />
さらにそれから直交行列を作る場合、長さでわって<br />
${\bf v}_1=\frac{1}{\sqrt{2}}\begin{pmatrix}1\\1\end{pmatrix}$<br />
${\bf v}_2=\frac{1}{\sqrt{2}}\begin{pmatrix}1\\-1\end{pmatrix}$<br />
とすることで、直交行列 $P=({\bf v}_1{\bf v}_2)$ を得る。<br />
<br />
<br />
例2<br />
ここで、簡単な例で、実対称行列で対角化可能だが、<br />
直交行列でなくても対角化できることを見てみます。<br />
$A$ として単位行列 $E$ を考えれば、<br />
任意の逆行列をもつ行列 $P$ に対して、$P^{-1}AP=E$ ですから<br />
対角化ができています。一般に、逆行列をもつ $P$ といっても<br />
直交行列とは限りません。<br />
<br />
次の例をみましょう。<br />
<br />
例3<br />
<br />
$\begin{pmatrix}3&-1\\2&0\end{pmatrix}$<br />
この行列は、実対称行列ではないので、直交行列によって対角化はできませんが、<br />
対角化は可能です。<br />
この固有値は、$1,2$ であり、それぞれの固有ベクトルは、<br />
$E-A=\begin{pmatrix}-2&1\\-2&1\end{pmatrix}$ ですから、<br />
連立方程式<br />
$(E-A){\bf v}={\bf 0}$ の解として、<br />
${\bf v}_1=\begin{pmatrix}1\\2\end{pmatrix}$が選べます。<br />
<br />
また、<br />
$2E-A=\begin{pmatrix}-1&1\\-2&2\end{pmatrix}$ ですから、<br />
連立一次方程式 $(2E-A){\bf x}={\bf 0}$ の解として、<br />
${\bf v}_2=\begin{pmatrix}1\\1\end{pmatrix}$ をとることができる。<br />
<br />
たしかに${\bf v}_1$ と ${\bf v}_2$ は直交しません。<br />
この行列 $A$ は、 $P=\begin{pmatrix}1&1\\2&1\end{pmatrix}$ によって<br />
対角化することはできます。<br />
<br />
例4<br />
次に、連立漸化式から数列の一般項を出す方法を考えます。<br />
<br />
$$x_1=3,y_1=1$$<br />
$$\begin{cases}x_{n+1}=4x_n+10y_n\\y_{n+1}=-3x_n-7y_n\end{cases}$$<br />
このとき、この漸化式を以下のように行列を用いて考えることができます。<br />
$$\begin{pmatrix}x_{n+1}\\y_{n+1}\end{pmatrix}=\begin{pmatrix}4&10\\-3&-7\end{pmatrix}\begin{pmatrix}x_n\\y_y\end{pmatrix}$$<br />
ここにでてきた $(2,2)$ 行列を $A=\begin{pmatrix}4&10\\-3&-7\end{pmatrix}$<br />
とすると、この式を $\begin{pmatrix}x_{n+1}\\y_{n+1}\end{pmatrix}=A\begin{pmatrix}x_n\\y_n\end{pmatrix}$ とすることができて、この式を繰り返し使うことで、<br />
$$\begin{pmatrix}x_n\\y_n\end{pmatrix}=A^{n-1}\begin{pmatrix}x_1\\y_1\end{pmatrix}$$<br />
をえることができます。<br />
<br />
ここで、$A^n$ の求め方は、前回やりましたから、ここで応用できますね。<br />
実際、この行列の固有値は、$-1,-2$ ですから、<br />
固有ベクトルを求めると、<br />
$-E-A=\begin{pmatrix}-5&-10\\3&6\end{pmatrix}$<br />
となりますから、この固有ベクトルは、$\begin{pmatrix}2\\-1\end{pmatrix}$<br />
であり、<br />
$-2E-A=\begin{pmatrix}-6&-10\\3&5\end{pmatrix}$<br />
となりますから、この固有ベクトルは、$\begin{pmatrix}5\\-3\end{pmatrix}$<br />
となります。<br />
よって、$P=\begin{pmatrix}2&5\\-1&-3\end{pmatrix}$ とすると、<br />
$P^{-1}AP=\begin{pmatrix}-1&0\\0&-2\end{pmatrix}$ となり、<br />
$$A^n=P\begin{pmatrix}(-1)^n&0\\0&(-2)^n\end{pmatrix}P^{-1}=\begin{pmatrix}6(-1)^n-5(-2)^n&10(-1)^n+5(-2)^{n+1}\\3(-1)^{n+1}+3(-2)^n&5(-1)^{n+1}-3(-2)^{n+1}\end{pmatrix}$$<br />
となります。<br />
よって、<br />
$$\begin{pmatrix}x_n\\y_n\end{pmatrix}=\begin{pmatrix}6(-1)^{n-1}-5(-2)^{n-1}&10(-1)^{n-1}+5(-2)^{n}\\3(-1)^{n}+3(-2)^{n-1}&5(-1)^{n}-3(-2)^{n}\end{pmatrix}\begin{pmatrix}3\\1\end{pmatrix}$$<br />
$$=\begin{pmatrix}28(-1)^{n-1}-25(-2)^{n-1}\\14(-1)^n+15(-2)^{n-1}\end{pmatrix}$$<br />
のように求めることができます。<br />
<br />
<br />
<u><span style="font-size: large;">まとめと対角化可能性について</span></u><br />
実対称行列であることは、</div>
行列が対角化可能であるための十分条件を<br />
与えます。(さらに、直交行列によっても対角化できますが...)<br />
他に、対角化可能であるためのわかりやすい条件として、<br />
固有値が $n$ 個ある(今の場合、$2$個ある場合です。<br />
<br />
つまり、$(2,2)$ 行列の固有値がちょうど2個ある場合、対角化可能です。<br />
どうしてかというと、固有値に対して必ず、固有ベクトルが存在するので、<br />
この場合、2個の固有ベクトルが存在します。<br />
もちろんそれらは平行ではありません。もし平行なら、同じ固有値を持つはずです。<br />
<br />
よって、固有ベクトルで作られる行列 $P$ は逆行列をもち、<br />
$P^{-1}AP$ は対角行列になるからです。<br />
<br />
よって、$(2,2)$ 行列で対角化されるかどうかわからないのは、固有多項式が重解をもつ場合、つまり、固有多項式が $(t-\lambda)^2$ の形にかける場合ということになります。<br />
<br />
もちろん固有多項式が重解だからといって、行列が対角化可能である場合も存在します。<br />
極端な例として単位行列 $E$ を考えてみてください。<br />
この行列の行列式は、$(t-1)^2$ です。<br />
ですから、固有多項式だけみて、最終的に対角化可能であるかどかわかるのは、<br />
それが重解をもたない場合のみです。<br />
<br />
しかし、$(2,2)$ 行列で、固有多項式が重解をもち、対角化可能である場合は、<br />
その行列がスカラー行列である場合に限られます。<br />
ですので、<br />
この場合、対角化可能ではない場合というのは、固有多項式が重解をもち、<br />
スカラー行列ではない場合ということになります。<br />
例えば、<br />
$\begin{pmatrix}1&1\\0&1\end{pmatrix}$ は対角化可能ではありません。<br />
3次以上の正方行列の場合はもう少し複雑です。</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-56374060623594061972020-05-27T02:50:00.005+09:002022-04-19T10:13:55.676+09:00数学リテラシー1(第7回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="color: black; font-family: "times new roman"; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(水曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="color: black; font-family: "times new roman"; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="text-align: left;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
今回は固有値、固有ベクトル、行列の対角化、行列の $n$ 乗についての内容でした。<br />
<br />
<span style="font-size: large;"><u>固有値・固有ベクトル</u></span><br />
<u></u><br />
<span style="font-family: inherit;">固有値と固有ベクトルについてまとめておきます。</span><br />
<br />
固有値と固有ベクトルというのは次のように定義されます。<br />
<span style="background-color: white; color: black; display: inline; float: none; font-family: "times new roman"; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;"><br /></span>
<span style="background-color: white;">${\bf v}$</span>が<b>固有ベクトル</b>であるというのは、行列 $A$ を左からかけたもの $A{\bf v}$<br />
が ${\bf v}$ の定数倍になっているような<br />
ゼロではないベクトルのことを言います。<br />
<br />
つまり、ある数 $\lambda$ を使って、<br />
$A{\bf v}=\lambda{\bf v}$ をみたす(ゼロではない)ベクトルのことです。<br />
ここで、$\lambda$ はある複素数になります。<br />
行列 $A$ が実数であっても、固有値が複素数になることはあります。<br />
<br />
この $\lambda$ のことを $A$ の<b>固有値</b>といいます。<span lang="EN-US"></span><br />
<br />
また固有ベクトルはかならず零ベクトルではないということに注意をしてください。<br />
<br />
もし零ベクトル を許すと、任意の複素数も固有値になってしまいます。<br />
$A{\bf 0}=\lambda{\bf 0}$ の関係式は任意の複素数$\lambda$ が満たすからです。<br />
<br />
このとき行列$A$に対して固有ベクトルは以下の方程式を満たすことわかります。<br />
$A{\bf v}=\lambda{\bf v}\Leftrightarrow (\lambda E-A){\bf v}={\bf 0}$<br />
<br />
つまり <span lang="EN-US">$\lambda E-A$ </span>という行列は 逆行列を持たないということになります。
<span lang="EN-US"></span><br />
<br />
なぜならば もし逆行列を持てば、$(\lambda E-A){\bf v}={\bf 0}$ という関係式に<br />
<span lang="EN-US">$\lambda E-A$ </span>の逆行列を左からかけることで
<span lang="EN-US">${\bf v}={\bf 0}$</span>という式が出てしまい、<br />
${\bf v}\neq {\bf 0}$ であることに矛盾するからです。<br />
<br />
よってわかったことは、$\lambda E-A$ <span style="font-family: inherit;">という行列が逆行列をもたないこと、</span><br />
<span style="font-family: inherit;">同値なことに、</span><br />
<span style="font-family: inherit;"><br /></span></div>
<div style="text-align: center;">
<span style="font-family: inherit;"><u>$\lambda E-A$ の行列式が零であるということ です。</u></span><span lang="EN-US"></span></div>
<span style="font-family: inherit;"><br /></span><span style="font-family: inherit;">つまり、行列 $A$の固有値$\lambda$ は、$\det(t E-A)=0$ </span>を満たす解ということになります。</div>
</div>
<br />
この式 $\det(t E-A)=0$ は多項式です。この多項式のことを<b>固有多項式</b>といいます。<br />
$A$ が $n$ 次正方行列であるなら、$\det(t E-A)$ は、$n$ 次固有多項式です。<br />
<br />
$A=\begin{pmatrix}a&b\\c&d\end{pmatrix}$<br />
のときに、固有多項式を求めてみると、<br />
$$\det(tE-A)=\det\begin{pmatrix}t-a&-b\\-c&t-d\end{pmatrix}$$<br />
$$=(t-a)(t-d)-bc=t^2-(a+d)t+ad-bc$$<br />
$$=t^2-\text{tr}(A)t+\det(A)$$<br />
となります。<br />
<br />
この多項式の根を、$\lambda_1,\lambda_2$ とします。<br />
$a,b,c,d$ が実数、つまり $A$ が実行列であるとすると、<br />
<ul style="text-align: left;">
<li>2つの異なる実数解の場合、</li>
<li>2つの異なる複素解の場合、</li>
<li>重解</li>
</ul>
の3パターンあります。<br />
<br />
では次の行列<br />
$$A=\begin{pmatrix}0&-1\\2&3\end{pmatrix}$$<br />
の固有ベクトルを考えましょう <span lang="EN-US"></span><br />
<br />
まずこの行列の固有値を求めましょう。<br />
最初に固有ベクトルを求めようとしてはいけません。 <span lang="EN-US"></span><br />
<br />
そのために固有多項式を求めます。<br />
$\text{tr}(A)=3$, $\text{det}(A)=2$<br />
ですから、$t^2-3t+2$ です。<br />
<br />
固有値は 固有多項式の根のですから この二次方程式を解いて、<br />
固有値は全部で、$1,2$ の2つあります。<br />
<br />
まず、<br />
固有値が $1$ の場合の 固有ベクトル求めましょう。<br />
<span lang="EN-US">このとき、$tE-A=E-A\begin{pmatrix}1&1\\-2&-2\end{pmatrix}$ </span><br />
<br />
固有ベクトルは 次のような 連立方程式の解です。<br />
<span lang="EN-US">$\begin{pmatrix}1&1\\-2&-2\end{pmatrix}\begin{pmatrix}x\\y\end{pmatrix}={\bf 0}$</span><br />
この連立方程式は、$x+y=0$と同じですから、この方程式を<br />
<br />
として、$c$ を任意の実数として、<br />
$\begin{pmatrix}x\\y\end{pmatrix}=c\begin{pmatrix}1\\-1\end{pmatrix}$<br />
<span lang="EN-US">とします。 </span>ここで、固有ベクトルは<br />
${\bf v}_1=\begin{pmatrix}1\\-1\end{pmatrix}$<br />
なります 。<span lang="EN-US"></span><br />
<br />
ここで、固有ベクトルは ゼロでないベクトルであれば何でもいいです。<span lang="EN-US"></span><br />
<br />
というのも、ある固有値 に対する固有ベクトルというのは定数倍をしても <span lang="EN-US"></span><br />
<br />
その固有値の固有ベクトルですから、本来その方向しか決まりません。<span lang="EN-US"></span><br />
ですので、固有ベクトルを与えるときは 連立方程式 を満たす。<br />
適当なゼロではないベクトルを選ぶことになります<br />
<br />
同様に固有値が $2$ の場合の 固有ベクトル求めましょう。<span lang="EN-US"> </span><br />
そのとき、固有ベクトルが満たす連立方程式は<br />
$(2E-A)\begin{pmatrix}x\\y\end{pmatrix}=\begin{pmatrix}2&1\\-2&-1\end{pmatrix}\begin{pmatrix}x\\y\end{pmatrix}={\bf 0}$<br />
であるから、<br />
ゼロではないベクトル求めると<span lang="EN-US"></span><br />
<br />
${\bf v}_2=\begin{pmatrix}-1\\2\end{pmatrix}$<br />
となります。この場合も、連立方程式を満たすゼロではないベクトルを選びました。<br />
<br />
<br />
<u><span style="font-size: large;">行列の対角化</span></u><br />
ここでは、行列の対角化について考えます。<br />
<br />
<span style="mso-ansi-language: EN-CA;">行列の対角化というのは 行列 <span lang="EN-CA">$A$ </span>に対して 行列 $P$<span lang="EN-CA"> </span>とその逆行列 $P^{-1}$</span><br />
<span style="mso-ansi-language: EN-CA;">を挟むことによって行列を対角行列にするということです。</span><br />
<span style="mso-ansi-language: EN-CA;">つまり、$P^{-1}AP$ を対角行列にするのです。</span><br />
<span style="mso-ansi-language: EN-CA;"><br /></span>
<span style="mso-ansi-language: EN-CA;">ここで先ほどの例を考えます。行列 <span lang="EN-CA">$P$ </span>として</span><br />
<span style="mso-ansi-language: EN-CA;">固有ベクトルを並べたものを考えましょう。</span><br />
<span style="mso-ansi-language: EN-CA;">つまり、$P=({\bf v}_1{\bf v}_2)=\begin{pmatrix}1&-1\\-1&2\end{pmatrix}$</span> です。</div>
<br />
このとき、$AP=A({\bf v}_1{\bf v}_2)=(A{\bf v}_1A{\bf v}_2)$<br />
となります。<br />
最後の行列は、$AP=({\bf v}_12{\bf v}_2)=\begin{pmatrix}1&0\\0&2\end{pmatrix}({\bf v}_1{\bf v}_2)=\begin{pmatrix}1&0\\0&2\end{pmatrix}P$<br />
となります。<br />
よって、$P$ の逆行列 $P^{-1}$ を左からかけることで、<br />
$$P^{-1}AP=\begin{pmatrix}2&0\\0&1\end{pmatrix}$$<br />
となるので、このとき、$A$ は $P$ によって対角行列にすることができたことになります。<br />
つまり、$P=\begin{pmatrix}1&-1\\-1&2\end{pmatrix}$ とするとき、<br />
$P^{-1}AP=\begin{pmatrix}1&0\\0&2\end{pmatrix}$ となります。<br />
固有ベクトルの順番を入れ替えて、$Q=\begin{pmatrix}-1&1\\2&-1\end{pmatrix}$<br />
としてやると、<br />
$Q^{-1}AQ=\begin{pmatrix}2&0\\0&1\end{pmatrix}$ が得られます。<br />
また、固有ベクトルを定数倍してやってやっても、対角化する行列($P$のこと)<br />
は違うものになるかもしれませんが、最終的な対角行列は同じものになります。<br />
ここで、対角化したときの対角行列は、固有値を対角成分に並べた行列になります。<br />
<br />
この計算は、この時だけうまくいったわけではなく、一般の行列 $A$ に対しても<br />
固有ベクトルを並べてできる行列 $P$ とその逆行列 $P^{-1}$ を持ってくると<br />
対角成分に固有値を並べた対角行列を求めることがわかります。<br />
<br />
ただし、対角化できるためには条件があって、<br />
固有ベクトルを並べて正方行列を作らなければならないということです。<br />
つまり、固有ベクトルが $n$ 個、この場合は、2個存在しないといけません。<br />
<br />
しかし、固有ベクトルは、定数倍をしても固有ベクトルですから、<br />
正確に言えば、固有ベクトルとして、一次独立な $n$ 個のベクトルを取る必要があります。<br />
そして、一次独立な固有ベクトルが $n$ 個(今は2個)とることができれば、<br />
$AP=PD$ となります。ここで、$D$ は対角行列です。<br />
<br />
今、$P$ 一次独立な $n$ 個のベクトル(今は2個のベクトル)から成っていたので、<br />
行列式がゼロではない、つまり、$P$ は逆行列を持つことになります。<br />
よって、行列 $A$ は<br />
$P^{-1}AP=D$ のように対角化することができます。<br />
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
<span style="font-size: large;"><u>行列の$n$ 乗</u></span><br />
<span style="background-color: white;"></span>
<br />
行列の対角化を利用して 行列の $<span lang="EN-US">n$ </span>乗を計算しましょう。<br />
<br />
正方行列 $A$ に対して、その $n$ 乗を求めてみます。<br />
$A^n$ は $A$ を $n$ 回かけて得られる行列ですが、対角化を求めることができます。<br />
<br />
$A$ は行列 $P=\begin{pmatrix}1&-1\\-1&2\end{pmatrix}$ とおくことで、<br />
$P^{-1}AP=\begin{pmatrix}1&0\\0&2\end{pmatrix}$ となります。<br />
ここで、$A^n$ をする代わりに、この行列の $n$ 乗を考えます。<br />
<br />
そうすると、$P^{-1}AP$ の $n$ 乗は、<br />
$(P^{-1}AP)^n=P^{-1}APP^{-1}AP\cdots P^{-1}AP=P^{-1}A^nP$ となり、<br />
$A^n$ が出現しました。<br />
一方、対角行列の $n$乗は、$\begin{pmatrix}1^n&0\\0&2^n\end{pmatrix}=\begin{pmatrix}1&0\\0&2^n\end{pmatrix}$<br />
となりますから、再び対角行列です。<br />
<br />
よって、$P^{-1}A^nP=\begin{pmatrix}1&0\\0&2^n\end{pmatrix}$<br />
となります、これはちょうど、$A^n$の対角化をしていることになります。<br />
<br />
この式から、両側から $P$ と $P^{-1}$ で挟むことで、<br />
$$A^n=P\begin{pmatrix}1&0\\0&2^n\end{pmatrix}P^{-1}$$<br />
$$=\begin{pmatrix}1&-1\\-1&2\end{pmatrix}\begin{pmatrix}1&0\\0&2^n\end{pmatrix}\begin{pmatrix}2&1\\1&1\end{pmatrix}$$<br />
$$=\begin{pmatrix}2-2^n&1-2^n\\-2+2^{n+1}&-1+2^n\end{pmatrix}$$<br />
<br />
となります。これが、$A^n$ の一般公式ということになります。<br />
<br />
このように行列の $n$ 乗を求めるのに、<br />
まず、行列を対角化 $P^{-1}AP=D$ をしておきます。<br />
この対角行列 $D$ には、その対角成分に固有値が並びます。<br />
この行列の $n$ 乗を求めることで、<br />
$P^{-1}A^nP=D^n$<br />
を得ることができます。$D^n$ は再び対角行列になっています。<br />
<br />
(このことから、すぐわかることは、$A$ が対角化できるのなら、$A^n$ も対角化を<br />
することができます。対角化をいつすることができるのかについては、上の<br />
行列の対角化の部分の最後を見てください。)<br />
<br />
この式に $P$ と $P^{-1}$ を両側からかけることによって、<br />
$A^n=PD^nP^{-1}$ を計算することができます。<br />
この式行列 $A$ の $n$ 乗の公式ということになります。<br />
<br />
このように、固有値は、行列の $n$ 乗を計算するのに大変役に立っている<br />
ということになります。</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-38000342232149513202020-05-25T11:42:00.002+09:002022-04-19T10:14:06.184+09:00数学リテラシー1(第6回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="-webkit-text-stroke-width: 0px; color: black; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(火曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="text-align: left;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
まずは正射影が表す線形変換について考えます。<br />
<br />
<span style="font-size: large;"><u>正射影とその表す行列</u></span><br />
<div style="text-align: left;">
<span style="font-family: inherit;">まず次の問題を解いてみましょう。</span><br />
<span style="font-family: inherit;"></span><br />
<span style="font-family: inherit;">問題</span><br />
<span style="font-family: inherit;">平面上の正射影の表す線形写像を行列 $A$ をもって表示せよ。</span><br />
<br />
<br />
<span style="font-family: inherit;">まず、平面上の<b>正射影</b>というのは次のような写像</span><br />
<span style="font-family: inherit;">$f:{\mathbb R}^2\to {\mathbb R}^2$ のことです </span><span lang="EN-US"></span><br />
<br />
<span style="font-family: inherit;">まず平面上の直線 $L$ を用意します 。</span><br />
<br />
<span style="font-family: inherit;">平面上に点 ${\bf x}$ を取ります。この点 ${\bf x}$ から 直線 $L$ への垂線を考え,</span><br />
<span style="font-family: inherit;">その垂線と $L$ の交わったところ (つまり垂線の足)を $f({\bf x})$ とするのです。</span><span style="font-family: inherit;"><br /></span><br />
<span style="font-family: inherit;">正射影というのは、ある直線へのベクトルの影を求める操作ということになります。</span><br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: inherit; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
この場合、直線 $L$ は原点を通る場合のみであることに注意しましょう。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: inherit; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
(一般に正射影といった場合、原点を通るとは限りません。)</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: inherit; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike></div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
今、<span style="font-family: inherit;">2通りのやり方で行列 $A$ を求めてみます。</span><span style="font-family: inherit;"><br /></span><br />
<span style="font-family: inherit;">
<br />
(1つ目のやり方)</span><br />
<span style="font-family: inherit;">1つ目の設定は正射影が何らかの方法で線形写像であることが分かったします。</span><br />
<span style="font-family: inherit;"><span style="font-family: inherit;"></span></span><br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span style="font-family: inherit;"><span style="font-family: inherit;"><span style="font-family: inherit;">その仮定の下で解いてみます </span><span lang="EN-US"></span></span></span></div>
<span style="font-family: inherit;"><span style="font-family: inherit;">
</span></span>
<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span style="font-family: inherit;"><span style="font-family: inherit;"><span style="font-family: inherit;"></span><span style="font-family: inherit;"></span></span></span></div>
<span style="font-family: inherit;"><span style="font-family: inherit;">
</span></span>
<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span style="font-family: inherit;"><span style="font-family: inherit;">まず直線 $L$ を $y=ax$ として与えておきます。 </span></span></div>
<span style="font-family: inherit;"><span style="font-family: inherit;">
</span><b></b><i></i><u></u><sub></sub><sup></sup><strike></strike></span><span style="font-family: inherit;">線形写像というのは行列の左から積で表されていたことは前回やりました。</span><br />
<br />
<span style="font-family: inherit;">また、行列 $A$ は標準基底(基本ベクトル)と像となるベクトルを並べた行列でした。</span><br />
<br />
<span style="font-family: inherit;">平面上の標準基底というのは、</span><br />
<span style="font-family: inherit;">${\bf e}_1=\begin{pmatrix}1\\0\end{pmatrix}$, ${\bf e}_2=\begin{pmatrix}0\\1\end{pmatrix}$</span><br />
<span style="font-family: inherit;">でした。</span><br />
<span style="font-family: inherit;">つまり、$f({\bf e}_i)={\bf a}_i$ ($i=1,2$) としたとき、</span><br />
<span style="font-family: inherit;">$({\bf a}_1{\bf a}_2)$ が求める行列 $A$ ということになります。</span><br />
<span style="font-family: inherit;"><br /></span>
<span style="font-family: inherit;">あとは、$f({\bf e}_1)$ と $f({\bf e}_2)$ を求めればよいのですが、</span><br />
<span style="font-family: inherit;">それをここでは三角比を用いて求めてみます。</span><span lang="EN-US"><span style="font-family: inherit;"> </span></span><br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgvw-V7Tr74UtDu83BnHJUqOief0DD4Nv_ekaR5XjII3mUtV5W2AisiIDdBGXB0P8tX1-W8ZdzxrojwViC8VyRk3FUMBkjH2fWQaBuFjTYQZU-IPle7HRTElUPNt6BrDl7zk415d1aU3JiL/s1600/sankaku.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="215" data-original-width="279" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgvw-V7Tr74UtDu83BnHJUqOief0DD4Nv_ekaR5XjII3mUtV5W2AisiIDdBGXB0P8tX1-W8ZdzxrojwViC8VyRk3FUMBkjH2fWQaBuFjTYQZU-IPle7HRTElUPNt6BrDl7zk415d1aU3JiL/s1600/sankaku.png" /></a></div>
<br />
<span lang="EN-US"><span style="font-family: inherit;"><span style="font-family: inherit;">ここで、$y=ax$ の傾きは正の数であるとします。</span></span></span><br />
<span lang="EN-US"><span style="font-family: inherit;"><span style="font-family: inherit;">$f({\bf e}_1)$ の像は、ベクトル $\vec{OA}$ なのですが、</span></span></span><br />
<span lang="EN-US"><span style="font-family: inherit;"></span></span><br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span lang="EN-US"><span style="font-family: inherit;"><span style="font-family: inherit;">その $x$ 座標は、$OH$ ですが 、長さ $OA$ は、$a=\tan\theta$ としたときの、</span></span></span></div>
<span lang="EN-US"><span style="font-family: inherit;">
</span></span>
<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span lang="EN-US"><span style="font-family: inherit;"><span style="font-family: inherit;">$\cos\theta$ に対応するから、$\cos $ を $\tan$ で表す式</span></span></span></div>
<span lang="EN-US"><span style="font-family: inherit;">
</span></span>
<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span lang="EN-US"><span style="font-family: inherit;"><span style="font-family: inherit;">$\cos\theta=\frac{1}{\sqrt{1+\tan^2\theta}}$ を用いて、</span><span style="font-family: inherit;">$OA=\frac{1}{\sqrt{1+a^2}}$ となります。</span></span></span></div>
<span lang="EN-US"><span style="font-family: inherit;">
</span></span>
<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span lang="EN-US"><span style="font-family: inherit;"><span style="font-family: inherit;">また </span>$\sin$ の方を求めておけば、</span></span></div>
<span lang="EN-US"><span style="font-family: inherit;">
</span></span>
<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span lang="EN-US"><span style="font-family: inherit;">$\sin\theta=\tan\theta\cos\theta=\frac{a}{\sqrt{1+a^2}}$ となります。</span></span><br />
<span lang="EN-US"><span style="font-family: inherit;">ここで、平方根はプラスの方向を取っている。つまり、</span></span><br />
<span style="font-family: inherit;">$\cos$ のうち、正の方を取っていますが、それは、</span><br />
<span style="font-family: inherit;">$a>0$ であることを暗に仮定しているからで、 $a<0$ であるときは、</span><br />
$\cos\theta=-\frac{1}{\sqrt{1+\tan^2\theta}}$ となります。</div>
<span lang="EN-US"><span style="font-family: inherit;">
</span></span><span style="font-family: inherit;"></span><span style="font-family: inherit;"></span><span style="font-family: "times new roman";"></span>
<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span lang="EN-US"><span style="font-family: inherit;"><br /></span></span>
<span lang="EN-US"><span style="font-family: inherit;">話を元に戻します。</span></span></div>
<span lang="EN-US"><span style="font-family: inherit;">
</span></span>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span lang="EN-US"><span style="font-family: inherit;">これにより、$OH=OA\cos\theta=\frac{1}{1+a^2}$ となります。</span></span></div>
<span lang="EN-US"><span style="font-family: inherit;">
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
