Skip to content

Commit

Permalink
更新至 Cauchy列与完备性
Browse files Browse the repository at this point in the history
  • Loading branch information
syvshc committed Sep 14, 2019
1 parent 4c36d1f commit a30fe48
Showing 1 changed file with 187 additions and 3 deletions.
190 changes: 187 additions & 3 deletions chapter1.tex
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@ \section{基本概念}
\item $ \tau $ 中任意多元素的并仍是 $ \tau $ 中元素 (任意并);
\item $ \tau $ 中有限多元素的交仍是 $ \tau $ 中的元素(有限交) .
\end{enumerate}
并称 $ (E, \tau) $ 为一个\textbf{拓扑空间}\index{T!拓扑空间}, $ \tau $ 中的元素称为\textbf{开集}\index{K!开集}.
并称 $ (E, \tau) $ 为一个\textbf{拓扑空间}\index{T!拓扑空间}, $ \tau $ 中的元素称为\textbf{开集}\index{K!开集}, 在不引起歧义的情况下, 可以简称 $ E $ 是一个拓扑空间.
\end{Def}
\begin{Rmk}\label{rmk:平凡离散}
对集合 $ E $ , 称 $ \tau=\{ E, \varnothing \} $\textbf{平凡拓扑}\index{P!平凡拓扑}, 称 $ \tau=2^{E} $\textbf{离散拓扑} \index{L!离散拓扑}.
Expand Down Expand Up @@ -59,7 +59,7 @@ \section{基本概念}
$ (E, d) $ 是一个拓扑空间, $ A\subset E $ , 若 $ A^{c} $ 是开集, 则称 $ A $$ E $ 上(关于 $ \tau $ )的\textbf{闭集}.\index{B!闭集}
\end{Def}
\begin{Ex}
$ (E, d) $ 是一个度量空间, 则闭球 $ \bar{B}(x, r)=\{ y: d(y, x)\leqslant r \} $ 是闭集.
$ (E, d) $ 是一个度量空间, 则闭球 $ \baro{B}(x, r)=\{ y: d(y, x)\leqslant r \} $ 是闭集.
\end{Ex}
\begin{Prop}\label{prop:闭集的性质}
闭集具有以下性质:
Expand All @@ -68,4 +68,188 @@ \section{基本概念}
\item 任意多个闭集的交仍是闭集(任意交);
\item 有限多个闭集的并仍是闭集(有限并).
\end{enumerate}
\end{Prop}
\end{Prop}

