Theorem 19.6 Let \(f: A \rightarrow \prod_{\alpha \in J} X_{\alpha}\) be given by the equation

\[
f(a) = (f_{\alpha}(a))_{\alpha \in J},
\]

where \(f_{\alpha}: A \rightarrow X_{\alpha}\) for each \(\alpha\). Let \(\prod X_{\alpha}\) have the product topology. Then the function \(f\) is continuous if and only if each function \(f_{\alpha}\) is continuous.

Comment: This is an extension of Theorem 18.4, where only two component spaces are involved.

Proof: a) First, we prove the projection map is continuous, which is defined on the Cartesian space constructed from a \(J\)-tuple of component spaces .

For all \(\beta \in J\), let \(\pi_{\beta}: \prod X_{\alpha} \rightarrow X_{\beta}\) be the projection map. For arbitrary open set \(V_{\beta}\) in \(X_{\beta}\), its pre-image under \(\pi_{\beta}\) is \(\pi_{\beta}^{-1}(V_{\beta})\), which is a subbasis element for the product topology on \(\prod X_{\alpha}\). Therefore, \(\pi_{\beta}^{-1}(V_{\beta})\) is open and the projection map \(\pi_{\beta}\) is continuous.

Next, we notice that for all \(\alpha \in J\), the coordinate function \(f_{\alpha}: A \rightarrow X_{\alpha}\) is a composition of the two continuous functions \(f\) and \(\pi_{\alpha}\), i.e. \(f_{\alpha} = \pi_{\alpha} \circ f\). Then according to Theorem 18.2 (c), \(f_{\alpha}\) is continuous.

Remark: Because the box topology is finer than the product topology, the projection map is also continuous when the box topology is adopted for \(\prod X_{\alpha}\). Therefore, this part of the theorem is true for both product topology and box topology.

b) Analysis: To prove the continuity of a function, showing that the pre-image of any subbasis element in the range space is open in the domain space is more efficient than using basis element or raw open set in the range space. In addition, the subbasis element for the product topology on \(\prod X_{\alpha}\) has the form \(\pi_{\beta}^{-1}(U_{\beta})\) with \(U_{\beta}\) being a single coordinate component and open in \(X_{\beta}\). This is the clue of the proof.

For all \(\beta \in J\) and arbitrary open set \(U_{\beta}\) in \(X_{\beta}\), we have \(f_{\beta}^{-1}(U_{\beta}) = f^{-1} \circ \pi_{\beta}^{-1}(U_{\beta})\). Because \(f_{\beta}\) is continuous and \(U_{\beta}\) is open, \(f_{\beta}^{-1}(U_{\beta})\) is open. In addition, \(\pi_{\beta}^{-1}(U_{\beta})\) is an arbitrary subbasis element for \(\prod X_{\alpha}\) with the product topology, whose pre-image under \(f\) is just the open set \(f_{\beta}^{-1}(U_{\beta})\), therefore \(f\) is continuous.

Remark: Part b) of this theorem really depends on the adopted topology for \(\prod X_{\alpha}\), which can be understood as below.

At first, we will show that for all \(\vect{U} = \prod U_{\alpha}\) being a subset of \(\prod X_{\alpha}\), \(f^{-1}(\vect{U}) = \bigcap_{\alpha \in J} f_{\alpha}^{-1}(U_{\alpha})\).

For all \(x \in f^{-1}(\vect{U})\), because \(f(x) \in \vect{U}\), then for all \(\alpha \in J\), \(f_{\alpha}(x) \in U_{\alpha}\), hence \(x \in \bigcap_{\alpha \in J} f_{\alpha}^{-1}(U_{\alpha})\) and \(f^{-1}(\vect{U}) \subset \bigcap_{\alpha \in J} f_{\alpha}^{-1}(U_{\alpha})\).

On the other hand, for all \(x \in \bigcap_{\alpha \in J} f_{\alpha}^{-1}(U_{\alpha})\), we have for all \(\alpha \in J\), \(f_{\alpha}(x) \in U_{\alpha}\). Therefore, \(f(x) \in \vect{U}\) and \(x \in f^{-1}(\vect{U})\). Hence \(\bigcap_{\alpha \in J} f_{\alpha}^{-1}(U_{\alpha}) \subset f^{-1}(\vect{U})\).