また、$f({\bf e}_1)$ の $y$ 座標は、$AH$ ですから、</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$AH=OA\sin\theta=\frac{a}{1+a^2}$ となり、</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$f({\bf e}_1)=\begin{pmatrix}\frac{1}{1+a^2}\\\frac{a}{1+a^2}\end{pmatrix}$</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: inherit; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
となります。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-variant: normal; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<br /></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
同様に、$f({\bf e}_2)$ を求めてみると、</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
まず、$f({\bf e}_2)$ の $x$ 座標は、ちょうど $AH$ ですから、$\frac{a}{1+a^2}$ </div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
と一致します。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$y$ 座標は、$OB-OH$ ですから、$1-\frac{1}{1+a^2}=\frac{a^2}{1+a^2}$ </div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
となります。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<br /></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
ゆえに、$f({\bf e}_2)=\begin{pmatrix}\frac{a}{1+a^2}\\\frac{a^2}{1+a^2}\end{pmatrix}$</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
となります。 </div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<br /></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
この2つのベクトル $f({\bf e}_1)$ と $f({\bf e}_2)$ を並べることで得られる行列</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$$\begin{pmatrix}\frac{1}{1+a^2}&\frac{a}{1+a^2}\\\frac{a}{1+a^2}&\frac{a^2}{1+a^2}\end{pmatrix}$$</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
は求める行列ということになります。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$a=\tan\theta$ を用いると、この行列は、</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$$\begin{pmatrix}\cos^2\theta&\cos\theta\sin\theta\\\cos\theta\sin\theta&\sin^2\theta\end{pmatrix}$$</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
となります。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<br /></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span style="font-family: inherit;">(2つ目)</span></div>
</span></span><br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-variant: normal; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
</div>
正射影が線形変換であるということを 用いないで $A$ を計算してみます。 <span style="mso-spacerun: yes;"> </span><span lang="EN-US"></span><br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-variant: normal; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
</div>
しかし この計算途中で正射影は線形変換であるということがわかります <span lang="EN-US"></span><br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-variant: normal; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span style="font-family: inherit;"><br /></span></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-variant: normal; letter-spacing: normal; margin: 0px; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span style="font-family: inherit;">正射影というのを3つ基本的な写像の合成だと考えます。</span></div>
<span style="font-family: inherit;">3つの写像は順番に、 $-\theta$ 回転、 $x$ </span><span style="font-family: inherit;">軸への射影、$\theta$ 回転</span><span style="font-family: inherit;">です。</span><br />
<span style="font-family: inherit;">この写像の合成は正射影を実現しています。</span><span lang="EN-US"></span><br />
<br />
<span lang="EN-US"><span style="font-family: inherit;"><br /></span></span>
<span lang="EN-US"><span style="font-family: inherit;"></span></span><br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: inherit; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span lang="EN-US"><span style="font-family: inherit;">この分解は 前回の鏡映を3つの行列の積で変えたことに少し似ていますね。</span></span></div>
<span lang="EN-US"><span style="font-family: inherit;">
</span></span>回転はわかりますが、 $x$ 軸への正射影は、<br />
<br />
$f(\begin{pmatrix}x\\y\end{pmatrix})=\begin{pmatrix}x\\0\end{pmatrix}$<br />
で、線形変換を行列を用いて<br />
$f(\begin{pmatrix}x\\y\end{pmatrix})=\begin{pmatrix}1&0\\0&0\end{pmatrix}\begin{pmatrix}x\\y\end{pmatrix}$ とあらわすことができるので、<br />
$x$ 軸への正射影は線形変換であることがわかります。<br />
<br />
<span lang="EN-US"><span style="font-family: inherit;">よって、この<span lang="EN-US">3</span>つはそれぞれ線形変換ですから その合成も線形変換になります。<br />
<br />
</span></span><span lang="EN-US"><span style="font-family: inherit;">これにより正射影変換は線形変換であるということがわかりました </span></span><span lang="EN-US"><span style="font-family: inherit;"><br /></span></span><br />
<br />
<span style="font-family: inherit;">この行列の積を求めると以下のようになります。</span><br />
$$\begin{pmatrix}\cos\theta&-\sin\theta\\\sin\theta&\cos\theta\end{pmatrix}\begin{pmatrix}1&0\\0&0\end{pmatrix}
\begin{pmatrix}\cos\theta&\sin\theta\\-\sin\theta&\cos\theta\end{pmatrix}$$<br />
$$=\begin{pmatrix}\cos^2\theta&\cos\theta\sin\theta\\\cos\theta\sin\theta&\sin^2\theta\end{pmatrix}$$<br />
となり、確かに上の計算と合いました。<br />
<br />
<br />
<span style="font-size: large;"></span><u></u><br />
<span style="font-size: large;"><u>直交行列によって対角化できる行列</u></span><br />
これまで ある線形変換 $A$ をある直交行列 $R$ を用いて <span lang="EN-US"></span><br />
<br />
$$RAR^{-1}$$<br />
のように 得られる線形変換について考えていました。<br />
$A$ に当たる行列は、簡単な行列、例えば、$\begin{pmatrix}1&0\\0&-1\end{pmatrix}$<br />
や、$\begin{pmatrix}1&0\\0&0\end{pmatrix}$ など、対角行列(対角成分以外はすべて $0$ )<br />
のような行列を考えていました。<br />
<br />
一般に、対角行列 $A$ は、$\begin{pmatrix}\lambda_1&0\\0&\lambda_2\end{pmatrix}$<br />
とすることができますが、<br />
このような $A$ に対して、$RAR^{-1}$ はどのような性質を持つのでしょうか?<br />
<br />
<br />
$RAR^{-1}$ の転置行列を取ってみます。<br />
ここで、$^t(XY)=^tY^tX$ となることに注意しましょう。<br />
そうすると、$^t(RR^{-1})=^t(R^{-1})^tR=E$ ですから、<br />
$^t(R^{-1})=(^tR)^{-1}$ となります。<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
つまり、逆行列を取る操作と転置行列を取る操作はどちらをさきに行っても同じということです。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
そういうわけで、$(^tR)^{-1}$ や $^t(R^{-1})$ も区別はなく、$^tR^{-1}$ と書いても</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: Times New Roman; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
差支えないということになります。</div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
また、$R$ が直交行列であるとすれば、$R^{-1}=^tR$ ですから、<br />
$^tR^{-1}=R$ ということにもなります。<br />
そうすると、<br />
<br />
$$^t(RAR^{-1})=^t(RA^tR)=R^t(RA)=R^tA^tR=RAR^{-1}$$<br />
となります。途中、行列 $A$ が対角行列であるから、$^tA=A$ であることを用いました。<br />
<br />
このことからわかることは、$X=RAR^{-1}$ という行列は、<br />
$^tX=X$ であることです。<br />
<br />
このように、転置行列を施すと、もとの行列に戻る行列を対称行列といいます。<br />
<br />
たしかに、さっき求めた行列は、$(1,2)$ 成分と $(2,1)$ 成分は一致していましたね。<br />
先週の鏡映変換も、$(1,2)$ 成分と $(2,1)$ 成分はどちらも $\sin2\theta$ でした。<br />
<br />
まとめると、対角行列 $A$ に対して、直交行列 $R$ を用いて<br />
$RAR^{-1}$ を求めると、対称行列になるということがわかりました。<br />
<br />
実は、この逆も成り立ちます。<br />
<br />
定理<br />
任意の対称行列は、ある直交行列 $R$ と対角行列 $A$ を用いて、<br />
$RAR^{-1}$ と書き表される。<br />
<br />
この定理は次回以降どこかで現れます。<br />
<br />
<br />
<span style="font-size: large;"><u>行列式は符号付面積であること </u></span><span lang="EN-US"></span><br />
<br />
これは行列式っていうのは、ある意味、符号付面積であると 言うことを考えたいと思います。<br />
<br />
ここで符号付面積というのは平面上の<span lang="EN-US">2</span>つのベクトル ${\bf a}_1$ と ${\bf a}_2$<br />
によって作られるようなで平行四辺形の符号付面積という意味です<br />
<br />
この2つが一致するということをここで見て行きます <span lang="EN-US"></span><br />
<br />
$A$ を $2\times 2$ 行列であるとし、その行列を縦ベクトルとして<br />
$A=({\bf a}_1{\bf a}_2)$ としてあらわされるとします。<br />
このとき、<br />
${\bf a}_1=\begin{pmatrix}r_1\cos\theta_1\\r_1\sin\theta_1\end{pmatrix}$<br />
<div style="color: black; font-family: "times new roman"; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
${\bf a}_2=\begin{pmatrix}r_2\cos\theta_2\\r_2\sin\theta_2\end{pmatrix}$</div>
のようにあらわしたとします。ここで、平面上の極座標表示を用いました。<br />
ここで、${\bf a}_1$ と ${\bf a}_2$ のうちどちらかが0ベクトルであるとすると、<br />
行列式はゼロなり、そのとき、符号付面積は0になりますので<br />
この2つは一致しているということになります。<br />
<br />
では、どちらもゼロベクトルではないときを考えます。そのとき、<br />
$\det({\bf a}_1{\bf a}_2)=r_1r_2(\cos\theta_1\sin\theta_2-\cos\theta_2\sin\theta_1)=r_1r_2\sin(\theta_2-\theta_1)$<br />
となります。<br />
<br />
ここで、$r_2|\sin(\theta_2-\theta_1)|$ は、${\bf a}_2 $を ${\bf a}_1$ に垂線を<br />
おろしたときにできる平行四辺形の高さになります。<br />
つまり、このとき、$\det(A)$ は、${\bf a}_1$ を底辺とする高さ $r_2|\sin(\theta_2-\theta_1)|$<br />
となる、${\bf a}_1$ と ${\bf a}_2$ によって作られる平行四辺形の面積ということになります。<br />
<br />
また、符号付面積というのはどういうことかというと、$\sin(\theta_2-\theta_1)$<br />
が負の数になることも考慮する必要があるということに対応します。<br />
つまり、$0<\theta_2-\theta_1<\pi$のときは、その値は正の数になりますが、<br />
$\pi<\theta_2-\theta_1<2\pi$ となると、負の数になります。<br />
つまり、ベクトル ${\bf a}_1$ と ${\bf a}_2$ がこの順番に角度が180度より小さく<br />
なるのなら、面積は正の数であり、${\bf a}_2$ と ${\bf a}_1$ の順に角度が180度<br />
より小さくなる時、面積は負の数になります。<br />
ということは、行列式が0になるのは、2つのベクトルが0度をなすとき、もしくは<br />
180度をなすときということになります。<br />
つまり、それは、いいかえれば、2つのベクトルが<br />
ちょうど平行になっているときです。<br />
<br />
<br />
<span style="font-size: large;"><u>行列式がゼロでないとき</u></span><br />
上で行列式がゼロでないとき、<br />
2つのベクトルは平行ではないということを意味していました。<br />
<br />
ここで、ベクトル ${\bf a}_1$ と ${\bf a}_2 $が<b>一次独立</b>であるということを<br />
実数 $c_1,c_2$ が $c_1{\bf a}_1+c_2{\bf a}_2={\bf 0}$ を満たすとき、$c_1=c_2=0$ である<br />
と定義します。<br />
一次独立でないことを<b>一次従属</b>といいます。<br />
<br />
ベクトル ${\bf a}_1$ と ${\bf a}_2$ が一次独立であることと、<br />
それらのベクトルが平行であることは同値です。<br />
<br />
もし、${\bf a}_1$ と ${\bf a}_2 $ が平行とすると、${\bf a}_1=k{\bf a}_2$<br />
もしくは、${\bf a}_2=k{\bf a}_1$ を満たす実数$k$ が存在することと同値です。<br />
また、ベクトル${\bf a}_1$ と ${\bf a}_2$ が一次従属であるとすると、<br />
$c_1{\bf a}_1+c_2{\bf a}_2={\bf 0}$ となる $(c_1,c_2)\neq (0,0)$が<br />
存在することと同値ですが、$c_1\neq 0$ であるとすると、 $c_1$ で割ることで、<br />
${\bf a}_1=k{\bf a}_2$ の形になります。<br />
同じように、 $c_2\neq 0 $であるときは、${\bf a}_2=k{\bf a}_1$ が成り立ちます。<br />
<br />
よって、ベクトル ${\bf a}_1$ と ${\bf a}_2$ が一次従属であるということは、<br />
それらが、平行であるということと同値となります。<br />
言いかえれば、ベクトルと一次が独立であることと、<br />
2つのベクトル${\bf a}_1$ と ${\bf a}_2$ <br />
が平行ではないことが同値であることがわかります。<br />
まとめますと、以下のようになります。<br />
$\det(A)\neq 0$ であることは、2つのベクトル${\bf a}_1$ と ${\bf a}_2$ が<br />
平行ではないとき、つまり、${\bf a}_1$ と ${\bf a}_2$ が一次独立であるときを意味します。<br />
また、第4回でもやったように、行列式 $\det(A)$ が0ではないということは、<br />
行列$A$ に逆行列が存在することと同値になります。<br />
よって、以下が同値であるということになります。</div>
</div>
</div>
<ul style="text-align: left;">
<li>行列 $A$ が逆行列をもつ</li>
<li>$\det(A)\neq 0$ である。</li>
<li>$A=({\bf a}_1{\bf a}_2)$ としたとき、${\bf a}_1$ と ${\bf a}_2$ は一次独立である。</li>
<li>${\bf a}_1$ と ${\bf a}_2$ は平行ではない。</li>
</ul>
<div>
また、この否定をとると、<br />
<ul style="text-align: left;">
<li>行列 $A$ が逆行列を持たない</li>
<li>$\det(A)=0$ である。</li>
<li>$A=({\bf a}_1{\bf a}_2)$ としたとき、${\bf a}_1$ と ${\bf a}_2$ は一次従属である。</li>
<li>${\bf a}_1$ と ${\bf a}_2$ は平行である。</li>
</ul>
<br />
ということになります。</div>
</div>
<span style="background-color: white; color: black; display: inline; float: none; font-family: "times new roman"; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
</span>
<ul style="text-align: left;"><span style="background-color: white; color: black; display: inline; float: none; font-family: "times new roman"; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
</span></ul>
<span style="background-color: white; color: black; display: inline; float: none; font-family: "times new roman"; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
</span><span style="background-color: white;"><span style="background-color: white; color: black; display: inline; float: none; font-family: "times new roman"; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;"></span></span>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-44823974481790480972020-05-19T12:58:00.002+09:002022-04-19T10:14:17.372+09:00数学リテラシー1(第5回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(水曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a><br />
<br />
前回は行列の一般論を行いました。<br />
<br />
今回からを用いて、一次変換(線形変換)の扱い方を学びます。<br />
$(2,2)$ 行列に一次変換の本質が詰まっています。<br />
<br />
ですので、$(2,2)$ 行列をプロトタイプとし、その後一般のサイズの一次変換に出あった<br />
ときにも同じように扱えるようにしたいと思います。<br />
<br />
前回で重要だったことは、<br />
線形写像 $f:{\mathbb R}^n\to {\mathbb R}^m$<br />
(任意のベクトル${\bf x},{\bf y}$ と任意の実数 $\lambda$ に対して、<br />
$f({\bf x}+{\bf y})=f({\bf x})+f({\bf y})$ かつ $f(\lambda {\bf x})=\lambda f({\bf x})$ が成り立つ写像のこと)<br />
<br />
は、必ずある行列 $(m,n)$ 行列を用いて、<br />
$f({\bf x})=A{\bf x}$ としてあらわされるということでした。<br />
<br />
<u></u><br />
<u><span style="font-size: large;">(2,2) 行列による一次変換</span></u><br />
線形写像が $f:{\mathbb R}^2\to {\mathbb R}^2$ の場合には、ある $(2,2)$ 行列を用いて<br />
$f({\bf x})=A\cdot {\bf x}$ としてあらわされることになります。<br />
この行列 $A$ はどのようにして計算できるか考えてみましょう。<br />
<br />
ベクトル ${\bf x}=\begin{pmatrix}x_1\\x_2\end{pmatrix}$ は、<br />
${\bf e}_1=\begin{pmatrix}1\\0\end{pmatrix}$, <br />
${\bf e}_2=\begin{pmatrix}0\\1\end{pmatrix}$<br />
を用いて、<br />
<br />
${\bf x}=x_1{\bf e}_1+x_2{\bf e}_2$<br />
のように書くことができます。<br />
このとき、$f$ の線形性から、<br />
<br />
$f({\bf x})=f(x_1{\bf e}_1+x_2{\bf e}_2)=x_1f({\bf e}_1)+x_2f({\bf e}_2)$<br />
のようになり、$f({\bf e}_1)={\bf a}_1$ かつ $f({\bf e}_2)={\bf a}_2$<br />
のように置きます。ここで、${\bf a}_1$ と ${\bf a}_2$ は<br />
2次元のユークリッド空間のベクトルです。<br />
<br />
そうすると、$f({\bf x})$ は、$x_1{\bf a}_1+x_2{\bf a}_2=({\bf a}_1{\bf a}_2)\begin{pmatrix}x_1\\x_2\end{pmatrix}$<br />
<br />
となり、$A=({\bf a}_1{\bf a}_2)$ とすれば、<br />
$$f({\bf x})=A{\bf x}$$<br />
ということになります。つまり、線形写像 $f$ に対して求めようと思っていた<br />
左からかける行列 $A$ は、$({\bf a}_1{\bf a}_2)$ と計算できることになります。<br />
<br />
この行列 $A$ は、2つの縦ベクトル ${\bf a}_1$ と ${\bf a}_2$ を並べてできる<br />
$(2,2)$ 行列です。<br />
<br />
もともと、${\bf a}_1$ と ${\bf a}_2$ が何だったかというと、<br />
${\bf e}_1$ と ${\bf e}_2$ の $f$ による行先でした。<br />
<br />
$A$ を求めたければ、この2つの単位ベクトル<br />
${\bf e}_1$ と ${\bf e}_2$ の像を並べてできる行列を求めればよいことになります。<br />
ベクトル ${\bf e}_1, {\bf e}_2$ のことを、2次元の<b>標準ベクトル</b>(<b>基底</b>)といいます。</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<br />
<span style="font-size: large;"><u>$x$ 方向と $y$ 方向への変倍の定数倍、回転、鏡映</u></span><br />
例3.2.1 <br />
平面上、 $x$ 方向に $\lambda_1$ 倍し、<br />
$y$ 方向に $\lambda_2$ 倍するような線形写像は、<br />
$$\begin{pmatrix}1\\0\end{pmatrix}\mapsto\begin{pmatrix}\lambda_1\\0\end{pmatrix},\begin{pmatrix}0\\1\end{pmatrix}\mapsto \begin{pmatrix}0\\\lambda_2\end{pmatrix}$$<br />
ですから、<br />
$${\bf e}_1\mapsto \lambda_1{\bf e}_1$$ であり、<br />
$${\bf e}_2\mapsto \lambda_2{\bf e}_2$$ <br />
ということですから、$A=({\bf a}_1{\bf a}_2)=\begin{pmatrix}\lambda_1&0\\0&\lambda_2\end{pmatrix}$<br />
となります。<br />
<br />
<br />
例3.2.2<br />
次は、平面上の原点 $O$ を中心とした回転を考えましょう。<br />
回転運動が一次変換であることは次のようにしてわかります。<br />
$f_\theta$ を原点中心の $\theta$ 回転の写像とします。<br />
<br />
このとき、$O$ と ${\bf x}$ と ${\bf y}$ と ${\bf x}+{\bf y}$<br />
は、ある平行四辺形をなします。<br />
<br />
このとき、この平行四辺形を原点 $O$ を中心として<br />
一斉に $\theta$ 回転をしたとすると、<br />
<br />
平行四辺形の各点は、<br />
$O$ と $f_\theta({\bf x})$, $f_\theta({\bf y})$, $f_\theta({\bf x}+{\bf y})$<br />
に移ります。<br />
平行四辺形は、回転しても平行四辺形ですから、<br />
<br />
$f_\theta({\bf x}+{\bf y})=f_\theta({\bf x})+f_{\theta}({\bf y})$ が成り立ちます。<br />
また、${\bf x}$ と ${\bf y}$ が平行で、平行四辺形が作れない場合は、<br />
つぶれた平行四辺形と考えれば同じことが言えます。<br />
<br />
また、$\lambda$ を実数として、${\bf x}$ と $\lambda{\bf x}$ は、$\theta$ 回転しても<br />
してもその関係は変わりません。<br />
というのも、回転というのは、長さと角度を変えないからです。<br />
よって、<br />
$$f_\theta(\lambda {\bf x})=\lambda f_\theta({\bf x})$$<br />
となります。<br />
<br />
つまり、回転というのは、一次変換ということになります。<br />
$f_\theta$ から定まる $(2,2)$ 行列 $R_\theta$ を求めていきます。<br />
<br />
上で求めた方法をとります。<br />
標準基底 ${\bf e}_1$ と ${\bf e}_2$ の像がどうなるかを調べれば<br />
よいことになります。<br />
<br />
${\bf e}_1$ の $f_\theta$ による行き先は、$\begin{pmatrix}\cos\theta\\\sin\theta\end{pmatrix}$であり、<br />
${\bf e}_2$ の $f_\theta$ による行き先は、$\begin{pmatrix}-\sin\theta\\\cos\theta\end{pmatrix}$<br />
となります。<br />
<br />
${\bf a}_1=f_\theta({\bf e}_1)=\begin{pmatrix}\cos\theta\\\sin\theta\end{pmatrix}$</div>
${\bf a}_2=f_\theta({\bf e}_2)=\begin{pmatrix}-\sin\theta\\\cos\theta\end{pmatrix}$</div>
<br />
ですから、$R_\theta$ は、<br />
<br />
$$R_\theta=\begin{pmatrix}\cos\theta&-\sin\theta\\\sin\theta&\cos\theta\end{pmatrix}$$<br />
となります。<br />
<br />
つまり、$\theta$ 回転を表す行列は<br />
$$f_\theta({\bf x})=\begin{pmatrix}\cos\theta&-\sin\theta\\\sin\theta&\cos\theta\end{pmatrix}{\bf x}$$<br />
となります。</div>
<br />
<br />
一次変換の合成に対応する、行列は、行列の積となります。<br />
前回やったように、$f_B\circ f_A=f_{BA}$ ですから、<br />
$f_{R_{\theta_2}}\circ f_{R_{\theta_1}}=f_{R_{\theta_1}R_{\theta_2}}$<br />
また、$\theta_1$ 回転をして、$\theta_2$ 回転をしてできる一次変換は<br />
$\theta_1+\theta_2$ 回転した一次変換ですから、<br />
$f_{R_{\theta_2}}\circ f_{R_{\theta_1}}=f_{\theta_1+\theta_2}=f_{R_{\theta_1+\theta_2}}$<br />
となります。<br />
よって、この2つから、<br />
$$R_{\theta_1+\theta_2}=R_{\theta_2}R_{\theta_1}$$<br />
が成り立ちます。<br />
よって、<br />
$$\begin{pmatrix}\cos\theta_2&-\sin\theta_2\\\sin\theta_2&\cos\theta_2\end{pmatrix}\begin{pmatrix}\cos\theta_1&-\sin\theta_1\\\sin\theta_1&\cos\theta_1<br />
\end{pmatrix}=<br />
\begin{pmatrix}\cos(\theta_1+\theta_2)&-\sin(\theta_1+\theta_2)\\\sin(\theta_1+\theta_2)&\cos(\theta_1+\theta_2)<br />
\end{pmatrix}<br />
$$<br />
が成り立ちますが、この式の各成分は、三角関数の加法定理を意味しています。<br />
<br />
例3.2.3<br />
直線 $y=(\tan\theta)x $ に沿った鏡映変換を考えましょう。<br />
鏡映変換とは、ある直線による線対称変換を意味します。<br />
この直線 $y=(\tan\theta)x$ による鏡映変換は、<br />
<br />
(1) 原点での $-\theta$ 回転、<br />
(2) $x$ 軸による線対称変換、<br />
(3) 原点での $\theta$ 回転<br />
<br />
のこの順番による合成になります。<br />
<br />
これらは、上の例ですでに見たものばかりです。<br />
$x$ 軸による線対称変換は、$x$ 方向は変わらず (1倍)、$y$ 方向に $-1$ 倍<br />
をする一次変換です。<br />
一次変換の合成も一次変換ですから、<br />
鏡映もやはり一次変換です。<br />
<br />
この3つの一次変換を合成することで得られる一次変換を $g_\theta$ とし、<br />
そのときの行列を $S_\theta$ とすると、<br />
$$S_\theta=R_\theta\begin{pmatrix}1&0\\0&-1\end{pmatrix}R_{-\theta}$$<br />
となります。<br />
ここで、かける順番に気を付けましょう。<br />
わからなくなったら、$f_B\circ f_A=f_{BA}$であることと、<br />
一次変換は、左から行列をかけることであったことを思い出しましょう。<br />
<br />
よって、$S_\theta$ を実際計算をすると、<br />
<br />
$$S_\theta=\begin{pmatrix}\cos 2\theta&\sin2\theta\\\sin2\theta&-\cos2\theta\end{pmatrix}$$<br />
となります。<br />
鏡映変換は、線対称変換ですから、2回同じ変換を行うと<br />
元に戻ります。<br />
これは、<br />
$$g_\theta\circ g_\theta=\text{id}_{{\mathbb R}^2}$$<br />
であることを示せばよいですが、行列の言葉に直せば、<br />
<br />
$$S_\theta^2=(R_{-\theta}\begin{pmatrix}1&0\\0&-1\end{pmatrix}R_\theta)^2=R_{-\theta}\begin{pmatrix}1&0\\0&-1\end{pmatrix}R_{\theta}R_{-\theta}\begin{pmatrix}1&0\\0&-1\end{pmatrix}R_{\theta}$$<br />
<br />
<br />
$$R_{-\theta}\begin{pmatrix}1&0\\0&-1\end{pmatrix}\begin{pmatrix}1&0\\0&-1\end{pmatrix}R_{\theta}=R_{-\theta}R_\theta=E$$<br />
よって、$g_\theta\circ g_\theta=\text{id}_{{\mathbb R}^2}$ が示せました。<br />
<br />
<br />
<span style="font-size: large;"><u>これらの行列の可換性</u></span><br />
これまで、定数倍($x$方向と$y$方向に変倍する)の変換、<br />
回転、鏡映変換などを考えました。<br />
これらの変換やその合成も一次変換です。<br />
<br />
回転による変換同士は可換であることはわかります。<br />
足し算の可換性が成り立つから、<br />
<br />
$$R_{\theta_1}R_{\theta_2}=R_{\theta_1+\theta_2}=R_{\theta_2+\theta_1}=R_{\theta_2}R_{\theta_1}$$<br />
となり可換です。<br />
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
$x$ 方向に $\lambda_1$ 倍、 $y$ 方向に $\lambda_2$ 倍する一次変換を<br />
する行列を $T_{\lambda_1,\lambda_2}$ としますと、<br />
$$T_{\lambda_1,\lambda_2}R_{\theta}=\begin{pmatrix}\lambda_1\cos\theta&-\lambda_1\sin\theta\\\lambda_2\sin\theta&\lambda_2\cos\theta\end{pmatrix}$$<br />
$$R_{\theta}T_{\lambda_1,\lambda_2}
=\begin{pmatrix}\lambda_1\cos\theta&-\lambda_2\sin\theta\\\lambda_1\sin\theta&\lambda_2\cos\theta\end{pmatrix}$$<br />
<br />
よって、$(2,1)$ 成分を比べることによって、<br />
$T_{\lambda_1,\lambda_2}R_\theta=R_\theta T_{\lambda_1,\lambda_2}$<br />
が成り立つためには、$\lambda_1=\lambda_2$ でなければなりません。<br />
また、$\lambda_1=\lambda_2=\lambda$ であれば、
つまり、$T_{\lambda,\lambda}$ は原点 $O$ を中心とした、$\lambda$ 拡大を表します。<br />
<br />
また、$T_{\lambda,\lambda}=\lambda E$ であり、<br />
スカラー倍は $R_{\theta}$ などあらゆる一次変換と可換ですから、<br />
$T_{\lambda,\lambda}$ と $R_{\theta}$ は可換となります。<br />
<br />
<u><span style="font-size: large;">直交変換</span></u><br />
ここで、直交変換を定義し、対応する行列の性質を考察して終わります。<br />
<b>直交変換</b>とは、長さを変えない一次変換ことを言います。<br />
$f$ を直交変換とし、${\bf e}_1$ と ${\bf e}_2$ を上記の標準基底とします。<br />
<br />
このとき、${\bf e}_1$ と ${\bf e}_2$ の長さは $1$ ですから、<br />
$f({\bf e}_1)=\begin{pmatrix}a\\b\end{pmatrix}$<br />
$f({\bf e}_2)=\begin{pmatrix}c\\d\end{pmatrix}$
としますと、<br />
$a^2+b^2=1$ かつ $c^2+d^2=1$ が成り立ちます。<br />
<br />
また、長さを変えないということは、角度(の絶対値)も変えないということです。<br />
どうしてかというと、<br />
<br />
三角形 $OAB$ を考えます。$O$ は原点、$A,B$ はそれ以外の点とします。<br />
そうすると、直交変換は長さを変えないのだから、この三角形 $OAB$<br />
は $OAB$ と合同な三角形 $OA'B'$ に移ります。<br />
ここで、一次変換であることから、原点は原点に移ります。<br />
(なぜなら $O$ の表すベクトルを ${\bf 0}$ とすると<br />
$f({\bf 0})=f(2{\bf 0})=2f({\bf 0})$より、$f({\bf 0})={\bf 0}$ となるからです。)<br />
よって、角 $AOB$ は $A'OB'$ に移ります。<br />
ただし、三角形 $OAB$ が裏返るかもしれないので、角度の絶対値<br />
は変わりません。<br />
<br />
よって、$\begin{pmatrix}a\\b\end{pmatrix}$ と $\begin{pmatrix}c\\d\end{pmatrix}$<br />
は直交しなければなりません。<br />
<br />