\subsection{邻域和邻域基}

\begin{Def}[邻域基]\label{def:邻域基}
$ (E, \tau) $ 是一个拓扑空间, $ \alpha\in E $,
\begin{enumerate}[(1)]
\item 若子集 $ V\subset E $ 满足
\[
\exists U\in \tau(x\in U\wedge U\subset V)
\]
则称 $ V $$ x $ 的一个\textbf{邻域}\index{L!邻域}, 并记 $ \mathcal{N}(x) $ 为全体 $ x $ 的邻域的集合, 称为 $ x $\textbf{邻域系};
\item$ \mathcal{N}(x) $ 的子集族 $ \mathcal{B}(x) $ 满足
\[
\forall V\in \mathcal{N}(x)\exists U\in \mathcal{B}(x)(U\subset V)
\]
则称 $ \mathcal{B}(x) $$ x $\textbf{邻域基}\index{L!邻域基}.
\end{enumerate}
\end{Def}
\begin{Rmk}
$ \mathcal{N}(x) $$ x $ 的邻域基; 包含 $ x $ 的所有开集的集合是 $ x $ 的邻域基.
\end{Rmk}
\begin{Ex}
$ (E, d) $ 是一个度量空间, $ x\in E $, 则
\[
\mathcal{B}(x)=\left\{ B(x,\frac{1}{n}): n\geqslant 1 \right\}
\]
$ x $ 的邻域基. (可用来描述极限)
\end{Ex}
\begin{Def}[粘着集]
$ (E, \tau) $ 是一个拓扑空间, $ A\subset E $
\begin{enumerate}[(1)]
\item$ x\in E $ , 若 $ \forall V\in \mathcal{N}(x)((V\cap A)\bs \{ x \}\neq\varnothing) $, 则称 $ x $$ A $\textbf{凝聚点}\index{N!凝聚点};
\item$ x\in A $$ A $ 的凝聚点, 则称 $ x $$ A $\textbf{粘着点}, 记 $ A $ 的粘着点的全体为 $ \baro{A} $ , 称为 $ A $\textbf{粘着集}\index{N!粘着集}, 也即
\[
x\in \baro{A}\Leftrightarrow \forall V\in \mathcal{N}(V\cap A\neq\varnothing).
\]
\end{enumerate}
\end{Def}
\begin{Rmk}
$ E $ 中包含 $ A $ 的最小闭集称为 $ A $\textbf{闭包}\index{B!闭包}, 可知 $ A $ 的闭包与粘着集是相同的, 它们从不同的角度描述了相同的集合.
\end{Rmk}
\begin{Def}[稠密]\label{def:稠密}
$ A\subset E $ , 若 $ \baro{A}=E $ , 则称 $ A $$ E $\textbf{稠密}\index{C!稠密}, 或称 $ A $$ E $ 的一个\textbf{稠子集}
\end{Def}
\begin{Ex}
有利数集 $ \Q $ 与无理数集 $ \mathbb{J} $ 都在 $ \R $ 中稠密.
\end{Ex}
\begin{Def}[内部]\label{def:内部}
$ A\subset E $, 若 $ A\in\mathcal{N}(x) $ , 则称 $ x $$ A $\textbf{内点}, $ A $ 的内点的全体称为 $ A $\textbf{内部}\index{N!内部}, 记作 $ \mathring{A} $$ \mathrm{Int}(A) $.
\end{Def}
\begin{Def}[边界]\label{def:边界}
$ A\subset E $, 称 $ \partial A = \baro{A}\bs\mathring{A} $$ A $\textbf{边界}\index{B!边界}.
\end{Def}
\begin{Ex}
$ E=\R $ , 取 $ A = (2, 3) $, 则 $ \partial A = \{ 2, 3 \} $ ; 取 $ A = [2, 3) $$ \partial A = \{ 2, 3 \} $.

$ E = \R^{2} $, 取 $ A = (2, 3)\times \{0\} $ , 则 $ \partial A = [2, 3]\times \{0\} $ , 此时 $ A\subset \partial A $ ; 取 $ \mathbb{D}=B(0, 1) $ , 则 $ \partial \mathbb{D}=\mathbb{S}^{1} $ (单位圆周).
\end{Ex}
\begin{Thm}\label{thm:闭包和内部的性质}
$ E $ 是一个拓扑空间, $ A, B\subset E $ , 则:
\begin{enumerate}[(1)]
\item $ \baro{\baro{A}}=\baro{A} $, $ \mathring{\mathring{A}}=\mathring{A} $; (幂等性)
\item $ E\bs\mathring{A}=\overline{E\bs A} $;
\item \label{item:闭包内部}$ \overline {A\cup B}=\baro{A}\cup \baro{B} $; $ \mathring{\widehat{A\cap B}}=\mathring{A}\cap\mathring{B} $.
\end{enumerate}
\end{Thm}
\begin{Rmk}
考虑定理~\ref{thm:闭包和内部的性质}~的~\ref{item:闭包内部}~,注意 $ \overline {A\cap B}=\baro{A}\cap \baro{B} $$ \mathring{\widehat{A\cup B}}=\mathring{A}\cup\mathring{B} $ 不一定成立. 比如取 $ A=\Q, B=\mathbb{J} $, 则此时 $ \overline{A\cap B}=\baro{\varnothing}=\varnothing $ , 而 $ \baro{A}\cap \baro{B}=\R\cap\R=\R\neq\overline {A\cup B} $ ; 同时 $ \mathring{\widehat{A\cap B}}=\mathring{\R}=\R $, 但是 $ \mathring{A}\cup\mathring{B}=\varnothing\cup\varnothing=\varnothing\neq \mathring{\widehat{A\cap B}} $
\end{Rmk}
\begin{Def}[拓扑比较]\label{def:拓扑比较}
$ \tau, \tau' $$ E $ 的拓扑, 若 $ \tau'\subset\tau $, 则称 $ \tau $$ \tau' $\textbf{强拓扑}\index{Q!强拓扑}, 也即: $ \tau' $--开集一定是 $ \tau $--开集.
\end{Def}
\begin{Ex}
$ E=\R $ 时, 平凡拓扑 $ \subset $ 自然拓扑 $ \subset $ 离散拓扑.
\end{Ex}
\begin{Def}[拓扑子空间]\label{def:拓扑子空间}
$ (E, \tau) $ 是一个拓扑空间, $ F\subset E $ , 在 $ F $ 上定义
\[
\tau_{F}=\{ U\cap F: U\in\tau \} ,
\]
$ \tau_{F} $ 是一个拓扑, 称为由 $ \tau $ 诱导的拓扑, 并称 $ (F, \tau_{F}) $$ (E, \tau) $\textbf{拓扑子空间}\index{T!拓扑子空间}.
\end{Def}
\begin{Ex}
$ E=\R $ , 取 $ F=(2, 3) $ , 则 $ \tau_{F} $ 中元素是 $ \R $ 中开集;