つまり、内積は0なので、$ac+bd=0$ となります。<br />
ここで、$R=\begin{pmatrix}a&c\\b&d\end{pmatrix}$ とすると、<br />
$$^tRR=\begin{pmatrix}a^2+b^2&ac+bd\\ac+bd&c^2+d^2\end{pmatrix}$$<br />
<br />
が成り立ち、この右辺はちょうど単位行列 $E$ となります。<br />
つまり、直交変換 $f$ の表す行列 $R$ は、$^tRR=E$ となります。<br />
<br />
このような行列 $R$ のことを<b>直交行列</b>といいます。</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-36657151461174624852020-05-16T03:52:00.006+09:002022-04-19T10:14:27.498+09:00数学リテラシー1(第4回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(水曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
前回は行列とその四則演算について行いましたが、</div>
<div style="font-size: 16px;">
今回は、行列の除法と一次変換について行いました。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
行列とは、縦に$m$個、横に$n$個、長方形の形に数を並べて、<br />
さらにカッコで括ったものを考えます。<br />
例えば、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$$\begin{pmatrix}1&-1&3&3\\4&2&-10&2\\3&1&0&-2\end{pmatrix}$$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
などです。</div>
<div style="font-size: 16px;">
行列の加法と減法はその成分ごとに行い、積 $AB$ は、<br />
$A\in M(m,n,{\mathbb R})$ と $B\in M(n,k,{\mathbb R})$ のとき定義されて、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$$AB=(a_{ij})(b_{ij})=(\sum_{p=1}^na_{ip}b_{pj})\in M(m,k,{\mathbb R})$$</div>
<div style="font-size: 16px;">
のようにして行います。<br />
このように、積は、$A$ の列数と $B$ の行数が一致している場合のみ定義されます。</div>
<div style="font-size: 16px;">
ここでは、その数($A$ の列数、$B$ の行数)は $n$ です。</div>
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<br /></div>
<u><span style="font-size: large;">行列の定数倍</span></u><br />
<div style="font-size: 16px;">
行列を定数倍するということを前回では書かなかったので、ここで</div>
<div style="font-size: 16px;">
書いておきます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$\alpha$ を実数とし、$A$ を $(m,n)$ 行列とします。ここでは、</div>
<div style="font-size: 16px;">
行列 $A$ を $(a_{ij})$ とします。このとき、$A$ に実数 $\alpha$ をかける</div>
<div style="font-size: 16px;">
という操作を $\alpha\cdot A$ と書き、$(\alpha\cdot a_{ij})$ と定義します。</div>
<div style="font-size: 16px;">
これが意味することは、$A$ の $mn$ 個の成分を一斉に$\alpha$ 倍するということです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$A$ を $mn$ 個の成分を持つベクトルと考えれば、ベクトルを $\alpha$ する</div>
<div style="font-size: 16px;">
という操作と同じです。</div>
<div style="font-size: 16px;">
例えば、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$2\cdot\begin{pmatrix}2&-1\\-1&2\end{pmatrix}=\begin{pmatrix}4&-2\\-2&4\end{pmatrix}$</div>
<div style="font-size: 16px;">
となります。もちろん、この逆操作で、共通する因子があれば、その数を括り出して</div>
<div style="font-size: 16px;">
$\begin{pmatrix}4&-2\\-2&4\end{pmatrix}=2\cdot\begin{pmatrix}2&-1\\-1&2\end{pmatrix}$</div>
<div style="font-size: 16px;">
とすることと同値です。くれぐれも、</div>
<div style="font-size: 16px;">
$\begin{pmatrix}4&-2\\-2&4\end{pmatrix}=2\cdot\begin{pmatrix}2&-2\\-1&4\end{pmatrix}$</div>
<div style="font-size: 16px;">
などと、一部の列や行だけ取り出すことは出来ませんので気をつけてください。</div>
<div>
<br /></div>
<div style="font-size: 16px;">
この定数倍のことを<b>スカラー倍</b>という言葉で書かれることもあります。</div>
<div style="font-size: 16px;">
<br /></div>
</div>
<div style="font-size: 16px;">
<span style="font-size: large;"><u>行列の単位元</u></span></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
これまで、行列の四則演算のうち、加減乗まで習ったわけですが、</div>
<div style="font-size: 16px;">
今回は除法について説明したいと思います。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
まず、行列の乗法の単位元 $1$ の役目をもつ行列を考えます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$$E=\begin{pmatrix}1&0&\cdots &\cdots&0\\0&1&0&\cdots &0\\\vdots&\ddots&\ddots&\ddots&\vdots\\0&\cdots&0&1&0\\0&\cdots &\cdots&0&1\end{pmatrix}$$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
としましょう。この行列は、対角成分、つまり、全ての $(i,i)$ 成分が $1$ で、<br />
それ以外の全ての成分で 0 となる行列です。</div>
<div style="font-size: 16px;">
このような行列を任意の行列 $A$ にかけてみると、$AE=EA=A$ となることがわかります。</div>
<div style="font-size: 16px;">
行列の単位元は、全ての成分が 1 の行列だと思った人あるかかもしれませんが、<br />
そうではありません。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$2\times 2$行列でやってみます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$$\begin{pmatrix}a&b\\c&d\end{pmatrix}\begin{pmatrix}1&1\\1&1\end{pmatrix}=\begin{pmatrix}a+b&a+b\\c+d&c+d\end{pmatrix}$$</div>
<div style="font-size: 16px;">
となり、確かにこの場合、単位元の役割を果たしていませんね。一方、</div>
<div style="font-size: 16px;">
$$\begin{pmatrix}a&b\\c&d\end{pmatrix}\begin{pmatrix}1&0\\0&1\end{pmatrix}=\begin{pmatrix}a&b\\c&d\end{pmatrix}$$</div>
<div style="font-size: 16px;">
となります。$EA=A$ も同様です。</div>
<div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
この行列を乗法の<b>単位元</b>と言います。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<span style="font-size: large;"><u>逆行列</u></span><br />
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
次に行列における実数の逆数に対応する概念を定義します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$n\times n$ 行列 $A$ の<b>逆行列</b>とは、</div>
<div style="font-size: 16px;">
$AB=E$ かつ $BA=E$</div>
<div style="font-size: 16px;">
を満たす $n\times n$ 行列 $B$ が</div>
<div style="font-size: 16px;">
存在することとして定義します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここで、$AB=E$ となる $B$ を<b>右逆行列</b>、$CA=E$ となる</div>
<div style="font-size: 16px;">
行列を<b>左逆行列</b>と呼ぶことにします。</div>
<div style="font-size: 16px;">
$A$ に逆行列が存在することは、右逆行列と左逆行列がともに存在し、</div>
<div style="font-size: 16px;">
それらが一致するという条件と同値なわけなんですが、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
実は、もし、$A$ に対して右逆行列と左逆行列が存在するなら、</div>
<div style="font-size: 16px;">
右逆行列と左逆行列は一致します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明) $AB=E$ かつ、$CA=E$ となるとします。</div>
<div style="font-size: 16px;">
このとき、$B=(CA)B=C(AB)=C$ となりますので、両者は一致します。(証明終了)</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここで本質的に用いているのは、積の結合法則です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
また、実は、 $A$ に右逆行列が存在するならば、左逆行列が存在することは</div>
<div style="font-size: 16px;">
正しいです。</div>
<div style="font-size: 16px;">
同じように、左逆行列が存在するならば、右逆行列が存在します。</div>
<div style="font-size: 16px;">
ここではこれらのことのみを言及しますが、証明はしません。</div>
<div>
この行列のことを勉強するうちにわかってくると思います。</div>
<div>
<br /></div>
<div>
<br /></div>
<div>
まとめますと、</div>
<div>
正方行列 $A$ に逆行列が存在するとは、</div>
<div>
$AB=BA=E$ となる正方行列 $B$ が存在することを意味します。</div>
<div>
このような $B$ は $A$ から一意的に定まり、それを、$A^{-1}$ と書きます。</div>
<div>
<br /></div>
<div>
なぜ一意的に定まるかというと、</div>
<div>
$AB=BA=E$ となる $B$ として、$B_1,B_2$ が取れたとすると、</div>
<div>
$B_1=B_1AB_2=EB_2=B_2$ となるからです。</div>
<div>
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<u><span style="font-size: large;">逆行列を持たない行列の例</span></u><br />
<div style="font-size: 16px;">
一方、正方行列 $A$ が逆行列が存在しない場合もあります。例えば、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$$\begin{pmatrix}0&0\\1&0\end{pmatrix}\begin{pmatrix}1&0\\0&0\end{pmatrix}=\begin{pmatrix}0&0\\1&0\end{pmatrix}$$</div>
<div style="font-size: 16px;">
となり、$\begin{pmatrix}0&0\\1&0\end{pmatrix}$ に逆行列 $B$ が存在するとすると、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$B\begin{pmatrix}0&0\\1&0\end{pmatrix}=E$ となり、</div>
<div style="font-size: 16px;">
$$B\begin{pmatrix}0&0\\1&0\end{pmatrix}\begin{pmatrix}1&0\\0&0\end{pmatrix}=E\begin{pmatrix}1&0\\0&0\end{pmatrix}=\begin{pmatrix}1&0\\0&0\end{pmatrix}$$</div>
<div style="font-size: 16px;">
$$B\begin{pmatrix}0&0\\1&0\end{pmatrix}\begin{pmatrix}1&0\\0&0\end{pmatrix}=B\begin{pmatrix}0&0\\1&0\end{pmatrix}=\begin{pmatrix}1&0\\0&1\end{pmatrix}$$</div>
<div style="font-size: 16px;">
となり、矛盾します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
また、等式</div>
<div style="font-size: 16px;">
$\begin{pmatrix}1&0\\0&0\end{pmatrix}\begin{pmatrix}0&0\\1&0\end{pmatrix}=O$</div>
<div style="font-size: 16px;">
が成り立つこともあり、両方 $O$ ではない行列 $A,B$ をかけて、$O$ になることがある</div>
<div style="font-size: 16px;">
ということになります。</div>
<div style="font-size: 16px;">
このようなことは、実数や複素数の時にはなかったことに注意しましょう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
この等式からも、$\begin{pmatrix}1&0\\0&0\end{pmatrix}$ や $\begin{pmatrix}0&0\\1&0\end{pmatrix}$ には逆行列が存在しないことを証明できます。</div>
<div style="font-size: 16px;">
$A=\begin{pmatrix}1&0\\0&0\end{pmatrix}$ と $B=\begin{pmatrix}0&0\\1&0\end{pmatrix}$</div>
<div style="font-size: 16px;">
とすると、$AB=O$ が成り立ちますが、</div>
<div style="font-size: 16px;">
$A$ に逆元が存在するとすると、$CA=E$ となる左逆元が存在し、</div>
<div style="font-size: 16px;">
$$CAB=O$$</div>
<div style="font-size: 16px;">
が成り立ちます。一方、</div>
<div style="font-size: 16px;">
$$CAB=EB=B$$</div>
<div style="font-size: 16px;">
にもなります。しかし、明らかに $B\neq O$ ですから、</div>
<div style="font-size: 16px;">
$A$ に逆行列が存在しないことになります。</div>
</div>
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$B$ に逆行列が存在しないことも証明できます。</div>
<div style="font-size: 16px;">
<br /></div>
<u><span style="font-size: large;">$2\times 2$ 行列の逆行列と行列式</span></u><br />
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここでは $2\times 2$ 行列の逆行列を求めてみます。</div>
<div style="font-size: 16px;">
$A=\begin{pmatrix}a&b\\c&d\end{pmatrix}$</div>
<div style="font-size: 16px;">
とし、$X=\begin{pmatrix}x&\ast\\y&\ast\end{pmatrix}$ とします。</div>
<div style="font-size: 16px;">
$\ast$ は、何かの実数が入ると思ってください。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
もし、$AX=E$ を満たすとすると、</div>
<div style="font-size: 16px;">
$$\begin{pmatrix}a&b\\c&d\end{pmatrix}\begin{pmatrix}x&\ast\\y&\ast\end{pmatrix}=\begin{pmatrix}ax+by&\ast\\cx+dy&\ast\end{pmatrix}$$</div>
<div style="font-size: 16px;">
よって、$ax+by=1$ かつ、$cx+dy=0$ を満たします。</div>
<div style="font-size: 16px;">
この連立一次方程式を加減法によって解きます。</div>
<div style="font-size: 16px;">
$adx+bdy=d$</div>
<div style="font-size: 16px;">
$bcx+bdy=0$ </div>
<div style="font-size: 16px;">
ですから、</div>
<div style="font-size: 16px;">
$(ad-bc)x=d$ となります。</div>
<div style="font-size: 16px;">
よって、$ad-bc\neq 0$ であれば、$x=\frac{d}{ad-bc}$ と</div>
<div style="font-size: 16px;">
同様に、$x$ を消すことによって、$y=\frac{-c}{ad-bc}$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
同様に、$X=\begin{pmatrix}\ast&x\\\ast&y\end{pmatrix}$ として、</div>
<div style="font-size: 16px;">
$AX=E$ を満たすとします。</div>
<div style="font-size: 16px;">
同様に</div>
<div style="font-size: 16px;">
$$\begin{pmatrix}a&b\\c&d\end{pmatrix}\begin{pmatrix}\ast&x\\\ast&y\end{pmatrix}=\begin{pmatrix}\ast&ax+by\\\ast&cx+dy\end{pmatrix}$$</div>
<div style="font-size: 16px;">
を満たすので、$ax+by=0$ かつ $cx+dy=1$ となります。</div>
<div style="font-size: 16px;">
これを加減法によって解きます。</div>
<div style="font-size: 16px;">
$adx+bdy=0$</div>
<div style="font-size: 16px;">
$bcx+bdy=b$ </div>
<div style="font-size: 16px;">
となりますから、$(ad-bc)x=-b$ となります。</div>
<div style="font-size: 16px;">
よって、$x=\frac{-b}{ad-bc}$ が成り立ち、</div>
<div style="font-size: 16px;">
$y$ を消すことによって、$y=\frac{a}{ad-bc}$ が成り立ちます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
これにより、共通して、$ad-bc\neq 0$ であるなら、</div>
<div style="font-size: 16px;">
$$X=\frac{1}{ad-bc}\begin{pmatrix}d&-b\\-c&a\end{pmatrix}$$</div>
<div style="font-size: 16px;">
が得られます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
よって、$A=\begin{pmatrix}a&b\\c&d\end{pmatrix}$ の逆行列</div>
<div style="font-size: 16px;">
を作るには、まず、$ad-bc$ が $0$ でなければ、</div>
<div style="font-size: 16px;">
$A$ には逆行列が存在して、</div>
<div style="font-size: 16px;">
$\tilde{A}=\begin{pmatrix}d&-b\\-c&a\end{pmatrix}$</div>
<div style="font-size: 16px;">
を作り、$\frac{1}{ad-bc}\tilde{A}$ を計算することで、$A$ の逆行列を計算することが</div>
<div style="font-size: 16px;">
できます。ここで、$\frac{1}{ad-bc}\tilde{A}$ は、$\tilde{A}$ に</div>
<div style="font-size: 16px;">
$\frac{1}{ad-bc}$ のスカラー倍をすることです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
本当にこの$\frac{1}{ad-bc}\tilde{A}$ が $A$ の逆行列となるかはご自分で確かめてください。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
また、$(2,2)$ 行列 $A$ の $ad-bc$ のことを $\det(A)$ とかいて、</div>
<div style="font-size: 16px;">
<b>行列式</b>と言います。英語では、determinant(ディターミナント)とも言います。</div>
<div style="font-size: 16px;">
略して、デットなど言ったりもします。</div>
<div style="font-size: 16px;">
(逆行列が存在するかどうかの判別をするので判別式とでも</div>
<div style="font-size: 16px;">
言いたいところですが、判別式というと、もうすでに多項式の重解を持つかどうか</div>
<div style="font-size: 16px;">
の式として先取されていますので、通常行列式と言います。)</div>
<div style="font-size: 16px;">
直訳すれば、決定式とも言えますが、日本語ではそれもやはり使われません。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
また、$a+d$ も、$(2,2)$ 行列を調べる上で重要なので、この数を</div>
<div style="font-size: 16px;">
$(2,2)$ 行列 $A$ の<b>トレース</b>、<b>跡</b>などと言い、tr$(A)$ などと書きます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
これまでのところでわかったことは、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$$det(A)\neq 0\Rightarrow A\text{ は逆行列が存在する}$$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
でした。実は、この逆が成り立ちます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
これは簡単に示すことができます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここで、次を示しておく必要があります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
問3.1.6</div>
<div style="font-size: 16px;">
$A,B\in M(2,{\mathbb R})$ であるとすると、</div>
<div style="font-size: 16px;">
$$\det(AB)=\det(A)\det(B)$$</div>
<div style="font-size: 16px;">
が成り立つ。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
これは、$A=\begin{pmatrix}a&b\\c&d\end{pmatrix}$ や $B=\begin{pmatrix}x&z\\y&w\end{pmatrix}$</div>
<div style="font-size: 16px;">
を入れて計算することで直接示すことができます。</div>
<div style="font-size: 16px;">
ここでは省略しておきます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
今、$A$ が逆行列を持つとすると、</div>
<div style="font-size: 16px;">
$AA^{-1}=E$ となります。</div>
<div style="font-size: 16px;">
そこで、この等式の両辺に行列式 $\det$ を取ってみると、</div>
<div style="font-size: 16px;">
$$\det(AA^{-1})=\det(A)\det(A^{-1})=\det(E)=1$$</div>
<div style="font-size: 16px;">
となりますので、特に、$\det(A)\neq 0$ であることがわかります。</div>
<div style="font-size: 16px;">
この等式から、$\det(A^{-1})=\frac{1}{\det(A))}$ であることもわかりますね。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<u><span style="font-size: large;">$n$ 次元ユークリッド空間</span></u><br />
<div style="font-size: 16px;">
2つの集合 $X,Y$ のペアの空間、</div>
<div style="font-size: 16px;">
$$X\times Y=\{(x,y)|x\in X,y\in Y\}$$</div>
<div style="font-size: 16px;">
を<b>直積集合</b>と定義します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
${\mathbb R}\times {\mathbb R}$ を ${\mathbb R}^2$ と書く</div>
<div style="font-size: 16px;">
ことにすれば、これは $\{(x,y)|x,y\in {\mathbb R}\}$ のことですので</div>
<div style="font-size: 16px;">
平面の集合を表します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
${\mathbb R}\times {\mathbb R}\times {\mathbb R}={\mathbb R}^3$ は、</div>
<div style="font-size: 16px;">
$\{(x,y,z)|x,y,z\in {\mathbb R}\}$ のことですので、</div>
<div style="font-size: 16px;">
空間の集合を表します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
次に、$n=4$ の場合はあまり想像ができませんが、</div>
<div>
<div style="font-size: 16px;">
${\mathbb R}^4$ は、$\{(x,y,z,w)|x,y,z,w\in {\mathbb R}\}$ のことであり、4次元空間のことを指します。</div>
</div>
<div>
<br /></div>
<div style="font-size: 16px;">
これ以上はさらに想像を超えるので、想像することを諦めて、実数が $n$ 個並んだもの</div>
<div style="font-size: 16px;">
を全て集めた集合として一般に、${\mathbb R}^n={\mathbb R}\times {\mathbb R}\times \cdots\times {\mathbb R}$</div>
<div style="font-size: 16px;">
として定義することで、</div>
<div style="font-size: 16px;">
${\mathbb R}^n$ を、$n$ この実数のペアを集めたものを考えます。</div>
<div style="font-size: 16px;">
つまり、$\{(x_1,x_2,\cdots, x_n)|x_i\in {\mathbb R}\}$ となる集合です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
今、$(x_1,x_2,\cdots, x_n)$ のことを、${\mathbb R}^n$ の元として、縦に</div>
<div style="font-size: 16px;">
$$\begin{pmatrix}x_1\\x_2\\\vdots\\x_n\end{pmatrix}$$</div>
<div style="font-size: 16px;">
のように並べて書くことにします。</div>
<div style="font-size: 16px;">
これは、横に書いていたものと本質的に同じものです。</div>
<div style="font-size: 16px;">
つまり今は、</div>
<div style="font-size: 16px;">
$${\mathbb R}^n=\left\{\begin{pmatrix}x_1\\x_2\\\vdots\\x_n\end{pmatrix}|x_i\in {\mathbb R}\right\}$$</div>
<div style="font-size: 16px;">
となるわけです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
この空間 ${\mathbb R}^n$ を平面ベクトルや空間ベクトルの一般化として考えたいです。</div>
<div style="font-size: 16px;">
つまり、${\mathbb R}^n$ の元をベクトルと考えたいのです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
そのために、</div>
<div style="font-size: 16px;">
$${\bf x}=\begin{pmatrix}x_1\\x_2\\\vdots\\x_n\end{pmatrix},{\bf y}=\begin{pmatrix}y_1\\y_2\\\vdots\\y_n\end{pmatrix}\in {\mathbb R}^n$$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
に対して、その和を</div>
<div style="font-size: 16px;">
$${\bf x}+{\bf y}=\begin{pmatrix}x_1+y_1\\x_2+y_2\\\vdots\\x_n+y_n\end{pmatrix}$$</div>
<div style="font-size: 16px;">
のように定義します。</div>
<div style="font-size: 16px;">
また、定数倍(スカラー倍)を</div>
<div style="font-size: 16px;">
$$\alpha\cdot \begin{pmatrix}x_1\\x_2\\\vdots\\x_n\end{pmatrix}=\begin{pmatrix}\alpha x_1\\\alpha x_2\\\vdots\\\alpha x_n\end{pmatrix}$$</div>
<div style="font-size: 16px;">
として定義します。</div>
<div style="font-size: 16px;">
この和とスカラー倍は、$(n,1)$ 行列としての和とスカラー倍と考えることもできます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このように、和とスカラー倍が定義された直積集合 ${\mathbb R}^n$ を</div>
<div style="font-size: 16px;">
<b>$n$ 次元ユークリッド空間</b>といいます。</div>
<div style="font-size: 16px;">
<b>$n$ 次元数ベクトル空間</b>ともいいます。</div>
<div style="font-size: 16px;">
このように縦に並んだ形の ${\mathbb R}^n$ の元のことを(<b>縦</b>)<b>ベクトル</b>と言います。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
次を定義しましょう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
定義3.1.1</div>
<div style="font-size: 16px;">
$n$ 次元ユークリッド空間 ${\mathbb R}^n,{\mathbb R}^m$ に</div>
<div style="font-size: 16px;">
対して、写像 $f:{\mathbb R}^n\to {\mathbb R}^m$ が</div>
<div style="font-size: 16px;">
以下を満たすとき、$f$ を<b>線形写像</b>と言います。</div>
<div style="font-size: 16px;">
${\bf x},{\bf y}$ を ${\mathbb R}^n$ の任意のベクトルとし、</div>
<div style="font-size: 16px;">
$\alpha$ を任意の実数とします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(1) $f({\bf x})+{\bf y})=f({\bf x})+f({\bf y})$</div>
<div style="font-size: 16px;">
(2) $f(\alpha \cdot {\bf x})=\alpha\cdot f({\bf x})$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここで、行列を用いた、線形写像を考えます。</div>
<div style="font-size: 16px;">
$A\in M(m,n,{\mathbb R})$ とします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このとき、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
${\bf x}\in {\mathbb R}^n$ とすると、</div>
<div style="font-size: 16px;">
$f_A:{\mathbb R}^n\to {\mathbb R}^m$ を ${\bf x}\mapsto A\cdot {\bf x}$</div>
<div style="font-size: 16px;">
として定義します。</div>
<div style="font-size: 16px;">
ここで、$A\cdot {\bf x}$ は、$(m,n)$ 行列と $(n,1)$ 行列の行列の積として</div>
<div style="font-size: 16px;">
考えてください。</div>
<div style="font-size: 16px;">
このとき、$A\cdot {\bf x}$ は、$M(m,1,{\mathbb R})$ つまり、${\mathbb R}^m$ となります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このとき、実は $f_A$ は線形写像となります。</div>
<div style="font-size: 16px;">
というのも、行列の分配法則を用いて、</div>
<div style="font-size: 16px;">
$f_A({\bf x}+{\bf y})=A({\bf x}+{\bf y})=A{\bf x}+A{\bf y}=f({\bf x})+f({\bf y})$</div>
<div style="font-size: 16px;">
また、</div>
<div style="font-size: 16px;">
$f_A(\alpha\cdot {\bf x})=A(\alpha{\bf x})=\alpha\cdot A{\bf x}=\alpha f({\bf x})$</div>
<div style="font-size: 16px;">
となります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このとき、次が成り立ちます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
命題3.1.1</div>
<div style="font-size: 16px;">
$f:{\mathbb R}^n\to {\mathbb R}^m$ が線形写像であるとすると、</div>
<div style="font-size: 16px;">
ある $(m,n)$ 行列 $A$ が存在して、$f=f_A$ が成り立つ。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このことの証明は教科書にありますが、</div>
<div style="font-size: 16px;">
簡単に説明しますと</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ベクトル ${\bf e}_i\in {\mathbb R}^n$ をその $i$ 番目が $1$ で、それ以外は $0$ となる</div>
<div style="font-size: 16px;">
ベクトルとします。このとき、</div>
<div style="font-size: 16px;">
$$f({\bf e}_i)=\begin{pmatrix}a_{1i}\\\vdots\\a_{ni}\end{pmatrix}$$</div>
<div style="font-size: 16px;">
とします。そうすると、</div>
<div style="font-size: 16px;">
$${\bf x}=\begin{pmatrix}x_{1}\\\vdots\\x_{n}\end{pmatrix}=x_1{\bf e}_1+\cdots+x_n{\bf e}_n$$</div>
<div style="font-size: 16px;">
となり、$f$ の線形性から、</div>
<div style="font-size: 16px;">
$$f({\bf x})=f(x_1{\bf e}_1+\cdots+x_n{\bf e}_n)=x_1f({\bf e}_1)+\cdots+x_nf({\bf e}_n)$$</div>
<div style="font-size: 16px;">
$$=\begin{pmatrix}a_{11}&\cdots &a_{1n}\\\cdots&\cdots&\cdots\\a_{n1}&\cdots&a_{nn}\end{pmatrix}\begin{pmatrix}x_1\\\vdots\\x_n\end{pmatrix}$$</div>
<div style="font-size: 16px;">
$$=A{\bf x}=f_A({\bf x})$$</div>
<div style="font-size: 16px;">
となります。ここで、$A$ は $A=(a_{ij})$ となる $(m,n)$ 行列となります。(証明終了)</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
よって、$f$ が線形写像であることと、それがある行列を用いて</div>
<div style="font-size: 16px;">
${\bf x}\mapsto A{\bf x}$ のように書かれることは同値であるということになります。</div>
<div style="font-size: 16px;">
つまり、それらは全く同じものを考えているということを意味します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
線形写像 $f:{\mathbb R}^n\to {\mathbb R}^m$ と $g:{\mathbb R}^m\to {\mathbb R}^k$</div>
<div style="font-size: 16px;">
があるときに、合成写像 $g\circ f:{\mathbb R}^n\to {\mathbb R}^k$ が得られます。</div>
<div style="font-size: 16px;">
$f=f_A$ であり、$g=f_B$ となる行列 $A\in M(m,n,{\mathbb R})$ かつ $B\in M(k,m,{\mathbb R})$</div>
<div style="font-size: 16px;">
となります。$f$ と $g$ の合成も線形写像になります。(証明してみてください。)</div>
<div style="font-size: 16px;">
なので、命題3.1.1を用いることで、$f_B\circ f_A$ はある行列 $C\in M(k,n,{\mathbb R})$</div>
<div style="font-size: 16px;">
が存在して、$f_B\circ f_A=f_{C}$ となります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
${\bf x}\in {\mathbb R}^n$ に対して、</div>
<div style="font-size: 16px;">
$$f_{B}\circ f_A({\bf x})=f_B(A{\bf x})=B(A{\bf x})=(BA){\bf x}$$</div>
<div style="font-size: 16px;">
となるので、$f_B\circ f_A=f_{BA}$ となります。</div>
<div style="font-size: 16px;">
つまり、行列の積は自然に線形写像の合成を意味するのです。</div>
<div>
<br /></div>
<div style="font-size: 16px;">
$n=m$ のとき、線形写像 $g:{\mathbb R}^n\to {\mathbb R}^n$ を<b>線形変換</b>もしくは<b>一次変換</b></div>
<div style="font-size: 16px;">
といいます。</div>
<div style="font-size: 16px;">
この時、命題3.1.1を使い、正方行列 $A$ が存在して、</div>
<div style="font-size: 16px;">
$f({\bf x})=A{\bf x}$ が成り立ちます。</div>
<div style="font-size: 16px;">
線形変換 $f_A$ と $f_B$ の合成は、$f_B\circ f_A=f_{BA}$ が成り立ちます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
単位行列 $E$ に対して、$f_E=\text{id}_{{\mathbb R}^n}$ が成り立つことに</div>
<div style="font-size: 16px;">
注意しておきます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$A$ が逆行列 $A^{-1}$ が存在するとき、</div>
<div style="font-size: 16px;">
$$f_A\circ f_{A^{-1}}=f_{AA^{-1}}=f_E=\text{id}_{{\mathbb R}^n}$$</div>
<div style="font-size: 16px;">
$$f_{A^{-1}}\circ f_A=f_{A^{-1}A}=f_E=\text{id}_{{\mathbb R}^n}$$</div>
<div style="font-size: 16px;">
よって、$f_{A^{-1}}$ は $f_A$ の逆写像となります。</div>
<div style="font-size: 16px;">
逆写像に関しては、<a href="https://motochans.blogspot.com/2020/04/12.html">第2回</a>を見てください。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
つまり、</div>
<div style="font-size: 16px;">
次は同値となります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(i) $f_A$ は逆写像を持つ</div>
<div style="font-size: 16px;">
(ii) $A$ が逆行列を持つ</div>
<div style="font-size: 16px;">
(iii) $f_A$ が全単射である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ちなみに、(i) と(iii)の同値性は、第2回で証明をしました。</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-23979474585407370212020-05-13T11:46:00.001+09:002022-04-19T10:14:36.646+09:00数学リテラシー1(第3回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(水曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
</div>
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="text-align: left;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
数学リテラシー3回目です。<br />
<br />
今日から行列や一次変換についての説明に入ります。<br />
行列というのは、高校では習わなかった対象だと思います。<br />
(一部の高等学校では教育課程を逸脱して教えているところもあるようですが...)<br />
しかし、行列は大学では線形代数の話でさらっと登場し、<br />
その後の数学の学習にはなくてはならない対象になります。<br />
数学に限らず、科学の基礎の部分でどの分野に進んでも<br />
そのテクニックを使うことになるでしょう。<br />
<br />
私が受験生のころは行列も高校までの範囲に入っていましたので<br />
「代数・幾何」という授業のなかで行列や一次変換を<br />
勉強をしたことを鮮明に覚えています。ちょっと変わった代数だなと思っただけで、<br />
その意味まではよくわかっていませんでした。<br />
しかし、大学に入って線形代数を系統的に学び、連立一次方程式を解いたり、<br />
固有値を用いて、いとも簡単に線形常微分方程式が解けたりなど<br />
様々な応用があることが分かってその奥深さに面白さを感じました。<br />
<br />
2次関数を平行移動したりすることは今でも高校で習うと思いますが、<br />
一方、行列を使えば、図形を回転させた時にそれがどのように移るか?<br />
ということも調べることができます。<br />
なので、これからの講義を聞いた後であれば、<br />
例えば、$y=x^2$ という2次関数を30度回転させたときに<br />
図形がどのような方程式を満たすかなど計算できるようになります。<br />
<br />
今日は、行列の四則演算について学びますが、<br />
今日以降の学習のどこかで、図形を変形させる変換として考え<br />
たりもします。<br />
そのような変換のことを<b>一次変換</b>といいます。<br />
<br />
行列を使うことによって、図形を拡大させたり、回転させたり、<br />
相似拡大をしたりするのも一次変換です。<br />
また、ある方向に歪ませたりする操作、<br />
たとえば、正方形を平行四辺形に歪ませるようなことですが、<br />
これも一次変換です。<br />
これらの変換を組み合わせたり、合成したりしたものも一次変換です。<br />
(実は逆に、行列とは、上のような操作を合成したものと考えることもできるのですが<br />
それはどこかでおいおいと...)<br />
<div>
<br /></div>
どのように変換するかを記述するために行列が必要となります。<br />
<br />
今回は行列についての最初の基本的な知識についてです。<br />
全体で7つに分かれています。<br />
では、はじめていきましょう。<br />
<br />
<span style="font-size: large;"><u>Part 1: 行列の定義、行列の成分</u></span><br />
<br />
<b>行列</b>とは縦に $m$ 個、横に $n$ 個だけ長方形状に数を並べたものをいいます。<br />
<div>
そのような行列のことを<b>$m$ 行 $n$ 列の行列</b>といったり、</div>
<div>
<b>$m\times n$ </b><b>行列</b>といったり、$(m,n)$ <b>行列</b>と言ったりします。</div>
下の例では、縦に3 個、横に4個の数を長方形の形に並べたものですから、<br />
$(3,4)$ 行列の例です。<br />
<br />
$$\begin{pmatrix}1&2&3&4\\5&6&7&8\\9&10&11&12\end{pmatrix}$$<br />
<br />
このとき、<b>第1行目</b>とは、<br />
$$\begin{pmatrix}1&2&3&4\end{pmatrix}$$<br />
をさし、<br />
<b>第2行目</b>とは、<br />
$$\begin{pmatrix}5&6&7&8\end{pmatrix}$$<br />
<div>
のことをさします。第 $3$ 行目はも同様に考えれば<br />
もうわかりますね。</div>
<div>
<br /></div>
<div>
また、<b>第1列目</b>といえば、</div>
<div>
$$\begin{pmatrix}1\\5\\9\end{pmatrix}$$</div>
<div>
のことであり、<b>第3列目</b>といえば、</div>
<div>
<div>
$$\begin{pmatrix}3\\7\\11\end{pmatrix}$$</div>
</div>
<div>
のことを指します。</div>
<div>
このように、<b>行</b>といえば、横に並んだ数のことをいい、</div>
<div>
<b>列</b>といえば、縦に並んだ数のことをいいます。</div>
<div>
<br /></div>
<div>
また、とりわけ、$m=n$ のときは、正方形的に数が並ぶため、</div>
<div>
<b>正方行列</b>といいます。この場合、$n$ <b>次正方行列</b>ともいいます。</div>
<div>
<br /></div>
<div>
<div>
行列に並んでいる、$mn$ 個のそれぞれの数のことを<b>成分</b>と言います。</div>
<div>
</div>
$m$ 行 $n$ 列の行列の場合、上から数えて $i$ 番目、左から数えて</div>
<div>
$j$ 番目にいる成分を、$(i,j)$ <b>成分</b>といいます。</div>
<div>
<br /></div>
<div>
例えば、上の $3\times 4$ 行列の $(1,3)$ 成分は $3$ ということになります。</div>
<div>
同じように考えることで、$(3,4)$ 成分は、$12$ということになります。</div>
<div>
<br /></div>
<div>
まずは、このような言い方に慣れてください。</div>
<div>
例えば、以下はうちにある衣装ケースですが、これは、$3\times 2$ 行列</div>
<div>
と考えられます。</div>
<div>
黄色いシールは成分の名前(例えば $(1,2)$ を書いており、貼り付けています。</div>
<div>
あなたのうちにもこのようなケースやタンスがあれば、一つ一つ指をさして、<br />
ここが、$(1,2)$ 成分、ここが、$(2,3)$ 成分などと確認しながら</div>
<div>
言ってみるのもよいでしょう。</div>
<div>
<br /></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh1fkVPyTQVKqHczznKJMJ5VY1ubJlUviyoxiMUFWI3_EtfzOd5YETnE-ddEHjcLPhsjID3YDVzevwuyI7QgQuWH62POd_fVdeESUigcHcH3AT3dJiD0jFM2V2wqBCavTRZJIifDfGDa5fM/s1600/WIN_20200512_17_17_33_Pro.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1558" data-original-width="1600" height="194" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh1fkVPyTQVKqHczznKJMJ5VY1ubJlUviyoxiMUFWI3_EtfzOd5YETnE-ddEHjcLPhsjID3YDVzevwuyI7QgQuWH62POd_fVdeESUigcHcH3AT3dJiD0jFM2V2wqBCavTRZJIifDfGDa5fM/s200/WIN_20200512_17_17_33_Pro.jpg" width="200" /></a></div>
<div>
また、$A$ という行列の $(i,j)$ 成分を $a_{ij}$ のように書くことがあります。</div>
<div>
$3\times 4$ 行列の場合、</div>
<div>
<br /></div>
<div>
$$\begin{pmatrix}a_{11}&a_{12}&a_{13}&a_{14}\\a_{21}&a_{22}&a_{23}&a_{24}\\a_{31}&a_{32}&a_{33}&a_{34}\end{pmatrix}$$</div>
<div>
となります。この並びを見ながら、2つの数字の組み合わせ $(i,j)$ がどのように<br />
移り変わっているかを追ってみると、数え方がわかるようになるかもしれません。</div>
<div>
<div>
このような行列 $A$ を、$(a_{ij})$ のように書くことがあります。</div>
</div>
<div>
<br /></div>
<div>
<span style="font-size: large;"><u>Part 2: 行列の略記、行列全体の集合</u></span></div>
<div>
さきほど、最後に書いたように、$A$ の $(i,j)$ 成分が $a_{ij}$ となる行列</div>
<div>
を $(a_{ij})$ のように書きました。</div>
<div>
<br /></div>
<div>
他の書き方として、$(a_{ij})_{1\le i,j\le n}$ のように書くこともあります。</div>
<div>
また、$i$ と $j$ の間にコンマを入れて、$a_{i,j}$ のように書くこともあります。</div>
<div>
これは、例えば、$a_{121}$ と書いてしまったとき、$(12,1)$ 成分なのか、$(1,21)$ 成分なのかわからないからですが、数字が2つまでしか並ばないのなら<br />
コンマはなくてもわかります。</div>
<div>
<br /></div>
<div>
また、$n=1$ のとき、$m\times 1$ 行列は、縦に並んだ一直線の行列<br />
$$\begin{pmatrix}a_{11}\\\vdots\\a_{m1}\end{pmatrix}$$</div>
<div>
ですが、このような行列をベクトルということがあります。<br />
とりわけ、この場合は縦ベクトルといいます。<br />
<br />
また、$m=1$ のとき、$1\times n$ 行列とは、横に並んだベクトル</div>
<div>
$$\begin{pmatrix}a_{11}&\cdots&a_{1n}\end{pmatrix}$$ </div>
<div>
を表すことになります。<br />
こちらは横ベクトルともいいます。<br />
<br /></div>
<div>
また、普通の数 $1$ や $\sqrt{2}$ なども、カッコをつけて、</div>
<div>
$(1)$ や $(\sqrt{2})$ のように表しておくことで $(1,1)$ 行列ということもあります。<br />
この場合、カッコをつけているだけで、実数を考えていることと<br />
何ら変わりはありません。</div>
<div>
<br /></div>
<div>
$(m,n)$ 行列を全て集めることで、集合を作ることができます。</div>
<div>
この集合、つまり、</div>
<div>
$$\{A|A\text{ は全ての成分が実数となる $m\times n$ 行列}\}$$</div>
<div>
を $M(m,n,{\mathbb R})$ と書きます。成分が全て実数の行列の集合ですのでこのように</div>
<div>
書きますが、成分が複素数であれば、${\mathbb C}$ を用いて、$M(m,n,{\mathbb C})$</div>
<div>
となります。<br />
正方行列の場合は、$M(n,n,{\mathbb R})$ のように $n$ を2回続けて書く必要もない<br />
と感ずれば、$M(n,{\mathbb R})$ と書くこともあります。</div>
<div>
<br /></div>
<div>
<span style="font-size: large;"><u>Part 3: 行列の相等・行列の加法、零行列</u></span></div>
<div>
次に、行列の四則演算について行います。</div>
<div>
これ以降、行列の成分は全て実数として扱います。</div>
<div>
複素数や他の数体を用いてもよいですが、本質的には違いはありません。</div>
<div>
<br /></div>
<div>
実数の次の四則演算(加法、減法、乗法、除法)を思い出しましょう。</div>
<div>
四則演算とは、</div>
<div>
$$a+b,\ a-b,\ a\times b,\ a\div b$$</div>
<div>
のことでした。これらの法則は、</div>
<div>
$a+b=b+a$ (<b>加法の交換法則</b>)</div>
<div>
$(a+b)+c=a+(b+c)$ (<b>分配法則)</b></div>
<div>
$a\times (b+c)=a\times b+a\times c$ (<b>加法の結合法則</b>)</div>
<div>
を満たします。</div>
<div>
<br /></div>
<div>
また、実数 $0$ は、$0+a=a+0=a$ などの性質を満たします。</div>
<div>
このような性質をもつ $0$ のことを<b>零元</b>もしくは<b>加法の単位元</b>と言います。</div>
<div>
そのような単位元が存在することを、<b>加法の単位元の存在</b>といいます。</div>
<div>
<br /></div>
<div>
また、実数 $a$ に対して、 $-a$ とは、$a+(-a)=0$ を満たす実数を表します。</div>
<div>
たとえば、$a=3$ に対しては、$-a$ とは、$3$ のことを表し、</div>
<div>
$a=-2$ に対して、$-a=2$ とし、$a=0$ に対しては、$-a=0$ とすると、<br />
上の関係 $a+(-a)=0$ が成り立ちます。</div>
<div>
$a$ に対して、$-a$ が存在することになりますが、<br />
$a$ に対する $-a$ のことを $a$ の<b>逆元</b>といいます。</div>
<div>
このように各 $a$ に対して $-a$ が存在することを <b>加法の逆元の存在</b><br />
といいます。</div>
<div>
<br /></div>
<div>
今、$A,B\in M(m,n,{\mathbb R})$ に対して、$A$ と $B$ の加法を定義します。</div>
<div>
$A=(a_{ij})$ と $B=(b_{ij})$ とします。</div>
<div>
この記号は、$(i,j)$ 成分がそれぞれ、$a_{ij}$ であり、$b_{ij}$ となる</div>
<div>
行列ということでした。</div>
<div>
このとき、$(m,n)$ 行列 $A+B$ を $(a_{ij}+b_{ij})$ と定義します。</div>
<div>
<br /></div>
<div>
つまり、成分ごとに和を取るということです。</div>
<div>
例えば、</div>
<div>
$$A=\begin{pmatrix}1&2&3\\4&5&6\end{pmatrix},\ B=\begin{pmatrix}7&8&9\\10&11&12\end{pmatrix}$$</div>
<div>
とすると、</div>
<div>
$$A+B=\begin{pmatrix}1+7&2+8&3+9\\4+10&5+11&6+12\end{pmatrix}$$</div>
<div>
$$=\begin{pmatrix}8&10&12\\14&16&18\end{pmatrix}$$</div>
<div>
となります。</div>
<div>
<br /></div>
<div>
このような成分同士足して得られる和は、ベクトルと同じですね。</div>
<div>
2つのベクトル $(a,b,c)$ と $(d,e,f)$ の和は、$(a+d,b+e,c+f)$ のように</div>
<div>
成分同士たすのでした。</div>
<div>
<br /></div>
<div>
つまり、行列は、$mn$ 個の実数の集まりをあたかもベクトルだと思って</div>
<div>
和をとったものが行列の和ということになります。</div>
<div>
ですので、<u>2つの行列の和ができるためには、</u><br />
<u>その行列の2つのサイズが一致していないといけません</u>。</div>
<div>
つまり、$A,B$ がどちらも $m\times n$ 行列でないと、$A+B$ ができません。</div>
<div>
<br /></div>
<div>
上でいう、実数は加法の単位元が存在することから、<br />
行列も加法の逆元が存在します。<br />
全ての成分が0になる行列を</div>
<div>
<br /></div>
<div>
$$O=\begin{pmatrix}0&0&\cdots& 0\\\cdots &\cdots&\cdots&\cdots \\0&0&\cdots &0\end{pmatrix}$$</div>
<div>
として定義してやると、これは、ベクトルでいう零ベクトルであることに対応し、<br />
$A+O=O+A=A$ が成り立ちます。<br />
この行列のことを $O$ と<b>零行列</b>といいます。</div>
<div>
このように単に $O$ として書くと、サイズがどのような零行列かわからないと思う</div>
<div>
かもしれませんが、$A$ と同じサイズの全ての成分が $0$ の行列ということになります。</div>
<div>
ややこしくなる場合なら、$O_{m,n}$ のように、しておけば、$(m,n)$ 行列の</div>
<div>
零行列とするのが良いかもしれません。</div>
<div>
<br /></div>
<div>
また、$A\in M(m,n,{\mathbb R})$ に対して、$-A$ を全ての成分が$A$ の成分の</div>
<div>
マイナス1倍として定義しておきます。<br />
つまり、$A=(a_{ij})$ としたときに、$-A$ を $(-a_{ij})$ として定義するのです。</div>
<div>
<div>
このとき、成分同士の和をとることで、$A+(-A)=O$ が成り立ちます。<br />
<br /></div>
<div>
</div>
このようにすると、実数の加法、減法について以下が成り立つことがわかりました。</div>
<div>
<br /></div>
<div>
$A+B=B+A$, </div>
<div>
$A+(B+C)=(A+B)+C$, </div>
<div>
$A+O=O+A=A$,<br />
$A+(-A)=O$</div>
<div>
<br /></div>
<div>
また、行列の相等というのは、2つの行列が等しい時はいつか?</div>
<div>
という問題ですが、$A=(a_{ij})$ と$B=(b_{ij})$ が</div>
<div>
等しいということは、各成分が等しいということ、</div>
<div>
つまり、任意の $i,j$ に対して、$a_{ij}=b_{ij}$ が成り立つこと</div>
<div>
として定義します。</div>
<div>
<br /></div>
<div>
<u><span style="font-size: large;">Part 4: 行列の乗法</span></u></div>
<div>
次に行列の乗法について説明をします。</div>
<div>
ベクトルには、行列の加法と乗法がありましたが、乗法は</div>
<div>
ありませんでした。</div>
<div>
しかし、行列には乗法を考えることができます。</div>
<div>
ただし、しかるべき条件が必要です。</div>
<div>
<br /></div>
<div>
$A\in M(m,n,{\mathbb R})$ と、$B\in M(n,k,{\mathbb R})$ に対して、</div>
<div>
その積 $A\cdot B$ を定義することができます。<br />
この積を表すドット $\cdot$ は以下しばしば省略されることがあります。</div>
<div>
<br /></div>
<div>
ここで、$A$ の列の数と $B$ の行の数が一致していることが条件です。</div>
<div>
ここで得られる $A\cdot B=C$ は、$(m,k)$ 行列とになります。</div>
<div>
丁度、2つの行列のサイズでかぶっていた $n$ が消去されて、</div>
<div>
$C$ の行数は $A$ の行数であり、$C$ の列数は $B$ の列数となります。</div>
<div>
<br />
<div>
$(m,n)$ 行列 $A$ と $(n,k)$ 行列 $B$ に対して、$(m,k)$ 行列</div>
<div>
$C=A\cdot B$ の、$(i,j)$ 成分を</div>
<div>
$$\sum_{p=1}^na_{ip}b_{pj}$$</div>
<div>
として定義します。つまり、公式として書くならば</div>
<div>
$$C=(\sum_{p=1}^na_{ip}b_{pj})$$</div>
<div>
ということになります。</div>
<div>
<br /></div>
<div>
積の公式はこれですが、実際理解するには、何回か</div>
<div>
積の練習をする必要があります。<br />
</div>
<br /></div>
<div>
例えば、$m=n=k=3$ の時を考えてみましょう。</div>
<div>
<br /></div>
<div>
$A=(a_{ij})$ とし、$B=(b_{ij})$ とします。</div>
<div>
このとき、</div>
<div>
$$C=A\cdot B=\begin{pmatrix}a_{11}&a_{12}&a_{13}\\a_{21}&a_{22}&a_{23}\\a_{31}&a_{32}&a_{33}\end{pmatrix}\begin{pmatrix}b_{11}&b_{12}&b_{13}\\b_{21}&b_{22}&b_{23}\\b_{31}&b_{32}&b_{33}\end{pmatrix}$$</div>
<div>
を計算します。</div>
<div>
$C$ もまた、$(3,3)$ 行列となるのですが、</div>
<div>
その、$(i,j)$ 成分がどのようになるかをみてみましょう。</div>
<div>
<br /></div>
<div>
例えば、$(1,2)$ 成分ですが、 $A$ の第1行目、 $B$ の第2列目をとります。</div>
<div>
それぞれ $\begin{pmatrix}a_{11}&a_{12}&a_{13}\end{pmatrix}$ と$\begin{pmatrix}b_{21}\\b_{22}\\b_{23}\end{pmatrix}$ となりますが、</div>
<div>
これらの内積を考えます。</div>
<div>
つまり、対応する成分同士の積をとり、足し上げるのです。</div>
<div>
結果的に、</div>
<div>
$$a_{11}b_{21}+a_{12}b_{22}+a_{13}b_{32}$$</div>
<div>
となります。</div>
<div>
これが、$C$ の $(1,2)$ 成分となります。</div>
<div>
同じように続けます。<br />
他にも、例えば、$C$ の $(2,3)$ 成分を計算するには、</div>
<div>
$A$ の第2行目と $B$ の第3列目をとります。</div>
<div>
<div>
このとき、それぞれ $\begin{pmatrix}a_{21}&a_{22}&a_{23}\end{pmatrix}$ と $\begin{pmatrix}b_{13}\\b_{23}\\b_{33}\end{pmatrix}$ となりますが、</div>
</div>
<div>
この対応する成分同士の積を考えてそれらを全て足すと、</div>
<div>
$C$ の第 $(2,3)$ 成分は、</div>
<div>
$$a_{21}b_{13}+a_{22}b_{23}+a_{23}b_{33}$$</div>
<div>
となります。つまり、$(i,j)$ 成分は、</div>
<div>
$$a_{i1}b_{1j}+a_{i2}b_{2j}+ a_{i3}b_{3j}=\sum_{k=1}^3a_{ik}b_{kj}$$</div>
<div>
となるわけです。</div>
<div>
<br /></div>
<div>
講義の方にも、教科書の方でも問題がいくつかあるので自分で練習してみてください。</div>
<div>
<br /></div>
<div>
例えば、</div>
<div>
<br /></div>
<div>
$$\begin{pmatrix}1&2&-1\\2&-4&3\end{pmatrix}\begin{pmatrix}3&2\\3&4\\2&3\end{pmatrix}=\begin{pmatrix}7&7\\0&-3\end{pmatrix}$$</div>
<div>
<br /></div>
<div>
などとなります。この計算も自分で確かめてみてください。</div>
<div>
<br /></div>
<div>
また、この行列の積の順番を入れ替えたとき、</div>
<div>
<br /></div>
<div>
<div>
$$\begin{pmatrix}3&2\\3&4\\2&3\end{pmatrix}\begin{pmatrix}1&2&-1\\2&-4&3\end{pmatrix}=\begin{pmatrix}7&-2&3\\11&-10&9\\8&-8&7\end{pmatrix}$$</div>
</div>
<div>
<br /></div>
<div>
となります。この例から分かる通り、積に関して $AB=BA$ が成り立ちません。</div>
<div>
そもそも行列のサイズさえ合っていません。<br />
<br />
通常の実数は積は入れ替えても答えは同じですからそのことを<br />
実数の積は<b>可換</b>であるといいます。<br />
可換ではないような積をもつ場合は<b>非可換</b>といいます。<br />
上で見た通り、行列の積は非可換です。<br />
<br />
しかし、その理由が行列のサイズの問題だけであるとすれば、<br />
行も列の一致している正方行列のときを考えるとどうでしょうか?<br />
<br />
ところが、たとえ、$A,B$ が正方行列であっても可換ではありません。</div>
<div>
例えば、</div>
<div>
$A=\begin{pmatrix}1&1\\0&1\end{pmatrix},\ B=\begin{pmatrix}1&0\\1&1\end{pmatrix}$</div>
<div>
とすると、</div>
<div>
$$AB=\begin{pmatrix}2&1\\1&1\end{pmatrix}$$</div>
<div>
<div>
$$BA=\begin{pmatrix}1&1\\1&2\end{pmatrix}$$</div>
</div>
<div>
<br /></div>
<div>
となり、やはり、$AB\neq BA$ となります。</div>
<div>
よって、正方行列だけに制限したとしても、行列は積に関して非可換である<br />
ということになります。</div>
<div>
<br /></div>
<div>
しかし、全ての $A,B$ に対していつでも $AB\neq BA$ というわけではありません。</div>
<div>
たとえば、次のような例を</div>
<div>
<br /></div>
<div>
$A=\begin{pmatrix}1&2\\2&1\end{pmatrix}$</div>
<div>
$B=\begin{pmatrix}1&-1\\-1&1\end{pmatrix}$ とすると</div>
<div>
$$AB=BA=\begin{pmatrix}-1&1\\1&-1\end{pmatrix}$$</div>
<div>
となり、この場合は2つの行列は可換となります。</div>
<div>
<br /></div>
<div>
ある積の演算を持った集合(例えば整数や実数や正方行列など)<br />
が積において可換であるというのは、</div>
<div>
そのような集合の全ての元 $A,B$ に対していつでも可換、つまり $AB=BA$ を</div>
<div>
満たさなければならないということに注意してください。<br />
<br />
非可換であるということは、つまり可換を否定しているワケだから、<br />
最初の論理の時にもやりましたが、<br />
一つでも非可換な $A,B$ があれば、その集合は非可換であるということになります。<br />
なので、上で非可換な $A,B$ の例を挙げたことから、<br />
行列 $M(2,{\mathbb R})$ は非可換な積をもつ集合であるということになります。<br />
実際、$n>1$ であれば、$M(n,{\mathbb R})$ は可換ではありません。</div>
<div>
<br /></div>
<div>
<u><span style="font-size: large;">Part 5: 行列の結合法則、分配法則</span></u></div>
<div>
行列は、実数の加法の性質から、自然に行列の加法に関するいくつかの法則が</div>
<div>
導かれました。</div>
<div>
加法の可換性や、加法の結合法則や加法の単位元の存在などです。</div>
<div>
<br /></div>
<div>
今、上で行列の積を定義しましたが、実数の積の法則で、</div>
<div>
行列の積でも同じ法則が成り立つ部分がありますのでそれを説明しておきます。<br />
<br />
上では実数では成り立つ、積の可換性は行列では成り立たない</div>
<div>
ことも示しました。</div>
<div>
それ以外の法則をみてみますと、</div>
<div>
<br /></div>
<div>
$a,b,c$ を実数とするとき、</div>
<div>
$(ab)c=a(bc)$ (積の結合法則)</div>
<div>
$(a+b)c=ac+bc$ (分配法則1)</div>
<div>
$a(b+c)=ab+ac$ (分配法則2)</div>
<div>
が成り立ちますが、実は、行列でも、<br />
<br />
$A\in M(m,n,{\mathbb R})$ かつ $B\in M(n,l,{\mathbb R})$ かつ $C\in M(l,p,{\mathbb R})$ であるとすると、<br />
$A(BC)=(AB)C$ (積の結合法則)<br />
<br />
$A,B\in M(m,n,{\mathbb R})$ かつ $C\in M(n,l,{\mathbb R})$ であるとすると、<br />
<div>
$(A+B)C=AC+BC$ (分配法則1)</div>
<div>
が成り立ちます。</div>
<div>
<br /></div>
<div>
$A\in M(m,n,{\mathbb R})$ かつ $B,C\in M(n,l,{\mathbb R})$ であるとすると、<br />
<div>
$A(B+C)=AB+AC$ (分配法則2)</div>
<div>
が成り立ちます。</div>
</div>
<div>
<br /></div>
</div>
<div>
この証明はそれぞれに行うことができますが、とりあえず</div>
<div>
最初の結合法則だけ証明しておきます。</div>
<div>
基本的に、実数の場合の証明がそのまま受け継がれるというものです。</div>
<div>
<br /></div>
<div>
$A=(a_{ij})\in M(m,n,{\mathbb R}),\ B=(b_{ij})\in M(n,l,{\mathbb R}),\ C=(c_{ij})\in M(l,p,{\mathbb R})$ とします。</div>
<div>
このとき、</div>
<div>
<br /></div>
<div>
$$A(BC)=(a_{ij})(\sum_{k=1}^lb_{ik}c_{kj})=(\sum_{q=1}^na_{iq}\sum_{k=1}^lb_{qk}c_{kj})$$</div>
<div>
$$=(\sum_{q=1}^n\sum_{k=1}^la_{iq}b_{qk}c_{kj})=(\sum_{q=1}^na_{iq}b_{qk}\sum_{k=1}^lc_{kj})$$</div>
<div>
$$=(\sum_{q=1}^na_{iq}b_{qj})(c_{ij})=((a_{ij})(b_{ij}))(c_{ij})$$</div>
<div>
$$=(AB)C$$</div>
<div>
となります。</div>
<div>
<br /></div>
<div>
<u><span style="font-size: large;">Part 6: 行列の分配法則の証明</span></u></div>
<div>
分配法則について証明しておきます。</div>
<div>
ここでは、分配法則の2について示しておきます。<br />
講義の方では1の方を示しています。</div>
<div>
<br /></div>
<div>
$A\in M(m,n,{\mathbb R})$ かつ $B,C\in M(n,l,{\mathbb R})$ であるとき、<br />
<div>
$A(B+C)=AB+AC$<br />
を示します。<br />
<br />
$A=(a_{ij})$, $B=(b_{ij})$, $C=(c_{ij})$ とします。<br />
$$A(B+C)=A(b_{ij}+c_{ij})=(\sum_{k=1}^na_{ik}(b_{kj}+c_{kj}))$$<br />
$$=(\sum_{k=1}^na_{ik}b_{kj}+\sum_{k=1}^na_{ik}c_{kj})$$<br />
$$=(\sum_{k=1}^na_{ik}b_{kj})+(\sum_{k=1}^na_{ik}b_{kj})$$<br />
$$=AB+AC$$<br />
となります。<br />
<br />
これらの証明を読む時に気をつけて欲しいことは、<br />
<u>なんとなく読み飛ばすのではなく、一つ一つ何を適用したのかを</u><br />
<u>考えていくことが大切です。</u><br />
一つ一つのステップは、実数の演算の法則や行列の和や積の定義や<br />
単なる並び替えなどしか使っていません。<br />
<br />
<span style="font-size: large;"><u>Part 7: 行列の転置</u></span><br />
$(m,n)$ 行列 $A=(a_{ij})$ に対して、$(i,j)$ 成分を、$A$ の $(j,i)$ 成分 $a_{ji}$<br />
であるような行列 $(a_{ji})$ をその転置行列といい、 $^tA$ (もしくは $A^T$) と書きます。<br />
<br />
このとき、転置行列 $^tA$ は $(n,m)$ 行列になります。<br />
<br />
例えば、$A=\begin{pmatrix}1&2&3\\4&5&6\end{pmatrix}$ ならば、<br />
$^tA=\begin{pmatrix}1&4\\2&5\\3&6\end{pmatrix}$<br />
となります。<br />
<br />
<br /></div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-33718585664150406032020-04-30T14:31:00.001+09:002022-04-19T10:14:47.517+09:00数学リテラシー1(第2回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(水曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px; margin: 0px;">
<div style="margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a></div>
</div>
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<span style="font-size: large;"><u>写像の一般論</u></span></div>
<div style="font-size: 16px;">
今日は写像についての講義でした。<br />
<br />
以下文字ばかりですいません。<br />
本当は図を載せるべきなのでしょうけれど、<br />
ここに図をのっけることはできるのですが、少々面倒なのでご勘弁を。<br />
できる限り、紙を用意して図を描きながら進めるとよいかもしれません。<br />
<br />
<br />
さぁて本題に入りましょう。<br />
まずは皆さんをのっけからこまらせた写像の説明からです。<br />
<br />
<br /></div>
<div style="font-size: 16px;">
$A, B$ を集合とします。$f:A\to B$ が<b>写像</b>であるとは、</div>
<div style="font-size: 16px;">
$A$ の<span style="font-size: large;">任意の元 $a$ に対して、ただ一つの</span> $B$ の元 $f(a)$ を定める</div>
<div style="font-size: 16px;">
規則のことを写像といいます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このとき、$a\mapsto f(a)$ とかくことで、その写像を定義することができます。</div>
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このように言われてもよくわからないかもしれませんね。</div>
<div style="font-size: 16px;">
皆さんがこれからずっと実践してほしいことは、</div>