$ F=[2, 3) $ , 则 $ \forall2<x<3, [2,x) $$ \tau_{F} $ 中元素, 但不是 $ \R $ 中开集.
\end{Ex}
\begin{Ex}
$ (E,d) $ 是度量空间, $ F\subset E $, 取 $ \delta=d|_{F\times F} $$ (F, \delta) $ 是度量空间, 且当 $ E, F $ 分别赋以 $ d, \delta $ 诱导的拓扑时, $ F $$ E $ 的拓扑子空间.
\end{Ex}
\begin{Prop}
$ F\subset E, A\subset F $$ A $$ F $ 中的闭集 $ \Leftrightarrow $ 存在闭集 $ B\subset E $ s.t. $ A=B\cap F $ .
\end{Prop}
\begin{Prf}
由闭集性质可以知道
\[
\begin{aligned}
A \text{\,\,} F \text{\,中闭集\,} & \Leftrightarrow F\bs A \text{\,\,} F \text{\,中开集\,} ;\\
& \Leftrightarrow \exists D\in \tau(F\bs A=D\cap F);\\
& \Leftrightarrow A=D^{c}\cap F
\end{aligned}
\]
其中$ D^{c} $是闭集. \qed
\end{Prf}

\subsection{分离空间}

\begin{Def}[Hausdorff空间]\label{def:Hausdorff空间}
$ (E, \tau) $ 是拓扑空间, 若
\[
\forall x, y\in E(x\neq y)\exists U\in \mathcal{N}(x)\exists V\in\mathcal{N}(x) (U\cap V)=\varnothing
\]
则称 $ (E, \tau) $\textbf{Hausdorff空间}\index{H!Hausdorff空间}或\textbf{分离空间}.
\end{Def}
\begin{Ex}
所有离散拓扑空间都是 Hausdorff空间, 所有度量空间都是 Hausdorff空间, 元素个数多于一个的平凡拓扑空间不是 Hausdorff空间(因为所有元素都只有一个邻域, 即 $ E $ ).
\end{Ex}
\begin{Prop}
以下4个命题相互等价
\begin{enumerate}[(1)]
\item $ E $ 是 Hausdorff空间;
\item $ \forall x\in E $, 其所有闭邻域的交为 $\{ x \}$;
\item $\bigcap \mathcal{N}(x)=\{x\}$;
\item $ \forall F\subset E $, $ F $ 也是Hausdorff空间.
\end{enumerate}
\end{Prop}