<div style="font-size: 16px;">
<br /></div>
<span style="font-size: large;"><b>抽象的なことは、より具体的に考え、</b></span><br />
<span style="font-size: large;"><b>具体的なことは、より抽象的に考える。</b></span></div>
<div style="font-size: 16px;">
<br />
言われた通りに理解しようとしない。<br />
ちょっと穿った見方というか、1つのことを多角的に見ることを考えましょう。<br />
研究者でも、理解しにくいことは多々ありますが、<br />
そんなとき、以下のことを試しています。<br />
<br />
<ul style="text-align: left;">
<li>もっといい理解の仕方があるのではないか?</li>
<li>別の見方があるのではないか?</li>
<li>それを満たす条件は何か?</li>
<li>具体例が作れるか?</li>
<li>条件を外したとき成り立つのか成り立たないか?</li>
<li>なぜこのようなことを考えなければならないか?</li>
<li>極端な例を考えてみる。(その概念を満たすぎりぎりのところの例)</li>
</ul>
<br />
<br />
そのため、定義やルールを決めたときに、大抵例を書くのが普通です。<br />
実践していきましょう。<br />
<br />
<br /></div>
<div style="font-size: 16px;">
<u>例1</u></div>
<div style="font-size: 16px;">
三角関数 $\sin x$ を用いて、$x\mapsto f(x)=\sin x$ のように定めることで</div>
<div style="font-size: 16px;">
実数から$[-1,1]$ への写像を作ることができます。</div>
<div style="font-size: 16px;">
$[-1,1]$ とは、$-1$ 以上 $1$ 以下の実数を表します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このとき、写像 $f$ は $f:{\mathbb R}\to [-1,1]$ であることになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
また、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>例2</u></div>
<div style="font-size: 16px;">
整数 ${\mathbb Z}$ から、$1$ と $-1$ からなる集合 $\{1,-1\}$ への</div>
<div style="font-size: 16px;">
写像 $f:{\mathbb Z}\to \{1,-1\}$ を $n\mapsto f(n)=(-1)^n$ のように定めることができます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
つまり、写像とは関数のようなものだと考えることができますが、</div>
<div style="font-size: 16px;">
高校までの関数と違って、値となるものが数とは限らず、何かの集合</div>
<div style="font-size: 16px;">
であったり、数字の組み合わせであったり、平面上の点であったりします。</div>
<div style="font-size: 16px;">
そういうのも含めた”関数”を写像といいます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例えば、少々難しい写像としては、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例3</div>
<div style="font-size: 16px;">
実数 $m$ に対して、方程式 $y=mx+1$ を満たす $xy$-平面上の直線 $L_m= \{(x,y)|y=mx+1\}$ を与える規則も写像となります。</div>
<div style="font-size: 16px;">
そのような写像を $F$ とすると、</div>
<div style="font-size: 16px;">
$$F:{\mathbb R}\to \{L|L\text{は平面上の直線}\}$$</div>
<div style="font-size: 16px;">
となり、$F$ は $m\mapsto L_m$ という定義ということになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
すいません、この例3が理解できなければ無視して先に行ってください。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
この例にしても、上の2つの例にしても思想は、関数と同じです。</div>
<div style="font-size: 16px;">
何か与えられた集合 $A$ の任意の元に対して、何か $B$ の値が出てくるということです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
しかし、注意して欲しいことは、</div>
<div style="font-size: 16px;">
$A$ の<b><span style="font-size: large;">どの元に対しても定義しなければならず</span></b>、</div>
<div style="font-size: 16px;">
$B$ の<b><span style="font-size: large;">ある元の一つだけに対応しなければならない</span></b></div>
<div style="font-size: 16px;">
ということです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例えば、$x\mapsto f(x)=\frac{1}{x}$ という対応を考えても、これは</div>
<div style="font-size: 16px;">
写像 $f:{\mathbb R}\to {\mathbb R}$ とは呼びません。</div>
<div style="font-size: 16px;">
$0$ の行き先 $f(0)$ が定義されていないからです。</div>
<div style="font-size: 16px;">
${\mathbb R}\setminus \{0\}$ から ${\mathbb R}$ への写像としては定義されています。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
また、$A=B={\mathbb R}$ として、$x\mapsto \pm \sqrt{x}$ という</div>
<div style="font-size: 16px;">
対応も写像ではありません。まず、$x$ が負の数であれば、$\sqrt{x}$ は実数では</div>
<div style="font-size: 16px;">
ありませんし、符号がどちらか決められていないので、${\mathbb R}$ の</div>
<div style="font-size: 16px;">
ただ一つの値としても決まってはいません。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
もう少しイメージしやすい例を考えましょう。</div>
<div style="font-size: 16px;">
小学生全体を $A$ として、学年の集合 $\{1,2,3,4,5,6\}$ を $B$ とします。</div>
<div style="font-size: 16px;">
今、写像 $f:A\to B$ をその小学生の属する学年に写す規則を考えます。</div>
<div style="font-size: 16px;">
つまり、$a\in A$ という人には、$a$ の所属する学年を $n$ としたとき、</div>
<div style="font-size: 16px;">
$a\mapsto n$ という規則で写像を作るのです。</div>
<div style="font-size: 16px;">
この規則は、全ての小学生に、ただ一つ学年が決められているわけですから、</div>
<div style="font-size: 16px;">
写像となります。</div>
<div style="font-size: 16px;">
これを<b>学年写像</b>と呼ぶことにしましょう。</div>
<div style="font-size: 16px;">
下でもこの名前を使います。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例えば、小学生全員に、「あなたの弟の学年のところまでいきなさい!」</div>
<div style="font-size: 16px;">
という指令を出してしまうと、場合によっては弟がそもそも存在しない場合もあるし、</div>
<div style="font-size: 16px;">
小学生の弟は在籍するのだが2人以上違う学年におり1つに決まらないなどの場合は、</div>
<div style="font-size: 16px;">
児童は路頭に迷い、逆に「変な指令をだすんじゃない!」と怒られてしまうわけです。</div>
<div style="font-size: 16px;">
もしかしたらこのような変な指令をだしたらほとんどの児童は動けないかもしれません。</div>
<div style="font-size: 16px;">
つまり写像とは自動的に1つに行き先が決まらないといけないわけです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
学年写像 $f$ は、運動会での全体演技後「自分の学年の場所に戻りなさい」</div>
<div style="font-size: 16px;">
でも同様なことが起こるわけです。</div>
<div style="font-size: 16px;">
このように言われるととすると、</div>
<div style="font-size: 16px;">
(素直な小学校の子供たちなら、)自動的に自分の元の持ち場 $\{1,2,3,4,5,6\}$</div>
<div style="font-size: 16px;">
に戻れるわけですよね。(児童なだけに...)</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u><span style="font-size: large;">単射と全射</span></u></div>
<div style="font-size: 16px;">
次は単射と全射です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
単射と全射を教えるとき、いつもわからない人が続出する部分です。<br />
twitterでも何人かの学生がわからなかったと言っていたようにみえました。</div>
<div style="font-size: 16px;">
単射と全射は初めてきく概念ですので、一度聞いてすぐわかることは難しいと思います。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
もう一度書いておきましょう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u><b>単射の定義</b></u></div>
<div style="font-size: 16px;">
写像 $f:A\to B$ が単射であるとは、</div>
<div style="font-size: 16px;">
$a,a’\in A$ に対して、</div>
<div style="font-size: 16px;">
もし、$a\neq a’$ であれば、$f(a)\neq f(a’)$ を満たす</div>
<div style="font-size: 16px;">
ことをいう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
まず、$f$ が写像であることが仮定として必要です。</div>
<div style="font-size: 16px;">
写像でないと単射かどうかの議論さえできません。</div>
<div style="font-size: 16px;">
結論を言えば写像でないものは単射になり得ません。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
これ以降、$f$ は写像であることを仮定します。</div>
<div style="font-size: 16px;">
そのとき、単射であることは、文字通り読めば、</div>
<div style="font-size: 16px;">
違う元同士は違う元に写るということです。</div>
<div style="font-size: 16px;">
しかし、写像によっては、違う元なのに同じ元に行ってしまうことがあります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例えば、一番簡単な例は、</div>
<div style="font-size: 16px;">
$f:{\mathbb R}\to {\mathbb R},x\mapsto x^2$ ですが、$1$ の行き先も $-1$ の行き先もどちらも $1$ です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
しかし、$g:{\mathbb R}\to {\mathbb R}$ として、$x\mapsto x^3$ であれば、</div>
<div style="font-size: 16px;">
これは単射です。</div>
<div style="font-size: 16px;">
$x,x’\in {\mathbb R}$ のとき、$x\neq x’$ ならば、$x<x’$ と</div>
<div style="font-size: 16px;">
仮定してやると、$x^3<(x’)^3$ であるから、とくに、$x^3\neq (x’)^3$</div>
<div style="font-size: 16px;">
が成り立ちます。$x’<x$ としても同様です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
なので、写像によって単射であったり、単射でなかったりします。</div>
<div style="font-size: 16px;">
たとえば、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$f:{\mathbb Z}\to {\mathbb Z}$ として、$n\mapsto f(n)=n+1$ という写像</div>
<div style="font-size: 16px;">
とすると、$n\neq n’\in {\mathbb Z}$ に対して、$f(n)=n+1\neq n’+1=f(n’)$</div>
<div style="font-size: 16px;">
が成り立つので、この $f$ も単射になります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
最後に、単射性を対偶を使って言い換えておきます。</div>
<div style="font-size: 16px;">
$p\Rightarrow q$ の対偶は、$\bar{q}\Rightarrow \bar{p}$ ですから、</div>
<div style="font-size: 16px;">
単射性の条件は</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$f(a)=f(a’)$ なら、$a=a’$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
となります。<br />
この関係は下で何度もでてきます。<br />
<br />
次は全射です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u><b>全射の定義</b></u></div>
<div style="font-size: 16px;">
写像 $f:A\to B$ が全射であるとは、$B$ の任意の元 $b$ に</div>
<div style="font-size: 16px;">
対して、$A$ の元 $a$ が存在して、$f(a)=b$ を満たす</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ポイントは、$B$ の任意の元(つまり全ての元)というところです。</div>
<div style="font-size: 16px;">
$B$ のありとあらゆる元をとっても、その元に写ってくる $A$ の元がいる</div>
<div style="font-size: 16px;">
ということです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
今、学年写像 $f$ を用いて小学生全体 $A$ に対して学年</div>
<div style="font-size: 16px;">
$B=\{1,2,3,4,5,6\}$ を割り当てるとき、</div>
<div style="font-size: 16px;">
$f$ が全射であるということは、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<span style="font-size: large;">どの学年にも一人以上は児童がいるという状況</span></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
これが全射ということです。</div>
<div style="font-size: 16px;">
例えば、4年生だけ一人もいないという状況もどこかの小学校でも</div>
<div style="font-size: 16px;">
起こり得るかもしれませんが、このような小学校における</div>
<div style="font-size: 16px;">
学年写像 $f$ は全射ではないということになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
上の写像の定義を小学生を用いて言い換えると、次のようになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>全射の定義(学年写像</u><u>版)</u></div>
<div style="font-size: 16px;">
学年写像 $f:A\to B$ が全射であるとは、任意の学年 $b\in \{1,2,3,4,5,6\}$ に</div>
<div style="font-size: 16px;">
対して、ある小学生 $a$ が存在して、$a$ の学年は $b$ になる。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
少し言い方は堅苦しいかもしれませんが、意味はわかるのではないでしょうか?</div>
<div style="font-size: 16px;">
この定義から上の全射の定義まで目を平行移動させて、もう一度全射の定義を</div>
<div style="font-size: 16px;">
読んでみてください。</div>
<div style="font-size: 16px;">
(この行為を抽象化といいます。)</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
数学を用いた全射の例は以下のものがあります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例4</div>
<div style="font-size: 16px;">
$f:{\mathbb Z}\to \{1,-1\}$</div>
<div style="font-size: 16px;">
を、$n\mapsto f(n)=(-1)^n$ とします。</div>
<div style="font-size: 16px;">
このとき、$1\mapsto -1$ であり、 $2\mapsto 1$ であるので、</div>
<div style="font-size: 16px;">
$1,-1$ のどちらにも、それぞれその数に写ってくる整数が存在します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u><b>全単射の定義</b></u></div>
<div style="font-size: 16px;">
写像 $f:A\to B$ が全射かつ単射であるとき、$f$ は<b>全単射</b>という。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
全単射というのは、$A$ と $B$ が完全に対応関係が定まる</div>
<div style="font-size: 16px;">
ことを表します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例えば、上の写像 $f:{\mathbb Z}\to {\mathbb Z}, n\mapsto f(n)=n+1$</div>
<div style="font-size: 16px;">
とするとすると、$n$ に対してただ一つだけ $f(n)$ が</div>
<div style="font-size: 16px;">
定まり、$m\in {\mathbb Z}$ に対して、$m=f(n)$ となる $n$ はただ一つ</div>
<div style="font-size: 16px;">
しか定まりません。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
実は、$f$ が全単射であることと、逆写像(逆向きの写像で</div>
<div style="font-size: 16px;">
$f$ の逆の割り当てをするもの)が定まることは同じことです。</div>
<div style="font-size: 16px;">
これは下の補題で証明します。この説明は中途半端ですが、やめておきます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このように、全ての元に自動的に答えが定まるようなものは全て写像です。</div>
<div style="font-size: 16px;">
そういうわけで、ある意味身の回りには写像であふれていると言っても良いです。</div>
<div style="font-size: 16px;">
家の中を見回してみると写像となっているものはすぐに見つかるの</div>
<div style="font-size: 16px;">
ではないでしょうか?</div>
<div style="font-size: 16px;">
また、意外なところに写像が存在してハッとなるかもしれません。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例えば人間のパーソナリティに関係するデータは全て写像を与えます。</div>
<div style="font-size: 16px;">
人間の集合から血液型全体 $\{A,B,O,AB\}$ への写像、年齢をとる規則も写像となるし、</div>
<div style="font-size: 16px;">
生まれ落ちた場所を与える規則も写像になります。</div>
<div style="font-size: 16px;">
人間全体の集合から地球上のある点 $p$ への写像。</div>
<div style="font-size: 16px;">
また、役所にデータ化されている住所だって一人一人に与えられた</div>
<div style="font-size: 16px;">
写像です。大学に入って一人暮らしを始めたとすると、</div>
<div style="font-size: 16px;">
日本に住む日本人全員を $A$ として、その住所を割り当てる写像を $f:A\to B$ とする</div>
<div style="font-size: 16px;">
($B$ は日本の住所全体)自分 $a$ の行き先が去年より少しずれることになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
人間がその日に何を食べたかだって食べ物全体の部分集合全体への写像をなします。</div>
<div style="font-size: 16px;">
何も食べない日があったとしても、空集合への写像となります。</div>
<div style="font-size: 16px;">
(この例は少々むずかしいか...)</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
単射の例をとしては、マイナンバー制度があります。</div>
<div style="font-size: 16px;">
日本人全体を $A$、12桁の数全体を $B$ とすると、</div>
<div style="font-size: 16px;">
$A\to B$ として、その人のマイナンバーを与える写像とします。</div>
<div style="font-size: 16px;">
このとき、違う人に同じ番号を割り当てることはないわけなので、</div>
<div style="font-size: 16px;">
単射ということになります。</div>
<div style="font-size: 16px;">
しかし、全射ではありません。12桁の数すべてに対して</div>
<div style="font-size: 16px;">
どなたか日本人が存在したとすると、</div>
<div style="font-size: 16px;">
日本人の数は12桁の数字の分だけあるということになってしまいます。</div>
<div style="font-size: 16px;">
12桁の数全体は、$10^{12}$ で、現在の日本の人口はおよそ $1.2\times 10^8$ </div>
<div style="font-size: 16px;">
くらいですからそういうことは起こりません。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
などなど、日常みてみると、値が数となるような関数より、何か点だったり、</div>
<div style="font-size: 16px;">
集合だったりで数ではないようなバリエーションがある写像に溢れています。</div>
<div style="font-size: 16px;">
用途に応じて単射、全射もいくつかあります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
次に、像と逆像の定義をしましょう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>像</u></div>
<div style="font-size: 16px;">
写像 $f:A\to B$ に対して、$C\subset A$ を考える。</div>
<div style="font-size: 16px;">
$f(C)=\{f(c)|c\in C\}$ を $C$ の $f$ による<b>像</b>という。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>逆像</u></div>
<div style="font-size: 16px;">
写像 $f:A\to B$ に対して、$D\subset B$ を考える。</div>
<div style="font-size: 16px;">
$f^{-1}(D)=\{a|f(a)\in D\}$ を $D$ の $f$ による<b>逆像</b>という。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
像というのは、$c\in C$ に対して、$f(c)$ となる元全体ということです。</div>
<div style="font-size: 16px;">
先ほど小学生の例を考えたので、その例を用いましょうか。</div>
<div style="font-size: 16px;">
$A,B,f$ は先ほどと同じで、$f$ は学年写像とします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
まず $C=A$ の場合は $f(C)=f(A)$ はその小学校の在籍する学年の集合ということに</div>
<div style="font-size: 16px;">
なります。</div>
<div style="font-size: 16px;">
次に、$C\subset A$ として $A$ より小さい集合をとります。</div>
<div style="font-size: 16px;">
例えば、$A$ の中で女子児童だけの集合を $C$ としましょうか。</div>
<div style="font-size: 16px;">
このとき、$f(C)$ とは、女子児童のいる学年の集合</div>
<div style="font-size: 16px;">
ということになります。そうすると、$f(C)$ はもしかしたら $f(A)$ より</div>
<div style="font-size: 16px;">
小さいかもしれません。</div>
<div style="font-size: 16px;">
<br />
仮に、ある学年には女子がおらず男子だけの学年があったとすると、</div>
<div style="font-size: 16px;">
真に、$f(C)\subset f(A)$ が成り立ち、そのような(女子なし)学年を除いた</div>
<div style="font-size: 16px;">
集合が $f(C)$ ということになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
逆像というのは、$f$ で写したときに、$D$ に入るような$A$ の集合ということです。</div>
<div style="font-size: 16px;">
これも上の例を出しましょうか。</div>
<div style="font-size: 16px;">
例えば、$D\subset B$ として偶数学年とするとします。</div>
<div style="font-size: 16px;">
$D=\{2,4,6\}$ ということです。</div>
<div style="font-size: 16px;">
そのとき、$f^{-1}(D)=\{a\in A|f(a)\in D\}$ は学年を聞いたときに偶数であるような</div>
<div style="font-size: 16px;">
児童全体ということですから、もちろん偶数学年の児童全体を表します。</div>
<div style="font-size: 16px;">
補集合を取れば、$f^{-1}(D^c)$ は奇数学年全体です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここまでで像と逆像という概念の説明を終わります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u><b>合成</b></u></div>
<div style="font-size: 16px;">
写像 $f:A\to B$ と $g:B\to C$ とすると、$a\in A$ に対して、$g(f(a))$ を</div>
<div style="font-size: 16px;">
与える写像を $g\circ f:A\to C$ と書くことにし、</div>
<div style="font-size: 16px;">
$f$ と $g$ の合成という。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
関数でいえば、合成関数のことですね。</div>
<div style="font-size: 16px;">
あまりここでは問題はないかと思いますので素通りします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここでは、スライド通り補題の説明を試みましょう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>補題</b></div>
<div style="font-size: 16px;">
写像 $f:A\to B$ が全単射とする。</div>
<div style="font-size: 16px;">
このとき、ある写像 $g:B\to A$が存在して、$g\circ f=\text{id}_A$ および</div>
<div style="font-size: 16px;">
$f\circ g=\text{id}_B$ となる。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
この補題が意味することは、<br />
<br />
「$f$ が全単射であること」は、<br />
「$f$ の逆写像 (逆向きの写像 $g:B\to A$ で、$f$ との合成が恒等写像となるもの)<br />
が存在すること」<br />
<br />
であるということを意味します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
なぜ、単射と全射が成り立てば、逆の対応関係があることを意味するのか</div>
<div style="font-size: 16px;">
ということを証明するわけです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$\text{id}_A$ は $A$ 上の恒等写像 $a\mapsto a$ を意味します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
では、証明しましょう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明)</div>
<div style="font-size: 16px;">
仮定は、$f:A\to B$ が全単射であるということです。</div>
<div style="font-size: 16px;">
結論は、逆写像 $g$ が構成できるということです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
つまり、写像 $f:A\to B$ が全単射とします。</div>
<div style="font-size: 16px;">
逆写像 $g:B\to A$ を作りたいのだから、$B$ の元 $b$ をとって、<br />
$g$ による $b$ の行き先となる$A$ の元を決めてやれば良いことになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここで$f$の全射性を用いましょう。</div>
<div style="font-size: 16px;">
全射はすべての $B$ の元 $b$ に対して $f(a)=b$ となる $a$ が存在するのだから、</div>
<div style="font-size: 16px;">
なにかしら $a$ が定まります。<br />
<br />
しかし、実は、$f(a)=b$ を満たす $a$ は一意に決まることを証明します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
なぜなら、$f(a’)=f(a)=b$ となる$a’$ がもしあったとしたら、</div>
<div style="font-size: 16px;">
$f$ の単射性から、$a=a’$ が成り立ちます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
つまり、$b$ に対して $f(a)=b$ を満たす $a$ を対応させること<br />
$b\mapsto a$ は写像になるということを言っています。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
この対応、$b\mapsto a$ は、写像となり、</div>
<div style="font-size: 16px;">
この写像を $g$ と書くことにすると、 $g:B\to A$ を与えたことになります。</div>
<div style="font-size: 16px;">
このとき、$f\circ g(b)=f(a)=b$ であるから、$f\circ g$ は恒等写像であり、</div>
<div style="font-size: 16px;">
$g\circ f(a)=g(b)=a$ であるから、$g\circ f$ も恒等写像となります。(証明終了)</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このとき、$f$ が逆写像をもつなら、$f$ が全単射であるのか?</div>
<div style="font-size: 16px;">
ということが気になるかもしれませんが、これは実際以下のようにして<br />
証明することができます。<br />
<br /></div>
<div style="font-size: 16px;">
(証明)</div>
<div style="font-size: 16px;">
$f:A\to B$ が逆写像 $g$ を持つとします。</div>
<div style="font-size: 16px;">
このとき、$f$ が全射で単射であることを示しましょう。</div>
<div style="font-size: 16px;">
$f$ が全射であることは、</div>
<div style="font-size: 16px;">
任意の$b\in B$ に対して、$g(b)=a$ であるとき、$f\circ g(b)=f(a)=b$</div>
<div style="font-size: 16px;">
ですから、$f(a)=b$ となる $a$ が見つかるということは $f$ が全射であるということ<br />
になります。</div>
<div style="font-size: 16px;">
<br />
また、$f$ が単射であることは、次のようにして証明します。</div>
<div style="font-size: 16px;">
$f(a)=f(a’)$ が成り立つとします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
そうすると、$g(f(a))=g(f(a’))$ も成り立ち、</div>
<div style="font-size: 16px;">
$g\circ f$ は恒等写像であるから、$a=a’$ となります。</div>
<div style="font-size: 16px;">
よって、$f$ は単射です。(証明終了)</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このことから、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<span style="font-size: large;">「$f$ が全単射であること」$\Leftrightarrow$ 「$f$ が逆写像をもつこと」</span><br />
<br />
<br /></div>
<div style="font-size: 16px;">
ことがわかりました。</div>
<div style="font-size: 16px;">
つまり、$f$ が逆写像を持つかどうかを確かめるには、$f$ が全射で単射であることを</div>
<div style="font-size: 16px;">
確かめればよいということにもなります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
次の補題は、</div>
<div style="font-size: 16px;">
合成に関する話です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
補題</div>
<div style="font-size: 16px;">
写像 $f:A\to B$ と $g:B\to C$ に対して、次が成り立つ。</div>
<div style="font-size: 16px;">
(1) $f$ と $g$ が単射ならば、$g\circ f$ も単射である。</div>
<div style="font-size: 16px;">
(2) $f$ と $g$ が全射ならば、$g\circ f$ も全射である。</div>
<div style="font-size: 16px;">
(3) $f$ と $g$ が全単射ならば、$g\circ f$ も全単射である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
証明をしておきましょう。<br />
(1) $f,g$ が単射であるとします。このとき、$a,a’\in A$ に対して、<br />
$g\circ f(a)=g\circ f(a’)$ が成り立つとする。(結論として$a=a’$ であることがわかればよい)<br />
このとき、$g$ の単射性から $f(a)=f(a’)$ が成り立ち、$f$ の単射性から<br />
$a=a’$ となります。<br />
<br />
<br />
(2) $f,g$ が全射であるとします。<br />
$g\circ f$ が全射であることを示します。<br />
任意の $c\in C$ に対して、$g(b)=c$ となる $b\in B$ が存在します。<br />
$f$ が全射であることから、$f(a)=b$ となる $a\in A$ が存在します。<br />
よって、$g\circ f(a)=g(b)=c$ となるので、$a\in A$ となります。<br />
<br />
(3) $f,g$ が全単射であるとすると、$g\circ f$ は(1) と(2) を用いて<br />
全射かつ単射になります。よって$g\circ f$ は全単射となります。(証明終了)<br />
<br />
<br />
<br />
一言で言って仕舞えば、<br />
写像が単射であれば、違うものは違うものに写るので<br />
それが2回合成されても違うものは違うものに行くということと、<br />
<br />
写像が全射であれば、集合全部に向けて写っている写像を2回<br />
合成しても、全体に向けてうつされるということです。<br />
<br />
<br />
(3) は(1)と(2) を合わせただけとなります。<br />
<br />
<br />
次の命題はこれまでと違ってさほど簡単ではないかもしれません。<br />
なので命題となっています。<br />
これまでの内容(特に定義)がわかればわかると思います。<br />
<br />
命題<br />
2つの写像 $f:A\to B$ と $g:B\to C$ に対して次がなりたつ。<br />
<br />
(1) $g\circ f:A\to C$ が単射ならば $f$ も単射<br />
(2) $g\circ f:A\to C$ が全射ならば $g$ も全射<br />
<br />
<br />
もう一度証明をしておきます。<br />
(1) $g\circ f$ を単射であるとします。<br />
このとき、$a,a’\in A$ をとります。$f(a)=f(a’)$ が<br />
が成り立つとする。(目標は、$a=a’$ となることなのですが...)<br />
このとき、$g\circ f(a)=g\circ f(a’)$ が成り立ちます。<br />
今、$g\circ f$ が単射であることから、$a=a’$ がわかります。<br />
よって、$f$ が単射であることがわかります。<br />
<br />
<br />
(2) $g\circ f$ が全射であるとします。<br />
このとき、示すことは、任意の$c\in C$ に対して $g(b)=c$ となる$b$ が<br />
存在することです。<br />
しかし、今 $g\circ f$ は全射であるので、$g\circ f(a)=c$ となる$a$<br />
が存在します。よって、$b=f(a)$ とすることで、$g(b)=c$ となる$b\in B$<br />
が存在することがわかります。(証明終了)<br />
<br />
<br />
最後に次の問題をときます。<br />
これはある意味集合の問題です。<br />
<span style="background-color: white; color: black; display: inline; float: none; font-family: "quot"; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;"><br /></span>
問題<br />
写像 $f:A\to B$ を考える。このとき、部分集合 $C_1,C_2\subset A$ に<br />
対して、<br />
(1) $f(C_1\cap C_2)\subset f(C_1)\cap f(C_2)$<br />
(2) $f(C_1\cup C_2)=f(C_1)\cup f(C_2)$<br />
<br />
が成り立つ。<br />
<br />
$X,Y$ が集合であるとします。$X\subset Y$ が成り立つということを証明するには、<br />
$X$ の任意の元 $x\in X$ が $x\in Y$ であることを証明をすれば良い<br />
です。<br />
<br />
また、集合 $X,Y$ に対して、$X=Y$</div>
<div style="font-size: 16px;">
であることは、$X\subset Y$ かつ $Y\subset X$ が成り立つことです。</div>
<div style="font-size: 16px;">
つまり、$x\in X$ ならば $x\in Y$ であり、$y\in Y$ ならば $y\in X$ </div>
<div style="font-size: 16px;">
が成り立つことを示せば良いです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(1) の証明</div>
<div style="font-size: 16px;">
$b\in f(C_1\cap C_2)$ であるとします。</div>
<div style="font-size: 16px;">
このとき、$b=f(a)$ となる $a\in C_1\cap C_2$ が存在します。</div>
<div style="font-size: 16px;">
よって、$a\in C_1$ かつ $x\in C_2$ が成り立ちます。</div>
<div style="font-size: 16px;">
よって、像の定義から、$f(a)\in f(C_1)$ であり、$f(a)\in f(C_2)$ </div>
<div style="font-size: 16px;">
が成り立ちます。よって、$b=f(a)\in f(C_1)\cap f(C_2)$ が成り立ちます。</div>
<div style="font-size: 16px;">
これで、$f(C_1\cap C_2)\subset f(C_1\cap C_2)$ が成り立ちます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(2) の証明</div>
<div style="font-size: 16px;">
これはレポート問題なのですが、</div>
<div style="font-size: 16px;">
方針のみ書いておきます。</div>
<div style="font-size: 16px;">
まず、$b\in f(C_1\cup C_2)$ ならば、$b\in f(C_1)\cup f(C_2)$ </div>
<div style="font-size: 16px;">
を示します。上と同じように、像の定義から、$b=f(a)$ となる</div>
<div style="font-size: 16px;">
$a\in C_1\cup C_2$ が存在します。そして.....</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
今度は逆に、$b’\in f(C_1)\cup f(C_2)$ とすると、$b\in f(C_1\cup C_2)$</div>
<div style="font-size: 16px;">
が成り立つことを示します。</div>
<div style="font-size: 16px;">
<br />
<span style="color: black; display: inline; float: none; font-family: "quot"; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">$b'\in f(C_1)\cup f(C_2)$ とすると、</span>$b’\in f(C_1)$ または、$b’\in f(C_2)$ が成りたちます。</div>
<div style="font-size: 16px;">
前者であれば、$b’=f(a’)$ となる $a’\in C_1$ が存在し、</div>
<div style="font-size: 16px;">
後者であれば、.....</div>
<div style="font-size: 16px;">
----------------</div>
<div style="font-size: 16px;">
このように、<br />
$f(C_1\cup C_2)\subset f(C_1\cup C_2)$</div>
<div style="font-size: 16px;">
と<br />
$f(C_1)\cup f(C_2)\subset f(C_1\cup C_2)$<br />
を両方示すことによって<br />
証明を行います。<br />
<br />
<br />
この内容でわからないことがありましたら、下のコメント欄か、<br />
いつでも連絡いただければと思います。<br />
連絡先は、<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1</a>の下部にあります。</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-78571711712100943582020-04-29T15:19:00.001+09:002022-04-19T10:14:57.177+09:00数学リテラシー1(第1回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所:manaba上(火曜日12:00〜)]</span></div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px; margin: 0px;">
<div style="margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/20/literacy1.html">数学リテラシー1のHP</a></div>
</div>
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<span style="font-size: large;"><u>集合の基礎</u></span></div>
<div style="font-size: 16px;">
<span style="font-size: large;"><br /></span>
まずは数学は集合からやっていかなくちゃならんのですが、</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<br />
そこで最初に躓くことは、この集合という抽象的な概念です。<br />
しかし、集合はわかってしまえばなんてことはありません。<br />
考え方をみにつけてしまえばたとえ小学生だって理解できます。<br />
<br />
抽象的な概念をやわらかくするには、<br />
<span style="font-size: large;"><b>頭の中によい例を思い描くことです。</b></span><br />
理解のためには具体的なイメージを持つことが大切です。<br />
<br />
まず、集合がどういうものか述べます。<br />
<br />
<b>集合</b>とはものの集まりをいいます。
<br />
<div style="font-size: 16px;">
ものというのは、数字や論理式を用いて定義された対象を<br />
集めたもののことをいいます。</div>
<div style="font-size: 16px;">
集合に入っている要素を<b>元</b>といいます。</div>
<div style="font-size: 16px;">
<br />
<br />
集合というと、「集合!!整列!!前ならえ!!」<br />
という体育の時間かと思いますがそうではないですね。<br />
確かに筑波は体育学群もありスポーツは盛んですが、<br />
数学だってやっています。ノーベル賞だって獲っています。<br />
魔法使いもいます。<br />
<br />
それはさておき、<br />
集合とは、物があつまった状態、そのものたち<br />
自身のことをいいます。<br />
<br />
イメージするものは、<br />
<div style="color: black; font-family: "quot"; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
例えば、りんごがバスケットの中にいくつか入っているイメージでしょうか。</div>
<div style="color: black; font-family: "quot"; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
りんごが一つの場所に集まっています。ものが集まっていますから<br />
この状態を集合といいます。</div>
<div style="color: black; font-family: "quot"; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
大事なことは、ものというのは、りんごのように一つ一つがちゃんと<br />
区別されている必要があります。<br />
おばけのように、考えているうちに消えてしまったりするようなものは<br />
集合とはなりません。<br />
<br />
なのでおばけの集合というものは考えることはできません。<br />
このへんは考え方次第かもしれませんが....<br />
少なくとも、みんながそれ!と確認ができるものでないといけません。<br />
<br /></div>
<div style="color: black; font-family: "quot"; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin: 0px; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<span style="font-size: xx-small;"><br /></span></div>
そのほか、ある幼稚園にいる園児だって一人一人を元とした集合をなします。<br />
目まぐるしくうごきまわってなかなか判別することは難しいかもしれませんが、<br />
カリスマ保育士がひとたび声をかけて、絵本を読みだせば、みんな静まり返って<br />
その絵本の読み聞かせに耳を傾けます。<br />
どちらにしても、そこに、一つ一つ区別できる対象が存在することが大事です。<br />
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
<br />
数学でよく扱う例をみましょう。<br />
<br /></div>
<div style="font-size: 16px;">
<u>例</u><br />
よく数学で使われるのは次の集合たちです。</div>
<div style="font-size: 16px;">
${\mathbb N}$ 自然数の集合</div>
<div style="font-size: 16px;">
${\mathbb R}$ 実数の集合</div>
<div style="font-size: 16px;">
${\mathbb Q}$ 有理数の集合</div>
<div style="font-size: 16px;">
${\mathbb Z}$ 整数の集合</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
この集合たちにはこのさき同じ記号を使っていくことになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
集合$X,Y$ に対して、$X$ の全ての元が $Y$ の元であるとき、<br />
<br />
$X\subset Y$ と表します。</div>
<div style="font-size: 16px;">
つまり、$X$ の全ての元が $Y$の元であるということです。</div>
<div style="font-size: 16px;">
$X\subset Y$ のような関係を<b>包含関係</b>といいます。</div>
<div style="font-size: 16px;">
$X$ のことを $Y$ の<b>部分集合</b>ということもあります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
上の例ですと、</div>
<div style="font-size: 16px;">
$${\mathbb N}\subset {\mathbb Z}\subset {\mathbb Q}\subset {\mathbb R}$$</div>
<div style="font-size: 16px;">
のような包含関係になります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
何も元を含まないものを<b>空集合</b>といいます。</div>
<div style="font-size: 16px;">
空集合も集合の一つです。$\emptyset$ と書きます。</div>
<div style="font-size: 16px;">
$\{x|x\not\in A\}$ のことを$A^c$ のように書いて、$A$ の補集合といいます。</div>
<div style="font-size: 16px;">
c とは、補集合を表すcomplementの頭文字です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>集合の書き方</u></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$A$ の集合で、命題 $P(x)$ を満たす集合を</div>
<div style="font-size: 16px;">
$$\{x\in A|P(x)\}$$</div>
<div style="font-size: 16px;">
のように書きます。<br />
命題については、のちに解説されます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例えば、1より大きい整数の集合は、</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$$\{x\in {\mathbb Z}|x>1\}$$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
となります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>集合の演算</u></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
集合同士の演算を以下のように定義します。</div>
<div style="font-size: 16px;">
$A\cup B=\{x|x\in A\text{または}x\in B\}$ (和集合)</div>
<div style="font-size: 16px;">
$A\cap B=\{x|x\in A\text{かつ}x\in B\}$ (共通集合)</div>
<div style="font-size: 16px;">
$A\setminus B=\{x|x\in A\text{かつ}x\not\in B\}$ (差集合)<br />
<br />
$A\setminus B=A\cap B^c$ と書くこともできることに注意しておきます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>ド・モルガンの法則</u></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
次が成り立ちます。</div>
<div style="font-size: 16px;">
$$(A\cap B)^c=A^c\cup B^c$$</div>
<div style="font-size: 16px;">
$$(A\cup B)^c=A^c\cap B^c$$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
補集合と集合の演算の間の分配法則ですが、</div>
<div style="font-size: 16px;">
$\cap$ と $\cup$ が互いに入れ替わることに注意をしてください。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>問題</u></div>
<div style="font-size: 16px;">
$$U=\{1,2,3,4,....20\}$$</div>
<div style="font-size: 16px;">
とする。</div>
<div style="font-size: 16px;">
$A=\{x\in U|x\text{は素数}\}$</div>
<div style="font-size: 16px;">
$B=\{x\in U|x\text{は60の約数}\}$</div>
<div style="font-size: 16px;">
$A\cup B$, $A\cap B$, $A\setminus B$ $(B\setminus A)^c$ はどうなるか?</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
とすると、以下の集合となります。</div>
<div style="font-size: 16px;">
$$A\cup B=\{1,2,3,4,5,6,7,10,11,12,13,15,17,19,20\}$$</div>
<div style="font-size: 16px;">
<div>
$$A\cap B=\{8,9,14,16,18\}$$</div>
</div>
<div style="font-size: 16px;">
$$A\setminus B=\{7,11,13,17,19\}$$</div>
<div style="font-size: 16px;">
$$(B\setminus A)^c=(B\cap A^c)^c=B^c\cup A=\{2,3,5,7,8,9,11,13,14,16,17,18,19\}$$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
問題</div>
<div style="font-size: 16px;">
(1) 3つの元からなる集合 $\{a,b,c\}$ の部分集合を全て書け。</div>
<div style="font-size: 16px;">
(2) また、元の数が $N$ 個の場合、その部分集合はいくつ作られるか?</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(解答)</div>
<div style="font-size: 16px;">
(1) $\emptyset$, $\{a\}$. $\{b\}$, $\{c\}$, $\{a,b\}$, $\{a,c\}$, $\{b,c\}$, $\{a,b,c\}$ (解答終了)<br />
<br /></div>
<div style="font-size: 16px;">
となります。ここで空集合も部分集合であることに注意しましょう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(2) $N$ この元からなる集合の部分集合を数える。</div>
<div style="font-size: 16px;">
$\{a_1,a_2,\cdots, a_N\}$ の部分集合は、それぞれの元についてそれが</div>
<div style="font-size: 16px;">
含まれるか含まれないかの</div>
<div style="font-size: 16px;">
2通りずつパターンがあるから、全部で、$2^N$ 通りあります。</div>
<div style="font-size: 16px;">
つまり部分集合は全部で $2^N$ 個作れる。(解答終了)</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<span style="font-size: large;"><u>命題</u></span><br />
命題・・・真偽が判定できるもの。<br />
<br />
$p,q$ をそれぞれ何かの命題とする。<br />
例えば、$p=$「$a$ は $-1$ 以下である」などである。<br />
この場合、$a$ に何か入れて成り立つので、<br />
$p(a)$ ということもあります。<br />
<br />
「$p$ かつ $q$」 を $p\land q$ とかきます。<br />
「$p$ または $q$」 を $p\lor q$ とかきます。<br />
「$p$ ならば $q$」 を $p\Rightarrow q$ とかく<br />
$p$ の否定命題を $\bar{p}$ とかく。<br />
<br />
よって、$p\Rightarrow q$ の対偶は、$\bar{q}\Rightarrow \bar{p}$ とかける。<br />
<br />
<br />
全体集合 $U$ を決めておきます。<br />
このとき、<br />
$P=\{x\in U|p(x)\}$ を命題 $p$ の<b>真理集合</b>という。<br />
命題$q$ の真理集合を $Q$ と書くと、<br />
$p\land q$ の真理集合は、$P\cap Q$ となり、<br />
$p\lor q$ の真理集合は、$P\cup Q$ となり、<br />
$\bar{p}$ の真理集合は、$P^c$ となります。<br />
また、<br />
$p\Rightarrow q$ であることは、<br />
$P\subset Q$ とかけます。<br />
<br />
<div>
ド・モルガンの法則は、命題を次のように書き換えることができます。</div>
<div>
$\overline{p\land q}=\bar{p}\lor \bar{q}$</div>
<div>
$\overline{p\lor q}=\bar{p}\land \bar{q}$</div>
<div>
ここでのイコールは、命題の真偽がこの両辺においていつでも一致する<br />
ことを表しています。<br />
<br />
また、対偶が等しい命題であることは、<br />
$p\Rightarrow q=\bar{q}\Rightarrow \bar{p}$<br />
と書くことができます。</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-41657325845859856162019-12-12T12:12:00.001+09:002022-04-19T10:36:09.979+09:00トポロジー入門(第7回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所1E303,203(月曜日3,4限)]</span></div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px; margin: 0px;">
<div style="margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/19/topclass.html">トポロジー入門のHP</a></div>
</div>
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/19/top.html">トポロジー入門演習のHP</a></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
前回は開基を定義しました。</div>
<div style="font-size: 16px;">
開基とは開集合の集合ですが、</div>
<div style="font-size: 16px;">
任意の開集合が開基に属する開集合を使って覆えるものをいいます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
今回は</div>
<u><span style="font-size: large;">可算公理</span></u><br />
<div style="font-size: 16px;">
についてやりました。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定義7.1</b></div>
<div style="font-size: 16px;">
$(X,\mathcal{O})$ が<b>第2可算公理</b>を満たすとは、</div>
<div style="font-size: 16px;">
$X$のある開基として、可算個のものが取れるときをいう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定義7.2</b></div>
<div style="font-size: 16px;">
$(X,\mathcal{O})$ が<b>可分</b>であるとは、ある可算部分集合 $D\subset X$ が存在して、</div>
<div style="font-size: 16px;">
$X=\overline{D}$ となることをいう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このような部分集合 $\overline{D}=X$ となる部分集合のことを<b>稠密部分集合</b>といいます。</div>
<div style="font-size: 16px;">
まず、次の定理を示しました。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定理7.1</b></div>
<div style="font-size: 16px;">
$(X,\mathcal{O})$ が第2可算公理をみたすとき、$X$ は</div>
<div style="font-size: 16px;">
第1可算公理を満たし、可分である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明)</div>
<div style="font-size: 16px;">
$\mathcal{B}$ を $X$ の可算開基とします。</div>
<div style="font-size: 16px;">
$\mathcal{N}^\ast(x)=\{U\in \mathcal{B}|x\in U\}$ とします。</div>
<div style="font-size: 16px;">
このとき、$\mathcal{N}^\ast(x)$ が $x$ の基本近傍系であることを示します。</div>
<div style="font-size: 16px;">
$\forall U\in \mathcal{N}(x)$ とします。</div>
<div style="font-size: 16px;">
このとき、$x\in V\subset U$ となる開集合$V$ が存在して、さらに開基の定義から</div>
<div style="font-size: 16px;">
$x\in W\subset V$ となる $W\in \mathcal{B}$ がなりたち、</div>
<div style="font-size: 16px;">
$\mathcal{N}^\ast(x)$ の定義から、$W\in \mathcal{N}^\ast(x)$ が成り立ちます。</div>
<div style="font-size: 16px;">
ゆえに、$\mathcal{N}^\ast(x)$ は基本近傍系です。</div>
<div style="font-size: 16px;">
$\mathcal{N}^\ast(x)$ は可算集合の部分集合なので、高々可算集合です。</div>
<div style="font-size: 16px;">
つまり、$X$ は第1可算公理を満たします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$\forall U\in \mathcal{B}$ に対して、$a_U\in U$ を選んでおきます。</div>
<div style="font-size: 16px;">
このとき、$D=\{a_U|U\in \mathcal{B}\}$ とすると、$D$ は可算集合です。</div>
<div style="font-size: 16px;">
$\forall x\in X $ と$\forall U\in\mathcal{N}(x)$ を取ると、$x\in V\in \mathcal{B}$ が存在して、</div>
<div style="font-size: 16px;">
$V\subset U$ となります。よって、$a_V\in V$ を満たし、$\emptyset\neq V\cap D\subset U\cap D$ ですので、とくに$U\cap D\neq \emptyset$ となります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
よって、$D$ は $X$ の稠密部分集合ですので、$X$ は可分空間となります。$\Box$</div>
<div style="font-size: 16px;">
さらに次の定理を示します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定理7.2</b> 距離位相空間 $(X,\mathcal{O}_d)$ は、可分であることと第2可算であることは</div>
<div style="font-size: 16px;">
同値である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明)定理7.1から可分であるなら、第2可算であることをしめせばよい。</div>
<div style="font-size: 16px;">
$A\subset X$ を可算稠密集合とします。</div>
<div style="font-size: 16px;">
$\mathcal{B}=\{B_d(x,\frac{1}{n}|x\in A,n\in\mathcal{N}\}$ </div>
<div style="font-size: 16px;">
は可算集合なので、$\mathcal{B}$ が開基であることを示せばよいことになります。</div>
<div style="font-size: 16px;">
$\forall U\in \mathcal{O}$ と $\forall x\in U$ に対して、</div>
<div style="font-size: 16px;">
$B_d(x,\frac{1}{n})\subset U$ となるような$n\in {\mathbb N}$ が存在します。</div>
<div style="font-size: 16px;">
また、$A$ は稠密なので、$a\in B_d(x,\frac{1}{2n})$ となる $a$ が存在します。</div>
<div style="font-size: 16px;">
よって、$x\in B_d(a,\frac{1}{2n})$ がなりたち、</div>
<div style="font-size: 16px;">
$B_d(a,\frac{1}{2n})\subset B_d(x,\frac{1}{n})$ がなりたちます。</div>
<div style="font-size: 16px;">
なぜなら、$z\in B_d(a,\frac{1}{2n})$ なら、</div>
<div style="font-size: 16px;">
$d(z,x)\le d(z,a)+d(a,x)<\frac{1}{2n}+\frac{1}{2n}=\frac{1}{n}$ となるからです。</div>
<div style="font-size: 16px;">
よって、$x\in B_d(x,\frac{1}{2n})\subset B_d(x,\frac{1}{n})\subset U$ となるので</div>
<div style="font-size: 16px;">
$\mathcal{B}$ が開基であることがわかります。$\Box$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
例を考えましょう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>例1</u></div>
<div style="font-size: 16px;">
$({\mathbb R},\mathcal{O}_{d^1})$ を考えると、距離位相空間なので、${\mathbb Q}$</div>
<div style="font-size: 16px;">
が可算稠密集合なので、可分空間になります。</div>
<div style="font-size: 16px;">
よって、第2可算公理を満たし、かつ第1可算公理も満たします。</div>
<div style="font-size: 16px;">
開基として $\mathcal{B}=\{B_d(q,\frac{1}{n})|q\in {\mathbb Q},n\in{\mathbb N}\}$ を</div>
<div style="font-size: 16px;">
とれますが、実際、$\{(a,b)|a,b\in {\mathbb R}\}$を取っておいても</div>
<div style="font-size: 16px;">
開基となります。つまり、開区間が開基となる位相空間です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>例2</u></div>
<div style="font-size: 16px;">
$({\mathbb R},\mathcal{O}_l)$ を以下のように定義します。</div>
<div style="font-size: 16px;">
$\mathcal{O}_l$ を $\mathcal{B}_l=\{[a,b)|a,b\in {\mathbb R}\}$ を開基とする</div>
<div style="font-size: 16px;">
位相空間とし、<b>下限位相</b>もしくは<b>ゾルゲンフライ直線</b>といいます。</div>
<div style="font-size: 16px;">
このときも、${\mathbb Q}$ が可算稠密集合になるので可分な位相空間です。</div>
<div style="font-size: 16px;">
しかし、第2可算を満たしません。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
なぜなら、$\mathcal{B}$ を可算開基として矛盾を導きます。</div>
<div style="font-size: 16px;">
開集合 $[x,x+1)$ に対して $x\in U_x\subset [x,x+1)$ となる開基 $U_x\in \mathcal{B}$ </div>
<div style="font-size: 16px;">
が存在します。よって、$\{U_x|x\in{\mathbb R}\}$ は $\mathcal{B}$ の</div>
<div style="font-size: 16px;">
部分集合ですが、可算濃度ではありません。</div>
<div style="font-size: 16px;">
というのも、$x\neq x’$ に対して $\min(U_x)=x$ であり、$\min(U_{x’})=x’$ であり</div>
<div style="font-size: 16px;">
$U_x\neq U_{x’}$ であるからです。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
よって、可分で第2可算公理を満たさないので、$({\mathbb R},\mathcal{O}_l)$ </div>
<div style="font-size: 16px;">
は距離化できない位相空間ということになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここで次の定義をします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定義7.3</b></div>
<div style="font-size: 16px;">
$X$ 上の位相 $\mathcal{O}_1$ と $\mathcal{O}_2$ が</div>
<div style="font-size: 16px;">
集合として $\mathcal{O}_1\subset \mathcal{O}_2$ を満たすとき、</div>
<div style="font-size: 16px;">
$\mathcal{O}_2$ は $\mathcal{O}_1$ より<b>強い位相</b>、また、</div>
<div style="font-size: 16px;">
$\mathcal{O}_1$ は $\mathcal{O}_2$ より<b>弱い位相</b>といいます。</div>
<div style="font-size: 16px;">
そのとき、$(X,\mathcal{O}_1)\le (X,\mathcal{O}_2)$ とかいたり、</div>
<div style="font-size: 16px;">
単に、$\mathcal{O}_1\le \mathcal{O}_2$ と書いたりします。</div>
<div style="font-size: 16px;">
$\mathcal{O}_1\subset \mathcal{O}_2$ かつ $\mathcal{O}_1\neq \mathcal{O}_2$</div>
<div style="font-size: 16px;">
が成り立つとき、$\mathcal{O}_2$ は $\mathcal{O}_1$ より<b>真に強い</b></div>
<div style="font-size: 16px;">
<b>位相</b>といい、$\mathcal{O}_1<\mathcal{O}_2$ と書きます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$X$ 上の位相はこの $<$ および $\le$ によって順序集合となります。</div>
<div style="font-size: 16px;">
一番強い位相は離散位相空間で、一番弱い位相は、密着位相空間です。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
実は、$\mathcal{O}_{d^1}<\mathcal{O}_l$ がなりたちます。</div>
<div style="font-size: 16px;">
これはレポート問題に出したので解いてみてください。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
最後に次の例を挙げて終わります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>例3</u></div>
<div style="font-size: 16px;">
$(X,\mathcal{O}_{cf})$ を $\mathcal{O}_{cf}=\{U\subset X||U^c|<\aleph_0\}$</div>
<div style="font-size: 16px;">
とします。</div>
<div style="font-size: 16px;">
この位相を補有限位相といいます。</div>
<div style="font-size: 16px;">
$X={\mathbb R}$ として考えます。</div>
<div style="font-size: 16px;">
この位相空間 $({\mathbb R},\mathcal{O}_{cf})$ は第1可算でないことがわかります。</div>
<div style="font-size: 16px;">
このとき、$\forall x\in {\mathbb R}$ に対して、基本近傍系 $\mathcal{N}^\ast(x)$ </div>
<div style="font-size: 16px;">
が可算集合であるとします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このとき、</div>
<div style="font-size: 16px;">
$\{x\}=\cap_{U\in \mathcal{N}(x)}U$ であることを示します。</div>
<div style="font-size: 16px;">
$\subset $ は明らかです。$y\neq x$ を取ると、</div>
<div style="font-size: 16px;">
$X\setminus\{y\}\in \mathcal{N}(x)$ であるから、$y\not\in \cap_{U\in\mathcal{N}(x)}U$</div>
<div style="font-size: 16px;">
となります。よって、$\supset$ がなりたちます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
また、$\{x\}=\cap_{U\in\mathcal{N}(x)}U=\cap_{U\in\mathcal{N}^\ast(x)}U$ </div>
<div style="font-size: 16px;">
がなりたちます。最後の等式は $\subset$ であることは、基本近傍系が</div>
<div style="font-size: 16px;">
近傍系の部分集合であることからわかります。</div>
<div style="font-size: 16px;">
$\supset$ であることは、$y\in \cap_{U\in\mathcal{N}(x)}U$ に対して、</div>
<div style="font-size: 16px;">