\section{完备性}
\subsection{序列的极限}
\begin{Def}[极限]\label{def:极限}
$ (E, \tau) $ 是拓扑空间, $ (x_{n})_{n\geqslant 1} $$ E $ 中的序列, $ x\in E $, 若
\[
\forall V\in\mathcal{N}(x)\exists n_{0}\in\N(n\geqslant n_{0}\Leftarrow x_{n}\in V)
\]
则称 $ (x_{n})_{n\geqslant1} $ \textbf{收敛}于 $ x $, 并称 $ x $ 是序列 $ (x_{n})_{n\geqslant1} $ 的极限, 记作
\[
\tau-\lim_{n\to \infty}x_{n}=x.
\]
当不引起歧义的时候, 前面的 $ \tau- $ 可以省略.
\end{Def}
\begin{Rmk}
\begin{enumerate}[(1)]
\item 上述定义中的 $ \mathcal{N}(x) $ 可换成 $ \mathcal{B}(x) $ ;
\item $ \R $ 中序列的极限至多只有一个, 例如 $\lim\limits_{n\to\infty}\frac{1}{n}=0 $, $ \lim\limits_{n\to\infty} =\infty$(无极限), 取 $ (x_{n})_{n\geqslant1}=\left\{ 1, -1, 1, -1\ldots \right\} $ 也无极限;
\item 一般地, Hausdorff空间中的序列至多只有一个极限, 而平凡拓扑空间中的任意序列收敛到任意元素;
\item 对度量空间 $ (E, d) $ , 有
\[
\lim_{n\to \infty}x_{n}=x \Leftrightarrow \lim_{n\to\infty}d(x_{n}, x)=0,
\]
取邻域基 $ \mathcal{B}(x)=\left\{ B(x, \frac{1}{n}): n\geqslant1 \right\} $即可.
\end{enumerate}
\end{Rmk}
\begin{Prop}
$ E $ 是度量空间, 则 $ A\subset E $是闭集的充要条件是对任意 $ A $中序列 $ (x_{n})_{n\geqslant1} $$ \lim\limits_{n\to\infty}x_{n}=x_{0} $ 时, 都有$ x_{0}\in A $.
\end{Prop}
\begin{Prf}
\textit{必要性}. 若 $ x_{0}\notin A $, 则 $ \exists r>0 (B(x, r)\cap A=\varnothing) $, 而由极限的定义知
\[
\exists n_{0}\in\N(n\geqslant n_{0}\Rightarrow x_{n}\in B(x, r))\Rightarrow x_{n}\in A.
\]
$ (x_{n})_{n\geqslant1}\subset A $ 矛盾.

\textit{充分性}. 只需证 $ A^{c} $ 是开集, 用反证法, 若 $ A^{c} $ 是闭集, 则有
\[
\exists x\in A^{c}\,\forall r>0(B(x, r)\cap A\neq\varnothing),
\]
$ r=1/n, n=1, 2\ldots $ , 则有 $ x_{n}\in B(x, 1/n)\cap A $ . 由构造可知 $ (x_{n})_{n\geqslant1}\subset A $$ x_{n}\to x\,(n\to\infty) $ , 又由已知条件知 $ x\in A $, 矛盾. 故 $ A $ 为闭集. \qed
\end{Prf}
\begin{Def}[序列的粘着值]\label{def:序列的粘着值}
$ (x_{n})_{n\geqslant1} $$ E $ 中的序列, $ x\in E $, 若
\[
\forall V\in\mathcal{N}(x)\,\forall n_{0}\in\N\,\exists n\geqslant n_{0}(x_{n}\in V),
\]
则称 $ x $ 是序列 $ (x_{n})_{n\geqslant1} $\textbf{粘着值}\index{N!粘着值}.
\end{Def}
\begin{Ex}
$ \R $ 中序列 $ (1, -1, 1, -1\ldots) $ 的粘着值为 $ 1, -1 $.
\end{Ex}
\begin{Rmk}
序列的粘着值与集合的粘着点没有关系, 如 $ \R $上的序列 $ (x_{n})_{n\geqslant1}=(1, 0.8, {1}/{2}, {1}/{2}-0.2, {1}/{3}, {1}/{3}-0.2, \ldots, {1}/{n}, {1}/{n}-0.2, \ldots) $, 该序列的粘着值 $ 0, -0.2 $ 均不在序列中, 而集合 $ A=\{ x_{n} \}_{n\geqslant1} $ 的粘着集为 $ A\cup\{ 0,-0.2 \} $.

再取 $ (x_{n})_{n\geqslant1}=(1, 1, \ldots) $ 则该序列的粘着值为$1$, 但 $ 1 $ 不是 $ A=\{ 1 \} $ 的凝聚点.
\end{Rmk}
\section{Cauchy列与完备性}

0 comments on commit a30fe48

Please sign in to comment.