$\forall U\in \mathcal{N}(x)$ に対して、ある $V\in \mathcal{N}^\ast(x)$ が</div>
<div style="font-size: 16px;">
存在して、$z\in V\subset U$ であるから、$z\in U$ となります。</div>
<div style="font-size: 16px;">
つまり、$z\in \cap_{U\in\mathcal{N}(x)}U$ であることになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
よって、</div>
<div style="font-size: 16px;">
$$\{x\}=\cap_{U\in\mathcal{N}(x)}U=\cap_{U\in\mathcal{N}^\ast(x)}U$$</div>
<div style="font-size: 16px;">
であり、この補集合をとると、</div>
<div style="font-size: 16px;">
$$\{x\}^c=\cup_{U\in \mathcal{N}(x)}U^c=\cup_{U\in\mathcal{N}^\ast(x)}U^c$$</div>
<div style="font-size: 16px;">
となります。ここで、$X$ が実数の場合、最左辺は非可算集合であり、</div>
<div style="font-size: 16px;">
最右辺は、高々可算集合であり、矛盾します。</div>
<div style="font-size: 16px;">
よって、$({\mathbb R},\mathcal{O}_{cf})$ は第1可算ではないということになります。</div>
<div style="font-size: 16px;">
つまり、距離化可能でもありません。</div>
<div style="font-size: 16px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-91807333459062265952019-12-12T06:24:00.001+09:002022-04-19T10:36:20.098+09:00トポロジー入門(第6回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所1E303,203(月曜日3,4限)]</span></div>
</div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px; margin: 0px;">
<div style="margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/19/topclass.html">トポロジー入門のHP</a></div>
</div>
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/19/top.html">トポロジー入門演習のHP</a></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
今回は近傍系を元に位相空間を構成することを行いました。</div>
<div style="font-size: 16px;">
前回最後に以下を定義しました。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定義5.5</b></div>
<div style="font-size: 16px;">
$X$ を集合とする。$\forall x\in X$ に対して $\mathcal{N}(x)\subset \mathcal{P}(X)$</div>
<div style="font-size: 16px;">
を次を満たすものとする。</div>
<div style="font-size: 16px;">
(1) $\mathcal{N}(x)\neq \emptyset\land (V\in \mathcal{N}(x)\to x\in V)$</div>
<div style="font-size: 16px;">
(2) $\forall V_1,V_2\in \mathcal{N}(x)(V_1\cap V_2\in \mathcal{N}(x))$</div>
<div style="font-size: 16px;">
(3) $\forall V\in \mathcal{N}(x)(V\subset W\to W\in \mathcal{N}(x))$</div>
<div style="font-size: 16px;">
(4) $\forall V\in \mathcal{N}(x)\exists W\in \mathcal{N}(x)(y\in W\to V\in \mathcal{N}(y))$</div>
<div style="font-size: 16px;">
このとき、$\mathcal{N}(x)$ を<b>$x$ の近傍系</b>といい、$\mathcal{N}(x)$ の元を</div>
<div style="font-size: 16px;">
$x$ の<b>近傍</b>という。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
また、近傍系を集めた集合 $\mathcal{N}=\{\mathcal{N}(x)|x\in X\}$ を<b>$X$ の近傍系</b>という</div>
<div style="font-family: -webkit-standard;">
<div style="font-size: 16px;">
ことにします。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここでの話題は</div>
<u><span style="font-size: large;">近傍系と開集合系の同値性</span></u><br />
<div style="font-size: 16px;">
です。</div>
</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
開集合系をもつ位相空間 $(X,\mathcal{O})$ から、<br />
ある近傍系を定義することができます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定義6.1</b> $(X,\mathcal{O})$ を位相空間とする。$\mathcal{N}_{\mathcal{O}}(x)$ を</div>
<div style="font-size: 16px;">
$$\{V\in \mathcal{P}(X)|\exists U\in \mathcal{O}(x\in U\subset V)\}$$</div>
<div style="font-size: 16px;">
として定義する。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このとき、以下を証明をしました。<br />
<br />
<div style="font-family: -webkit-standard;">
<b>定理6.1</b> $(X,\mathcal{O})$ を位相空間とする。$\forall x\in X$ に対して<br />
$\mathcal{N}_{\mathcal{O}}(x)$ は $x$ の近傍系である。</div>
<div style="font-family: -webkit-standard;">
<br />
(証明) まず、開集合 $U\in \mathcal{O}$ と $x\in U$ に対して<br />
$U\in \mathcal{N}_{\mathcal{O}}(x)$ であることは簡単にわかります。<br />
<br />
近傍系の条件(1)から(4)を満たすことを示します。<br />
(1) $X\in \mathcal{N}_{\mathcal{O}}(x)$ を満たすので、成り立ちます。<br />
(2) $V_1,V_2\in \mathcal{N}_{\mathcal{O}}(x)$ とすると、<br />
$x\in U_1\subset V_1$ かつ $x\in U_2\subset V_2$ を満たす $U_i\in \mathcal{O}$ を<br />
が存在します。よって<br />
$x\in U_1\cap U_2\subset V_1\cap V_2$ かつ $U_1\cap U_2\in \mathcal{O}$<br />
を満たすので、$V_1\cap V_2\in \mathcal{N}_{\mathcal{O}}(x)$ となります。<br />
(3) $V\in \mathcal{N}_{\mathcal{O}}(x)$ に対して、$U\in \mathcal{O}$ が存在して、<br />
$x\in U\subset V$ が存在します。$V\subset W$ とすると、<br />
$U\subset V\subset W$ が成り立ち、とくに、$x\in U\subset W$ となるので、<br />
$W\in \mathcal{N}_{\mathcal{O}}(x)$ となります。<br />
(4) $\forall V\in\mathcal{N}_{\mathcal{O}}(x)$ に対して $x\in W\subset V$ を満たす<br />
$W\in \mathcal{O}$ が存在します。<br />
証明の最初の記述から $\forall y\in W$ に対して、<br />
$W\in \mathcal{N}_{\mathcal{O}}(y)$ となります。<br />
よって、$W\subset V$ であるから、$V\in \mathcal{N}_{\mathcal{O}}(y)$ となります。<br />
$\Box$<br />
<br /></div>
<br />
<div style="font-family: -webkit-standard;">
各点 $x$ に対して $x$ の近傍系 $\mathcal{N}(x)$ を</div>
<div style="font-family: -webkit-standard;">
定義したとき、逆に $X$ に位相を定義することができます。</div>
<div style="font-family: -webkit-standard;">
<br /></div>
<div style="font-family: -webkit-standard;">
<b>定義6.2</b> 各点 $x$ に対して $\mathcal{N}(x)$ を近傍系とすると、</div>
<div style="font-family: -webkit-standard;">
$\mathcal{O}_{\mathcal{N}}$を $\{U\subset X|\forall x\in U(U\in\mathcal{N}(x))\}$ と定義する。<br />
<br /></div>
そのとき、以下を示しました。<br />
<br /></div>
<div style="font-family: -webkit-standard;">
<div style="font-size: 16px;">
<b>定理6.2</b> $\mathcal{O}_{\mathcal{N}}$ は開集合系である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明) (I) の成立は省略します。</div>
<div style="font-size: 16px;">
(II) を証明をします。$U_1,U_2\in \mathcal{O}_{\mathcal{N}}$ とします。</div>
<div style="font-size: 16px;">
$\forall x\in U_1\cap U_2$ とすると、$x\in U_1$ かつ $x\in U_2$ が成り立ち、</div>
<div style="font-size: 16px;">
$U_1\in \mathcal{N}(x)$ かつ $U_2\in \mathcal{N}(x)$ を満たします。</div>
<div style="font-size: 16px;">
よって、近傍系の定義の(2)から $U_1\cap U_2\in \mathcal{N}(x)$ を満たすので、</div>
<div style="font-size: 16px;">
$U_1\cap U_2\in \mathcal{O}_{\mathcal{N}}$ が成り立ちます。</div>
<div style="font-size: 16px;">
(III) を証明します。$\{U_\lambda\in \mathcal{O}_{\mathcal{N}}|\lambda\in \Lambda\}$ とします。</div>
<div style="font-size: 16px;">
このとき、$x\in \cup_{\lambda\in \Lambda}U_\lambda$ とすると、</div>
<div style="font-size: 16px;">
$\exists \lambda\in \Lambda(x\in U_\lambda)$ となり、$U_\lambda\in \mathcal{N}(x)$</div>
<div style="font-size: 16px;">
となります。よって 近傍の条件(3) から$\cup_{\lambda\in\Lambda}U_\lambda\in \mathcal{N}(x)$ ですから、</div>
<div style="font-size: 16px;">
$\cup_{\lambda\in \Lambda}U_\lambda\in \mathcal{O}_{\mathcal{N}}$ が成り立ちます。</div>
<div style="font-size: 16px;">
よって、位相の条件が満たされるので、$\mathcal{O}_{\mathcal{N}}$ は開集合系となります。$\Box$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
まとめると $X$ を集合としたとき、写像</div>
<div style="font-size: 16px;">
$$\{\mathcal{O}|\mathcal{O}\text{は $X$ 上の開集合系}\}\to \{\mathcal{N}|\mathcal{N}\text{は$X$ の近傍系}\}$$</div>
<div style="font-size: 16px;">
$$\mathcal{O}\mapsto \mathcal{N}_{\mathcal{O}}$$</div>
<div style="font-size: 16px;">
を定義しました。また、この逆向きの写像</div>
<div style="font-size: 16px;">
$\mathcal{N}\mapsto \mathcal{O}_{\mathcal{N}}$ も定義できました。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
実際、これらの写像が逆写像の関係であることを示しましょう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定理6.3</b> $\mathcal{O}_{\mathcal{N}_{\mathcal{O}}}=\mathcal{O}$ である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明) $U\in \mathcal{O}_{\mathcal{N}_{\mathcal{O}}}$ とします。</div>
<div style="font-size: 16px;">
このとき、$\forall x\in U$ に対して $U\in \mathcal{N}_{\mathcal{O}}(x)$ が成り立ちます。</div>
<div style="font-size: 16px;">
よって $x\in V_x\subset U$ となる $V_x\in \mathcal{O}$ が存在して、</div>
<div style="font-size: 16px;">
$U=\cup_{x\in U}V_x$ が成り立ちます。</div>
<div style="font-size: 16px;">
(このイコールの証明は省略します。)</div>
<div style="font-size: 16px;">
位相の条件 (III) より、$U\in \mathcal{O}$ となります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
逆に、$U\in \mathcal{O}$ が成り立つとすると、$\forall x\in U$ に対して $U\in \mathcal{N}_{\mathcal{O}}(x)$ </div>
<div style="font-size: 16px;">
が成り立ちます。これは、$U\in \mathcal{O}_{\mathcal{N}_{\mathcal{O}}}$ </div>
<div style="font-size: 16px;">
が成り立つことを意味します。$\Box$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
また、次が示されます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定理6.4</b> $\mathcal{N}_{\mathcal{O}_{\mathcal{N}}}=\mathcal{N}$ である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
$\forall x$ に対して、$V\in \mathcal{N}_{\mathcal{O}_{\mathcal{N}}}(x)$ とします。</div>
<div style="font-size: 16px;">
このとき、$\exists U\in \mathcal{O}_{\mathcal{N}}(x\in U\subset V)$ が成り立ちます。</div>
<div style="font-size: 16px;">
また、$U\in \mathcal{N}(x)$ ですから、(3)から $V\in \mathcal{N}(x)$ が成り立ちます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
逆に、$V\in\mathcal{N}(x)$ とします。</div>
<div style="font-size: 16px;">
$V’=\{y\in X|V\in \mathcal{N}(y)\}$ と定義します。</div>
<div style="font-size: 16px;">
このとき、$z\in V’$ に対して $V\in \mathcal{N}(z)$ であるから、とくに $z\in V$ であり、</div>
<div style="font-size: 16px;">
$V’\subset V$ が成り立ちます。</div>
<div style="font-size: 16px;">
また、$\forall y\in V’$ とすると、$V\in \mathcal{N}(y)$ が成りたち、(4)から</div>
<div style="font-size: 16px;">
$\exists W\in \mathcal{N}(y)$ かつ $z\in W\to V\in \mathcal{N}(z)$ が成り立ちます。</div>
<div style="font-size: 16px;">
よって、$z\in V’$ であるから、</div>
<div style="font-size: 16px;">
$W\subset V’$ が成り立ちます。(3) より、$V’\in \mathcal{N}(y)$ が成り立ちます。</div>
<div style="font-size: 16px;">
これは、$V’\in \mathcal{O}_{\mathcal{N}}$ であることがわかります。</div>
<div style="font-size: 16px;">
よって、$V\in \mathcal{N}_{\mathcal{O}_{\mathcal{O}}}$ がわかりました。$\Box$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここで以下を示します。</div>
<div style="font-size: 16px;">
<b>定理6.5</b></div>
<div style="font-size: 16px;">
$O=\{\mathcal{O}|\mathcal{O}\text{は $X$ 上の開集合系}\}$</div>
<div style="font-size: 16px;">
とし、</div>
<div style="font-size: 16px;">
$M=\{\mathcal{N}|\mathcal{N}\text{は$X$ の近傍系}\}$</div>
<div style="font-size: 16px;">
としたとき、$\phi:O\to M\ \ (\mathcal{O}\mapsto \mathcal{N}_{\mathcal{O}})$</div>
<div style="font-size: 16px;">
としたとき $\phi$ は全単射である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明)</div>
<div style="font-size: 16px;">
$\forall \mathcal{O}_1, \mathcal{O}_2\in O$ を取ります。</div>
<div style="font-size: 16px;">
このとき、$\phi(\mathcal{O}_1)=\phi(\mathcal{O}_2)$ であるとすると、</div>
<div style="font-size: 16px;">
$\mathcal{N}_{\mathcal{O}_1}=\mathcal{N}_{\mathcal{O}_2}$ であるから、$\mathcal{O}_{\mathcal{N}_{\mathcal{O}_1}}=\mathcal{O}_{\mathcal{N}_{\mathcal{O}_2}}$ </div>
<div style="font-size: 16px;">
であり、定理6.3から $\mathcal{O}_1=\mathcal{O}_2$ が成り立ちます。</div>
<div style="font-size: 16px;">
よって $\phi$ は単射であることがわかります。$\phi$ が全射であることは</div>
<div style="font-size: 16px;">
定理6.4からわかります。 $\Box$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定理6.6 </b></div>
<div style="font-size: 16px;">
$(X,\mathcal{N}_X)$ と $(Y,\mathcal{N}_Y)$ を $X,Y$ 上の位相空間とする。</div>
<div style="font-size: 16px;">
このとき、$f:X\to Y$ が連続であることの必要十分条件は、$\forall x\in X$</div>
<div style="font-size: 16px;">
において、$\forall V\in \mathcal{N}_Y(f(x))\to \forall f^{-1}(V)\in \mathcal{N}_X(x)$</div>
<div style="font-size: 16px;">
が成り立つことである。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明) </div>
<div style="font-size: 16px;">
$f$ が連続であるとします。</div>
<div style="font-size: 16px;">
このとき、$\forall V\in \mathcal{N}(f(x))$ とすると、</div>
<div style="font-size: 16px;">
$f(x)\in U\subset V$ となる $U\in \mathcal{O}_{\mathcal{N}_Y}(f(x))$ が存在して、</div>
<div style="font-size: 16px;">
$x\in f^{-1}(U)\subset f^{-1}(V)$ が成り立ち、連続性から $f^{-1}(U)\in \mathcal{O}_X$ となります。</div>
<div style="font-size: 16px;">
$f^{-1}(V)\in \mathcal{N}(x)$ が成り立ちます。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
逆を示します。$\forall U\in \mathcal{O}_Y$ とします。</div>
<div style="font-size: 16px;">
このとき、$\forall x\in f^{-1}(U)$ に対して $U\in \mathcal{N}_Y(f(x))$</div>
<div style="font-size: 16px;">
ですから条件より、$f^{-1}(U)\in \mathcal{N}_X(x)$ となるので、</div>
<div style="font-size: 16px;">
$f^{-1}(U)\in \mathcal{O}$ が成り立ちます。</div>
<div style="font-size: 16px;">
よって $f$ は連続となります。$\Box$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
次に</div>
<span style="font-size: large;"><u>基本近傍系</u></span><br />
<div style="font-size: 16px;">
について言及しました。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<b>定義6.3</b></div>
<div style="font-size: 16px;">
$\mathcal{N}(x)$ を $x$ の近傍系とする。</div>
<div style="font-size: 16px;">
$\mathcal{N}^\ast(x)\subset \mathcal{N}(x)$ が基本近傍系であるとは、</div>
<div style="font-size: 16px;">
$\forall V\in \mathcal{N}(x)$ に対して $\exists V’\in \mathcal{N}^\ast(x)(V’\subset V)$ を</div>
<div style="font-size: 16px;">
満たすものをいう。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
つまり、基本近傍系とは、近傍系の部分集合で近傍として基本的なものを<br />
含んでいるものです。基本的というのは、いくらでも小さい近傍が含まれるということを意味します。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
<u>例</u></div>
<div style="font-size: 16px;">
例えば、$X$ 上の距離位相空間 $\mathcal{O}_d$ において、</div>
<div style="font-size: 16px;">
$\mathcal{N}_d=\mathcal{N}_{\mathcal{O}_d}$ と定義します。</div>
<div style="font-size: 16px;">
$\mathcal{N}^\ast_d(x)=\{B_d(x,\epsilon)\subset X|\epsilon>0\}$ としますと、</div>
<div style="font-size: 16px;">
$\mathcal{N}^\ast_d(x)$ は$\mathcal{N}_d$ の基本近傍系となります。<br />
<br />
<b>定義6.3.5</b><br />
$(X,\mathcal{O})$ が第1可算空間であるとは $\forall x\in X$ に<br />
対して高々可算個の基本近傍系をもつことをいう。<br />
<br />
<u>例</u><br />
例えば、任意の距離位相空間 $\mathcal{O}_d$ は第1可算空間です。<br />
基本近傍系 $\mathcal{N}^\ast(x)$ として、$\mathcal{N}^\ast(x)=\{B_d(x,\frac{1}{n})|n\in\mathcal{N}\}$ とすると<br />
$\mathcal{N}^\ast(x)$ は基本近傍系になります。<br />
というのは、$U\mathcal{N}(x)$ に対して、$B_d(x,\epsilon)\subset U$ となる$\epsilon>0$<br />
が存在しますが、$\frac{1}{n}<\epsilon$ を取れば、$B_d(x,\frac{1}{n})\subset B_d(x,\epsilon)\subset U$ となるからです。<br />
<br />
開基を定義します。<br />
<br />
<b>定義6.4</b><br />
$(X,\mathcal{O})$ を位相空間とし、$\mathcal{B}\subset \mathcal{O}$ が開基<br />
であるとは、$\forall U\in \mathcal{O}$ と $\forall x\in U$ に対して、$B\in \mathcal{B}$<br />
が存在して、$x\in B\subset U$ となることをいう。<br />
<br />
最後に次の定理を示して終わりました。<br />
<br />
<b>定理6.7</b><br />
$(X,\mathcal{O})$ を位相空間とする。$\mathcal{B}\subset \mathcal{O}$<br />
が開基であることは、$\forall U\in \mathcal{O}$ に対してある $\mathcal{B}’\subset \mathcal{B}$<br />
が存在して、$U=\cup\mathcal{B}’$ とできることをいう。<br />
<br />
証明<br />
$\mathcal{B}\subset \mathcal{O}$ を開基とします。<br />
このとき $\forall U\in \mathcal{O}$ に対して、$\mathcal{B}’=\{V\subset X|V\in \mathcal{B},x\in V\subset U\}$<br />
とすると、$\cup\mathcal{B}’=U$ となります。<br />
というのは、$\subset$ は、$\forall V\in \mathcal{B}’$ に対して、$V\subset U$<br />
となり、$\supset$ は、$\forall x\in U$ に対して$x\in V\in \mathcal{B}’$<br />
からです。<br />
逆を示します。$\forall U\in \mathcal{B}$ に対して $U=\cup\mathcal{B}’$ かつ $\mathcal{B}’\subset \mathcal{B}$<br />
となる$\mathcal{B}’$ が存在するので、<br />
$\forall x\in U$ に対して、$V\in \mathcal{B}’$ が存在して、$x\in V$ となるので、<br />
$x\in V\subset U$ となります。$\Box$</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-21272749955767293572019-11-23T23:43:00.001+09:002022-04-19T10:36:29.418+09:00トポロジー入門(第5回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<span style="color: orange; font-family: inherit;">[場所1E303,203(月曜日3,4限)]</span></div>
</div>
</div>
</div>
</div>
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="text-align: left;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/19/topclass.html">トポロジー入門のHP</a></div>
</div>
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/19/top.html">トポロジー入門演習のHP</a></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
前回残した定義があったのでそれを説明をしました。<br />
<br />
定義5.1<br />
<div>
$(X,\mathcal{O})$ を位相空間とする。</div>
$F\subset X$ が $F^c\in\mathcal{O}$ であるとき、$F$ を閉集合という。<br />
<br />
<u>定義5.2</u><br />
<div>
閉集合全体からなる集合を<b>閉集合系</b>という。</div>
<div>
閉集合系とは、$\mathcal{C}=\{F\subset X|F^c\in \mathcal{O}\}$ であり、</div>
<div>
以下を満たす。</div>
<div>
(I) $X,\emptyset\in \mathcal{C}$</div>
<div>
(II) $F_1\cdots, F_n$ が有限個の閉集合とすると、$F_1\cup\cdots \cup F_n\in \mathcal{C}$</div>
<div>
を満たす。</div>
<div>
(III) $\{F_\lambda\in \mathcal{C}|\lambda\in \Lambda\}$ を閉集合族とすると$\cup_{\lambda\in \Lambda}F_\lambda\cap \in \mathcal{C}$を満たす。</div>
<div>
<br /></div>
<div>
<br /></div>
<div>
<u>定義5.3</u></div>
<div>
$(X,\mathcal{O}_X)$ と $(Y,\mathcal{O}_Y)$ を位相空間とする。</div>
<div>
$\mathcal{C}_X,\mathcal{C}_Y$ を $X,Y$ の閉集合系とする。</div>
<div>
写像 $f:X\to Y$ が </div>
<div>
$\forall U\in \mathcal{O}_X\Rightarrow f(U)\in \mathcal{O}_Y$ を満たすとき$f$ は<b>開写像</b>という。</div>
<div>
また、</div>
<div>
$\forall F\in \mathcal{C}_X\Rightarrow f(F)\in \mathcal{C}_Y$ を満たすとき $f$ は<b>閉写像</b>という。</div>
<div>
<br /></div>
<div>
また、$f$ が全単射であり、$\forall U\in \mathcal{O}_X\Leftrightarrow f(U)\in \mathcal{O}_Y$</div>
<div>
が成り立つとき、$f$ は同相写像という。</div>
<div>
<br /></div>
<div>
例</div>
<div>
$(0,1)\to {\mathbb R}_{>0}$ を</div>
<div>
$x\mapsto \tan(x)$ は $(0,1)$ と ${\mathbb R}_{>0}$ の間の同相写像を与えます。</div>
<div>
全単射であることはすぐわかります。</div>
<div>
また、連続であることは、$\tan (x)$ が連続関数であることからわかります。</div>
<div>
(連続関数であることは位相空間同士の連続写像であることと同値であるから)</div>
<div>
また、$\text{Arctan}(x)$ が連続であることから、 $\tan(x)$ の逆写像も連続となります。</div>
<div>
このようにして$(0,1)$ と ${\mathbb R}_{>0}$ が連続であることがわかります。</div>
<div>
<br /></div>
<div>
また、${\mathbb R}\to {\mathbb S}^1=\{(x,y)\in{\mathbb R}^2|x^2+y^2=1\}$ </div>
<div>
が全射な連続な開写像であることがわかるのですが、</div>
<div>
これはまた後日行います。</div>
<div>
<br /></div>
<div>
例</div>
<div>
$(X,\mathcal{O})$ を密着位相ではない位相空間とします。</div>
<div>
このとき、</div>
<div>
$i:(X,\mathcal{O})\to (X,\{\emptyset,X\})$ を恒等写像とすると、</div>
<div>
$i$ は連続な全単射で、同相写像ではありません。</div>
<div>
もし同相なら、開集合系の濃度は特に等しくなります。</div>
<div>
<br /></div>
次の定理を示しました。<br />
<br />
定理5.1<br />
$(X,\mathcal{O}_X)$ と $(Y,\mathcal{O}_Y)$ $(Z,\mathcal{O}_Z)$ を位相空間とする。<br />
$f:X\to Y$ と $g:Y\to Z$ が連続写像とする。<br />
このとき、$g\circ f$ も連続写像となる。<br />
<br />
(証明)<br />
$\forall U\in \mathcal{O}_Z$ とすると、$f^{-1}(U)\in \mathcal{O}_Y$ が成り立ち、<br />
さらに、$g^{-1}(f^{-1}(U))=(f\circ g)^{-1}(U)\in \mathcal{O}_X$ が成り立つので、<br />
$g\circ f$ は連続写像となります。<br />
<br />
<br />
今回は、<br />
<u><span style="font-size: large;">位相空間の内部、閉包、境界</span></u><br />
についてやりました。<br />
<br />
まずは、内部と閉包と境界を定義します。<br />
<u>定義5.4</u> $(X,\mathcal{O})$ を位相空間とする。<br />
$A^\circ$ を $A$ に包まれる最大の開集合と定義する。<br />
$\bar{A}$ を $A$ を包む最小の閉集合と定義する。<br />
$\partial A=\bar{A}\setminus A^\circ$ と定義する。<br />
<br />
とくに、$A^\circ$ は開集合であり、$\bar{A}$ は閉集合になります。<br />
ここで次の定理を示しておきます。<br />
<br />
定理5.2<br />
(1) $A^\circ=\{a\in X|\exists U\in \mathcal{O}(a\in U\subset A)\}$<br />
(2) $\bar{A}=\{a\in X|\forall U\in\mathcal{O}(a\in U\to A\cap U\neq \emptyset\}$<br />
(3) $\partial A=\{a\in X|\forall U\in \mathcal{O}(a\in U\to (A\cap U\neq \emptyset\land A^c\cap U\neq \emptyset)\}$<br />
<br />
<span style="font-family: inherit;">(証明)</span><br />
<span style="font-family: inherit;">(1) まず、(1) の右辺を $A’$ とします。</span><br />
<span style="font-family: inherit;">$A’=\cup_{U\subset A,U\in \mathcal{O}}U$ となることを示します。</span><br />
<span style="font-family: inherit;">$x\in A’$ ならば、$U\in \mathcal{O}$ が存在して $x\in U\subset A$ を満たします。</span><br />
<span style="font-family: inherit;">とくに、$x\in \cup_{U\subset A,U\in \mathcal{O}}U$ が成り立ちます。</span><br />
<span style="font-family: inherit;">一方、$x\in \cup_{U\subset A,U\in \mathcal{O}}U$とすると、$\exists U\in \mathcal{O}$</span><br />
<span style="font-family: inherit;">であり、$x\in U$ であるが、$U\subset A$ であることから $x\in A’$ となり、</span><br />
<span style="font-family: inherit;">合わせて、$A’=\cup_{U\subset A,U\in \mathcal{O}}U$ が示せました。</span><br />
<span style="font-family: inherit;"><br /></span>
<span style="font-family: inherit;">最後に、$A’$ が $A$ に包まれる最大の開集合であることを証明をします。</span><br />
<span style="font-family: inherit;">まず、$A’$ は開集合の和集合なので、開集合です。</span><br />
<span style="font-family: inherit;">もし、$A'\subset A''\subset A$ となる開集合 $A’’$ が存在したとすると、</span><br />
<span style="font-family: inherit;">$A’’$ は $A’’\subset A$ かつ $A\in \mathcal{O}$ を満たすので、</span><br />
<span style="font-family: inherit;">$A’’\subset \cup_{U\subset A,U\in \mathcal{O}}U=A'$ であるから、</span><br />
<span style="font-family: inherit;">$A’’=A’$ となります。</span><br />
<span style="font-family: inherit;">つまり、$A’’$ は$A$ に包まれる最大の開集合ということになります。</span><br />
<span style="font-family: inherit;"><br /></span>
<span style="font-family: inherit;">(2) この(2) の右辺を $B’$ とすると、</span><br />
<span style="font-family: inherit;">$B’=\cap_{A\subset F,F\in \mathcal{C}}F$ となることを示します。</span><br />
$(B’)^c=\{a\in X|\exists U\in \mathcal{O}(a\in U\to A\cap U=\emptyset)\}=\{a\in X|\exists U\in \mathcal{O}(a\in U\subset A^c)\}=\cup_{U\subset A^c,U\in\mathcal{O}}U$<br />
よって、<br />
$B’=\cap_{U\subset A^c,U\in\mathcal{O}}U^c=\cap_{A\subset F,F\in\mathcal{C}}F$<br />
<span style="font-family: inherit;">となります。</span><br />
ここで、$A\subset B’’\subset B’$ となる閉集合とすると、<br />
$B’’\supset \cap_{A\subset F,F\in \mathcal{C}}F=B’$ となるので、<br />
$B’=B’’$ となります。<br />
よって、$B’$ は $A$ を包む最小の閉集合ですので、$B’=\bar{A}$ となります。<br />
ゆえに(2) が成り立ちます。<br />
<br />
(3) は省略します。<br />
<br />
このとき、<br />
$A^\circ$ を $A$ の<b>内部</b>といい、$A^\circ$ の点を $A$ の<b>内点</b>といい、<br />
$\bar{A}$ を $A$ の<b>閉包</b>といい、$\bar{A}$ の点を $A$ の<b>触点</b>といいます。<br />
また、$\partial A$ を $A$ の<b>境界</b>といい、$\partial A$ の点を $A$ の<b>境界点</b>といいます。<br />
<br />
次を証明をしました。<br />
定理5.3<br />
$A\in \mathcal{O}\Leftrightarrow A=A^\circ$<br />
$A\in \mathcal{C}\Leftrightarrow A=\bar{A}$<br />
<br />
(証明)<br />
$A\in \mathcal{O}$ であるとすると、$A$ に包まれる開集合の最大は $A$ 自身であり、<br />
$A^\circ =A$ がなりたち、逆に $A=A^\circ$ であるなら $A^\circ$ は開集合であるから<br />
$A\in \mathcal{O}$ が成り立ちます。<br />
<br />
$A\in\mathcal{C}$ であるなら、$A$ を包む最小の閉集合は $A$ 自身が<br />
存在するので、$\bar{A}=A$ となります。逆に、<br />
$\bar{A}=A$ であるなら、$\bar{A}$ は閉集合であるから $A\in \mathcal{C}$ です。<br />
<br />
<br />
定理5.4<br />
$f:X\to Y$ が連続であることは以下とそれぞれ同値である。<br />
(i) $\forall V\in \mathcal{C}_Y$ ならば $f^{-1}(V)\in \mathcal{C}_X$ である。<br />
(ii) $A\subset X\Rightarrow f(\bar{A})\subset \overline{f(A)}$ である。<br />
<br />
(証明) (i) と同値であることはすぐわかるので省略します。<br />
(ii) と同値であることを示します。<br />
もし $f$ が連続であるとします。$A\subset X$ に対して<br />
$f^{-1}(\overline{f(A)})$ は閉集合であり、$A$ を包むので、<br />
$\bar{A}\subset f^{-1}(\overline{f(A)})$ となります。<br />
よって、$f(\bar{A})\subset \overline{f(A)}$ となります。<br />
<br />
もし、$f(\bar{A})\subset \overline{f(A)}$ を満たすとします。<br />
$\forall F\in \mathcal{C}_Y$ とします。<br />
$f(\overline{f^{-1}(F)})\subset\overline{f(f^{-1}(F))}=\overline{F}=F$<br />
よって、$\overline{f^{-1}(F)}\subset f^{-1}(F)\subset \overline{f^{-1}(F)}$ となりますので<br />
$f^{-1}(F)$ は閉集合となります。<br />
よって $f$ は連続となります。$\Box$<br />
<br />
最後に近傍系を定義しました。<br />
定義5.5<br />
$X$ を集合とする。$\forall x\in X$ に対して $\mathcal{N}(x)\subset \mathcal{P}(X)$<br />
を次を満たすものとする。<br />
(1) $\mathcal{N}(x)\neq \emptyset\land (V\in \mathcal{N}(x)\to x\in V)$<br />
(2) $\forall V_1,V_2\in \mathcal{N}(x)(V_1\cap V_2\in \mathcal{N}(x))$<br />
(3) $\forall V\in \mathcal{N}(x)(V\subset W\to W\in \mathcal{N}(x))$<br />
(4) $\forall V\in \mathcal{N}(x)\exists W\in \mathcal{N}(x)(y\in W\to V\in \mathcal{N}(y))$<br />
このとき、$\mathcal{N}(x)$ を <b>$x$ の近傍系</b>といい、$\mathcal{N}(x)$ の元を<br />
$x$ の<b>近傍</b>という。<br />
<br /></div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-31674743775561092622019-11-22T17:43:00.003+09:002022-04-19T10:36:37.022+09:00トポロジー入門(第4回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="font-family: quot; font-size: 16px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<span style="color: orange; font-family: inherit;">[場所1E303,203(月曜日3,4限)]</span></div>
</div>
</div>
</div>
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
<div style="font-family: quot; margin: 0px;">
<div style="margin: 0px;">
<div style="text-align: left;">
<div style="margin: 0px;">
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="margin: 0px;">
<div style="font-size: 16px; margin: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/19/topclass.html">トポロジー入門のHP</a></div>
<div style="margin: 0px;">
<div style="font-size: 16px;">
<div style="font-size: 16px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/19/top.html">トポロジー入門演習のHP</a></div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
今日は位相空間に入ったのですが、その前に</div>
<div style="font-size: 16px;">
前回で残されていた部分をやりました。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
定理4.1 $(X,d)$ を距離空間とする。</div>
<div style="font-size: 16px;">
$A\subset X$ を部分集合とする。</div>
<div style="font-size: 16px;">
$\bar{A}=\{x|d(x,A)=0\}$である。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
(証明) $A’=\{x|d(x,A)=0\}$ と定義します。$\bar{A}=A’$ であることを</div>
<div style="font-size: 16px;">
示します。</div>
<div style="font-size: 16px;">
$x\in \bar{A}$ ならば、$\forall \epsilon>0(B_d(x,\epsilon)\cap A\neq \emptyset)$</div>
<div style="font-size: 16px;">
ですから、$a\in B_d(x,\epsilon)\cap A$ とすると、</div>
<div style="font-size: 16px;">
$0\le d(x,a)<\epsilon$ が成り立ちます。</div>
<div style="font-size: 16px;">
よって、$0\le \inf\{d(x,a)|a\in A\}\le d(x,a)<\epsilon$ であり、$\epsilon>0$ は</div>
<div style="font-size: 16px;">
任意にとることにより、</div>
<div style="font-size: 16px;">
$\inf\{d(x,a)|a\in A\}=0$ でなければならない。</div>
<div style="font-size: 16px;">
よって、$x\in A’$ である。</div>
<div style="font-size: 16px;">
逆に、$x\in A’$ であるとすると、$\forall \epsilon>0$ に対して、</div>
<div style="font-size: 16px;">
$\epsilon$ は、$\{d(x,a)|a\in A\}$ の下界にはならないから</div>
<div style="font-size: 16px;">
ある $a\in A$ が存在して、</div>
<div style="font-size: 16px;">
$0\le d(x,a)<\epsilon$ となります。</div>
<div style="font-size: 16px;">
(もし任意の $a\in A$ に対して、$\epsilon\le d(x,a)$ なら、$\epsilon$ は、$\{d(x,a)|a\in A\}$ の下界ということになって $\epsilon>0$ が下界でないということに矛盾します。)</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
よって、$a\in B_d(x,\epsilon)\cap A$ であるから、$B_d(x,\epsilon)\cap A\neq \emptyset$</div>
<div style="font-size: 16px;">
となる。よって、$a\in \bar{A}$ であることがわかります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
よって、$\bar{A}=A’$ であることがわかりました。$\Box$</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
ここからいよいよ位相空間を始めます。</div>
<div style="font-size: 16px;">
<br /></div>
<u><span style="font-size: large;">位相空間</span></u><br />
<div style="font-size: 16px;">
定理3.2では距離空間の間の連続写像を定義しました。</div>
<div style="font-size: 16px;">
そのとき、距離を用いて定義されましたが</div>
<div style="font-size: 16px;">
そのあと、連続性の条件を、距離を直接使うのではなく、開集合系についての条件として<br />
書き直しました(定理3.2)。<br />
つまり連続性というのは、距離ではなく、開集合が大事だということになります。</div>
<div style="font-size: 16px;">
<br /></div>
<div style="font-size: 16px;">
このことから、なんらかの開集合の定義があれば、<br />
連続性というのは定義できるのだということが<br />
わかります。<br />
距離空間の定義からくる開集合の性質(I),(II),(III)をもつ集合の<br />
集まりを開集合として定義できないか?<br />
となるのです。そして、次の定義に至ります。<br />
<br />
定義4.1<br />
$X$ を集合とする。$\mathcal{O}\subset \mathcal{P}(X)$ が<b>開集合系</b>であるとは<br />
以下を満たすものをいう。<br />
(I) $\emptyset\in \mathcal{O}$ かつ $X\in \mathcal{O}$<br />
(II) $n\in {\mathbb N}$ に対して、$U_1,\cdots,\cap U_n\in \mathcal{O}$ であるとき、$U_1\cap \cdots U_n\in \mathcal{O}$である。<br />
(III) $\{U_\lambda\in \mathcal{O}|\lambda\in \Lambda\}$ であるなら、$\cup_{\lambda\in \Lambda}U_\lambda\in \mathcal{O}$である。<br />
$\mathcal{O}$ を開集合系としたとき、空間と開集合系のペア $(X,\mathcal{O})$<br />
を<b>位相空間</b>という。<br />
<br />
定義4.2<br />
$(X,\mathcal{O}_X)$, $(Y,\mathcal{O}_Y)$ が位相空間とします。<br />
写像 $f:X\to Y$ が<b>連続</b>であるとは、<br />
$\forall U\in \mathcal{O}_Y$ に対して $f^{-1}(U)\in \mathcal{O}_X$<br />
を満たすものをいう。<br />
<br />
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
次の定理を示しましょう。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<br /></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
定理4.2</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$A\in \mathcal{O}\Leftrightarrow \forall a\in A\exists U\in \mathcal{O}(a\in U\subset A)$</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<br /></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
この定理は、距離空間の開集合の定義、</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$U\subset X\Leftrightarrow \forall x\in U\exists\epsilon>0(B_d(x,\epsilon)\subset U)$ </div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
を一般の位相空間に拡張したものと考えられます。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<br /></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
(証明)</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$\Rightarrow$ ですが、$U$ として $A$ 自信をとればよい。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$\Leftarrow$ は、$A\subset X$ が $x\in A\exists U\in \mathcal{O}(x\in U\subset A)$</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
を満たす $U$ を $U_x$ としておきます。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
このとき、$A=\underset{x\in A}\cup U_x$ が成り立ちます。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$A\supset \underset{x\in A}\cup U_x$ かつ $A\subset \underset{x\in A}\cup U_x$ </div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
が成り立つことを示します。</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<br /></div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
(というのも、$U_x\subset A$ であることから $\supset$ が成り立ち、</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-ligatures: normal; font-variant-position: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
$\forall x\in A$ に対して、$x\in U_x$ であることから、$\subset$ が成り立ちます。</div>
よって、$A=\underset{x\in A}{\cup}U_x$ であり、$A$ は開集合のいくつかの<br />
和集合によって得られるから、$A\in \mathcal{O}$ が<br />
成り立ちます。$\Box$<br />
<br />
ここで位相空間の例を与えます。<br />
<br />
例1<br />
$X$ を集合とし、$(X,\mathcal{P}(X))$ は位相の条件を満たすので位相空間です。<br />
このような位相空間を<b>離散位相空間</b>といいます。<br />
<br />
例2<br />
$X$ を集合とし、$(X,\{\emptyset,X\})$ は位相の条件を満たすので位相空間です。<br />
このような位相空間を<b>密着位相空間</b>といいます。<br />
<div>
<br />
例3</div>
</div>
<div style="font-size: 16px;">
$(X,d)$ を距離空間とします。$\mathcal{O}_d$ を距離空間の開集合とします。<br />
距離空間の開集合の定義は前回を見てください。<br />
このとき、$(X,\mathcal{O}_d)$ は位相空間としての開集合系の条件を満たすので、<br />
位相空間となります。<br />
このような位相空間を<b>距離位相空間</b>といいます。<br />
<br />
ある位相空間 $(X,\mathcal{O})$ がこのように $X$ 上の何かの距離 $d$ からくる<br />
距離位相空間と一致するつまり、$\mathcal{O}=\mathcal{O}_d$ となるとき、<br />
$(X,\mathcal{O})$ は<b>距離化可能</b>であるといいます。<br />
<br />
距離の性質をもつ空間を考えたのだから、距離空間が自然に位相空間になることは<br />
わかりますが、距離化可能ではない空間が構成できるのでしょうか。<br />
実際、距離化可能ではない例が存在することを証明します。<br />
<br />
例4<br />
$X=\{1,2\}$ とします。$X$ の上に位相空間を考えます。<br />
$\mathcal{P}(X)=\{\emptyset,\{1\},\{2\},X\}$ ですから、この部分集合として<br />
位相を与えるものを考えることで位相空間が構成できます。<br />
まず、(I)から、$\mathcal{O}$ には $\emptyset$ と $X$ は必ずふくまれるので、<br />
$\{1\}$ が含まれるか含まれないか、$\{2\}$ が含まれるか含まれないか<br />
4パターンあります。<br />
そのうち、どちらも含む場合が離散位相空間で、<br />
どちらも含まない場合は密着位相空間です。<br />
<br />
$\mathcal{O}=\{\emptyset,\{1\},X\}$ としてやると、これも位相の条件を満たします。<br />
この位相空間は、離散位相空間でも密着位相空間でもないですが、<br />
実際距離化可能ではありません。<br />
<br />
<br />
<br />
<br />
定理4.3<br />
$X$ を有限集合とする。$X$ 上の開集合 $\mathcal{O}$<br />
が距離位相空間であるなら、$X$ は離散位相空間である。<br />
<br />
そのために次の命題を用意します。<br />
<br />
命題<br />
$X$ が離散位相空間であることの必要十分条件は、<br />
$\forall x\in X(\{x\}\in \mathcal{O})$ であることである。</div>
<div style="font-size: 16px;">
<br />
(証明)$\Rightarrow$ は、離散位相は、$\mathcal{O}=\mathcal{P}(X)$ ですから<br />
当然 $\forall \{x\}\in \mathcal{O}$ が成り立ちます。<br />
$\Leftarrow$ は、$\forall U\in \mathcal{P}(X)$ に対して、<br />
$U=\underset{x\in U}{\cup}\{x\}$ であり、位相の条件(III)から<br />
$U\in \mathcal{O}$ が成り立ちます。<br />
<br />
上の定理4.2を証明をしましょう。<br />
(証明)$X$ が有限集合とし、その上の距離空間を考えます。<br />
$\delta=\min\{d(x,y)|x,y\in X,x\neq y\}$ をとります。<br />
$X$ の有限性から$\delta>0$ が成り立ちます。<br />
このとき、$B_d(x,\delta/2)=\{x\}$ であることがわかります。<br />
<br />
故に、任意の1点は開集合ですから、上の命題から、$X$ 上のこの位相は<br />
離散位相空間となります。<br />
<br />
よって例4の位相空間 $(\{1,2\},\{\emptyset,\{1\},\{1,2\}\})$ は<br />
距離空間とは一致しないことになります。<br />
<br />
このようにして、距離空間を見本にして距離空間を一般化した位相空間<br />
を定義しましたが、距離空間とは違う空間を位相空間として取り入れることが<br />
できたことになります。<br />
<br />
最後に次の例を考えます。<br />
<br />
例5<br />
距離空間 $({\mathbb R}^2,d_M)$ と $({\mathbb R}^2,d_2)$ を<br />
$d_2({\bf x},{\bf y})=\sqrt{(x_1-y_1)^2+(x_2-y_2)^2}$ と定義し一方、<br />
$d_M({\bf x},{\bf y})=\sum_{i=1}^2|x_i-y_i|$ と定義します。<br />
ここで、${\bf x}=(x_1,x_2)$, ${\bf y}=(y_1,y_2)$ です。</div>
<div style="font-size: 16px;">
このとき、この2つの距離が決める距離位相空間は一致します。<br />
つまり、$\mathcal{O}_{d_M}=\mathcal{O}_{d_2}$ となります。<br />
<br />
(証明) $U\in \mathcal{O}_{d_M}$ とします。<br />
$\forall x\in U$ に対して $x\in B_{d^2}(x,\epsilon)\subset U$ となる $\epsilon>0$ が<br />
存在します。また、$x\in B_{d_M}(x,\epsilon)\subset B_{d_2}(x,\epsilon)$ が成り立ちます。<br />
なぜなら、$\forall z\in B_{d_M}(x,\epsilon)$ とし、$z=(z_1,z_2)$ とすると、<br />
$(|z_1-x_1|+|z_2-x_2|)^2-((z_1-x_1)^2+(z_2-x_2)^2)=2|z_1-x_1||z_2-x_2|\ge 0$<br />
が成りたつからです。<br />
<br />
よって、<br />
$$\epsilon\ge |z_1-x_1|+|x_2-x_2|\ge \sqrt{(z_1-x_1)^2+(z_2-x_2)^2}$$<br />
が成り立ち、$z\in B_{d_2}(x,\epsilon)$ となり、$B_{d_M}(x,\epsilon)\subset B_{d_2}(x,\epsilon)$ となります。<br />
よって、$\mathcal{O}_{d_2}\subset \mathcal{O}_{d_M}$ が成り立ちます。<br />
<br />
一方、$U\in \mathcal{O}_{d_M}$ に対して、<br />
$\forall x\in U$ に対して $B_{d_M}(x,\epsilon)$ となる$\epsilon>0$ が存在し、<br />
$B_{d_M}(x,\epsilon)\subset U$ が成り立ちます。<br />
このとき、$B_{d_2}(x,\frac{\epsilon}{\sqrt{2}})\subset B_{d_M}(x,\epsilon)$ が成り立ちます。<br />
なぜなら、$\forall z\in B_{d_2}(x,\epsilon)$ とすると、<br />
$$2((z_1-x_1)^2+(z_2-x_2)^2)-(|z_1-x_1|+|z_2-x_2|)^2$$<br />
$$=(z_1-x_1)^2+(z_2-x_2)^2-2|z_1-x_1||z_2-x_2|$$<br />
$$\ge (|z_1-x_1|-|z_2-x_2|)^2\ge 0$$<br />
が成りたつからです。<br />
<br />
よって、<br />
$$\frac{\epsilon}{\sqrt{2}}\ge \sqrt{(z_1-x_1)^2+(z_2-x_2)^2}\ge\frac{1}{\sqrt{2}}(|z_1-x_1|+|z_2-x_2|)$$<br />
が成り立つので、$z\in B_{d_M}(x,\epsilon)$<br />
よって、$U\in \mathcal{O}_{d_M}$ が成り立ちます。<br />
つまり、$\mathcal{O}_{d_2}\subset \mathcal{O}_{d_M}$ となり、<br />
<br />
$\mathcal{O}_{d_2}=\mathcal{O}_{d_M}$ が成り立ちます。$\Box$<br />
<br />
このようにして、違う距離でも同じ距離位相空間になってしまう例があります。<br />
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0tag:blogger.com,1999:blog-962211057314600203.post-32750768972929894502019-11-22T11:06:00.002+09:002019-11-22T11:06:49.261+09:00微積分演習F(第2回)<div dir="ltr" style="text-align: left;" trbidi="on">
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<span style="color: orange; font-family: inherit;">[場所1E102(水曜日5限)]</span></div>
</div>
</div>
</div>
</div>
</div>
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<br /></div>
</div>
</div>
</div>
</div>
</div>
</div>
<div style="-webkit-text-stroke-width: 0px; color: black; font-family: &quot; font-size: 16px; font-style: normal; font-variant: normal; font-weight: 400; letter-spacing: normal; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px; orphans: 2; text-align: left; text-decoration: none; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="text-align: left;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="font-size: 16px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<div style="margin-bottom: 0px; margin-left: 0px; margin-right: 0px; margin-top: 0px;">
<a href="http://www.math.tsukuba.ac.jp/~tange/jugyo/19/bis.html">微積分演習FのHP</a></div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<b></b><i></i><u></u><sub></sub><sup></sup><strike></strike><br />
今回はガンマ関数とベータ関数についてやりました。<br />
<br />
<u><span style="font-size: large;">ガンマ関数とベータ関数の定義</span></u><br />
$s>0$を満たす実数とし、$p,q>0$を満たす実数とします。<br />
このとき、ガンマ関数とベータ関数を広義積分<br />
$$\Gamma(s)=\int_0^\infty e^{-x}x^{s-1}dx$$<br />
$$B(p,q)=\int_0^1t^{p-1}(1-t)^{q-1}dt$$<br />
として定義します。<br />
<br />
まず、この広義積分ですが、条件 $s>0$ $p,q>0$ において<br />
これらの広義積分は収束します。<br />
<br />
まずガンマ関数の方からいきます。<br />
$x=\infty$ で広義積分を考えます。<br />
$\int_1^\infty e^{-x}x^{s-1}dx$ が収束するかどうか考えます。<br />
<br />
$s+1<n$ となる自然数 $n$ を取ります。<br />
そのとき、指数関数のテイラー展開から、$e^x\ge \frac{x^{n}}{n!}$<br />
が成り立つので、<br />
$|e^{-x}x^{s+1}|\le |e^{-x}x^n|\le \frac{x^n}{\frac{x^{n}}{n!}}\le n!$<br />
が成り立ちます。<br />
よって、$|e^{-x}x^{s-1}|\le \frac{n!}{x^2}$<br />
であり、広義積分 $\int_1^{\infty}\frac{n!}{x^2}dx$ は収束するので、<br />
優関数法から $\int_1^\infty e^{-x}x^{s-1}dx$ は収束します。<br />
<br />
$x=0$ での広義積分を考えます。<br />
$\int_0^1e^{-x}x^{s-1}dx$ を考えますが、<br />
$s\ge 1$ であれば、$e^{-x}x^{s-1}$ は有限な値ですから広義積分ではなく<br />
通常の積分となり、値は求まります。<br />
$0<s<1$ の場合は $|e^{-x}x^{s-1}|\le \frac{1}{x^{1-s}}$<br />
であり、広義積分 $\int_0^1\frac{1}{x^{1-s}}dx$ は収束するので<br />
やはりこのときも広義積分は収束します。<br />
<br />
ベータ関数についてもやってみます。<br />
$p,q\ge 1$ であれば、<br />
$$\int_0^1t^{p-1}(1-t)^{q-1}dt$$<br />
の被積分関数は $x=0,1$ でも有限な値を持つので、<br />
広義積分ではありません。<br />
つまり、通常の積分として求めることができます。<br />
よって、$0<p,q<1$ であると仮定しておきます。<br />
例えば、$t= 0$ のときの広義積分を考えましょう。<br />
$\int_0^{\frac{1}{2}}t^{p-1}(1-t)^{q-1}dt$<br />
を考えますと、$|\frac{1}{t^{1-p}}(1-t)^{q-1}|\le \frac{1}{2^{q-1}t^{1-p}}$<br />
となり、この積分<br />
$$\int_0^{\frac{1}{2}}\frac{1}{2^{q-1}t^{1-p}}dt=\frac{1}{2^{q-1}}\int_0^{\frac{1}{2}}\frac{1}{t^{1-p}}dt$$<br />
は前回書いたように収束する広義積分でした。<br />
よって、広義積分 $\int_0^{\frac{1}{2}}t^{p-1}(1-t)^{q-1}dt$ も収束することがわかります。<br />
<br />
この関数 $\Gamma(s)$ や $B(p,q)$ を用いて多くの積分を書いていきましょう。<br />
まず、この関数の性質を調べてみると、<br />
以下のことが知られています。<br />
<br />
$$B(p,q)=\frac{\Gamma(p)\Gamma(q)}{\Gamma(p+q)}$$<br />
$$\Gamma(a+1)=a\Gamma (a)\ \ (a>0)$$<br />
$$\Gamma\left(\frac{1}{2}\right)=\sqrt{\pi},\ \ \Gamma(1)=1$$<br />
とくに、$n$ が自然数のときに、<br />
$$\Gamma(n)=(n-1)!$$<br />
となります。<br />
この中で比較的わかりやすいのは、$\Gamma(1)$ であり、<br />
$$\Gamma(1)=\int_0^\infty e^{-x}dx=\left[-e^{-t}\right]_0^\infty=1$$<br />
として直接計算できます。<br />
また、$\Gamma(a+1)=a\Gamma(a)$ も、<br />
$$\Gamma(a+1)=\int_0^\infty x^{a}t^{-x}dx=\left[-x^{a}e^{-x}\right]_0^\infty+a\int_0^\infty x^{a-1}t^{-x}dx=a\Gamma(a)$$<br />
として部分積分だけで求められます。<br />
ここで、$\lim_{x\to \infty }x^{a}e^{-x}=0$<br />
なる極限を使いましたが、これは、$a<n$ となる自然数を取っておいて<br />
$$|x^{a}e^{-x}|=\frac{x^n}{e^x}<\frac{x^n}{\frac{x^{n+1}}{(n+1)!}}\le\frac{(n+1)!}{x}\to 0\ \ (x\to \infty)$$<br />
となるので、挟み撃ちの原理により<br />
$x\to \infty$ において<br />
$$x^ae^{-x}\to 0$$<br />
となることがわかります。<br />
その他の公式についてはここでは詳しくできませんが、この演習の中で<br />
そのうちでてくる方法を用いれば証明をすることができます。<br />
注意してほしいことは、$\Gamma(0)$ の値は求まらないことです。<br />
今のところ、ガンマ関数 $\Gamma(s)$ は $s>0$ だけです。<br />
上の公式を用いると、<br />
$$\Gamma(s)=\frac{\Gamma(s+1)}{s}$$<br />
ですが、$s\to 0$ とすると、右辺の分子は $1$ の有限の値に<br />
収束しますが、分母は $0$ に近づいてしまうので、<br />
結局、$\lim_{s\to 0}\Gamma(s)=\infty$ となってしまいます。<br />
<br />
<span style="font-size: large;"><u>ガンマ関数やベータ関数の公式を用いて積分を計算する</u></span><br />
実際、これらの公式を用いていろいろな積分を求めてみます。<br />
授業中やった計算をもう一度してみます。<br />
$\sin^2x=t$ とおきます。すると、$dt=2\sin x\cos x=2\sqrt{t(1-t)}dx$ ですから、<br />
$$\int_0^{\frac{\pi}{2}}\sin^{2n}xdx=\int_0^1t^{n}\frac{dt}{2\sqrt{t(1-t)}}=\frac{1}{2}\int_0^1t^{n-\frac{1}{2}}(1-t)^{-\frac{1}{2}}dt$$<br />
$$=\frac{1}{2}B\left(n+\frac{1}{2},\frac{1}{2}\right)=\frac{\Gamma\left(n+\frac{1}{2}\right)\Gamma\left(\frac{1}{2}\right)}{2\Gamma\left(n+1\right)}=\frac{(n-\frac{1}{2})(n-\frac{3}{2})\cdots \frac{1}{2}\Gamma(\frac{1}{2})\Gamma(\frac{1}{2})}{2(n!)}$$<br />
$$=\frac{(2n-1)!!}{2^{n+1}n!}\pi=\frac{(2n-1)!!}{(2n)!!}\frac{\pi}{2}$$<br />
となります。ここで、二重階乗は<br />
$(2n-1)!!=(2n-1)(2n-3)\cdots 3\cdot 1$<br />
$(2n)!!=(2n)(2n-2)\cdots 4\cdot 2$<br />
を表します。<br />
<br />
また、$\int_0^\infty e^{-x^2}dx$ も、$x^2=t$ とすると、<br />
$dt=2xdx$<br />
$$\int_0^\infty e^{-x^2}dx=\int_0^\infty e^{-t}\frac{1}{2\sqrt{t}}dt=\frac{1}{2}\int_0^\infty t^{-\frac{1}{2}}e^{-t}dt=\frac{\sqrt{\pi}}{2}$$<br />
となります。<br />
<br />
また $\int_0^1\frac{dx}{\sqrt{1-x^3}}$ は、$x^3=t$ とすることで、$dt=3x^2dx$ であり、<br />
$$\int_0^1\frac{dx}{\sqrt{1-x^3}}=\int_0^1\frac{1}{3\sqrt[3]{t^2}}(1-t)^{-\frac{1}{2}}dt$$<br />
$$=\frac{1}{3}B(\frac{1}{3},\frac{1}{2})=\frac{\Gamma(\frac{1}{3})\Gamma(\frac{1}{2})}{3\Gamma(\frac{5}{6})}=\frac{\Gamma(\frac{1}{3})}{\Gamma(\frac{5}{6})}\frac{\sqrt{\pi}}{3}$$<br />
となります。<br />
<span style="font-size: large;"></span><u></u><br />
<span style="font-size: large;"><u>曲線の長さ</u></span><br />
次に、曲線の長さについての演習を行いました。<br />
平面上に $(x(t),y(t))$ のパラメータをもつ曲線 $C$ の $a\le t\le b$ のときの長さ $l(C)$ を<br />
$$l(C)=\int_a^b\sqrt{(x'(t))^2+(y'(t))^2}dt$$<br />
として計算できます。<br />
<br />
授業中に最後まで計算できなかった計算をしておきます。<br />
(すいません、三角関数で置換し、計算を間違えました。)<br />
<br />
以下もう一度計算しなおしました。<br />
$(t,t^2)$ として定義できる2次関数のグラフの<br />
$0\le t\le 1$ の部分 $C$ の長さ $l(C)$ は<br />
$l(C)=\int_0^1\sqrt{1+4t^2}dt$ のように計算できます。<br />
また、$2t=\sinh \theta$ とおくと、$2dt=\cosh \theta d\theta$ であり、<br />
$2=\sinh \theta$ となるとき、<br />
$4=e^{\theta}-e^{-\theta}\Leftrightarrow e^{2\theta}-4e^\theta-1=0\Leftrightarrow e^\theta=2+\sqrt{5}\Leftrightarrow \theta=\log (2+\sqrt{5})$<br />
なので、$\text{Arcsinh}(2)=\log(2+\sqrt{5})$ となります。<br />
ここで、$\text{Arsinh}(x)$ は $\sinh(x)$ の逆関数を表すことにします。<br />
<br />
また、$\sinh(2z)=2\sinh(z)\cosh(z)$ や、$\cosh^2(z)-\sinh^2(z)=1$ であることを用いると、<br />
$$l(C)=\int_0^{\text{Arsinh}(2)}\frac{\cosh^2\theta}{2} d\theta=\frac{1}{2}\int_0^{\text{Arsinh}(2)}\frac{1+\cosh (2\theta)}{2}d\theta $$<br />
$$=\frac{1}{2}\left[\frac{\theta}{2}+\frac{2\sinh(\theta)\cosh(\theta)}{4}\right]_0^{\text{Arsinh}(2)}=\frac{\log(2+\sqrt{5})}{4}+\frac{4\sqrt{1+4}}{8}$$<br />
$$=\frac{\log(2+\sqrt{5})}{4}+\frac{\sqrt{5}}{2}$$<br />
となります。</div>
Motoo Tange's bloghttp://www.blogger.com/profile/14983497108982514046noreply@blogger.com0