Next, if we assign the product topology to \(\prod X_{\alpha}\), for any \(\vect{U} = \prod U_{\alpha}\) with \(U_{\alpha}\) open in \(X_{\alpha}\) and only a finite number of them not equal to \(X_{\alpha}\), it is a basis element of the product topology. Let the set of all indices with which \(U_{\alpha} \neq X_{\alpha}\) be \(\{\alpha_1, \cdots, \alpha_n\}\) and also notice that when \(U_{\alpha} = X_{\alpha}\), \(f_{\alpha}^{-1}(U_{\alpha}) = A\), we have

\[
f^{-1}(\vect{U}) = \bigcap_{\alpha \in J} f_{\alpha}^{-1}(U_{\alpha}) = \bigcap_{i=1}^n f_{\alpha_i}^{-1}(U_{\alpha_i}),
\tag{*}
\label{eq:intersection}
\]

where those \(f_{\alpha}^{-1}(U_{\alpha})\) with \(\alpha \notin \{\alpha_1, \cdots, \alpha_n\}\) do not contribute to the intersection. This indicates that \(f^{-1}(\vect{U})\) is a finite intersection of open sets which is still open. Hence \(f\) is continuous.

However, if the box topology is adopted for \(\prod X_{\alpha}\), qualitatively speaking, because the topology for the range space becomes finer, according to our previous post, it makes a function to be continuous more difficult. Specifically in this theorem, \(f^{-1}(\vect{U})\) in \eqref{eq:intersection} can be an intersection of infinite number of open sets \(U_{\alpha}\) not equal to \(X_{\alpha}\). Thus \(f^{-1}(\vect{U})\) may not be open anymore.

After understanding this point, it is not difficult to construct a counter example for part b) as below.

Let \(f: \mathbb{R} \rightarrow \mathbb{R}^{\omega}\) be defined as \(f(t) = (t, t, \cdots)\). Select a basis element \(\vect{U}\) in \(\mathbb{R}^{\omega}\) such that the intersection of all its coordinate components is not open. For example, \(\vect{U} = \prod_{n=1}^{\infty} (-\frac{1}{n}, \frac{1}{n})\), which is a neighborhood of \(f(0) = (0, 0, \cdots)\).

For any basis element \((a, b)\) in \(\mathbb{R}\) containing \(0\), with \(a < 0\) and \(b > 0\), by letting \(\delta = \min\{-a, b\}\), we have \((-\delta, \delta) \subset (a, b)\) and \(0 \in (-\delta, \delta)\). The image of \((-\delta, \delta)\) under \(f\) is \(\prod_{n=1}^{\infty} (-\delta, \delta)\). Then there exist an \(n_0 \in \mathbb{Z}_+\) such that \((-\delta, \delta)\) is not contained in \((-\frac{1}{n_0}, \frac{1}{n_0})\). Therefore, \(\pi_{n_0}(f((-\delta, \delta)))\) is not contained in \(\pi_{n_0}(\vect{U})\) and \(\pi_{n_0}(f((a, b)))\) is not contained in \(\pi_{n_0}(\vect{U})\). Hence the image of \((a, b)\) under \(f\) is not contained in \(\vect{U}\). This contradicts Theorem 18.1 (4) and \(f\) is not continuous.

James Munkres Topology: Theorem 19.6的更多相关文章

  1. James Munkres Topology: Theorem 20.4

    Theorem 20.4 The uniform topology on \(\mathbb{R}^J\) is finer than the product topology and coarser ...

  2. James Munkres Topology: Theorem 20.3 and metric equivalence

    Proof of Theorem 20.3 Theorem 20.3 The topologies on \(\mathbb{R}^n\) induced by the euclidean metri ...

  3. James Munkres Topology: Theorem 16.3

    Theorem 16.3 If \(A\) is a subspace of \(X\) and \(B\) is a subspace of \(Y\), then the product topo ...

  4. James Munkres Topology: Sec 18 Exer 12

    Theorem 18.4 in James Munkres “Topology” states that if a function \(f : A \rightarrow X \times Y\) ...

  5. James Munkres Topology: Sec 22 Exer 6

    Exercise 22.6 Recall that \(\mathbb{R}_{K}\) denotes the real line in the \(K\)-topology. Let \(Y\) ...

  6. James Munkres Topology: Sec 22 Exer 3

    Exercise 22.3 Let \(\pi_1: \mathbb{R} \times \mathbb{R} \rightarrow \mathbb{R}\) be projection on th ...

  7. James Munkres Topology: Lemma 21.2 The sequence lemma

    Lemma 21.2 (The sequence lemma) Let \(X\) be a topological space; let \(A \subset X\). If there is a ...

  8. James Munkres Topology: Sec 37 Exer 1

    Exercise 1. Let \(X\) be a space. Let \(\mathcal{D}\) be a collection of subsets of \(X\) that is ma ...

  9. James Munkres Topology: Sec 22 Example 1

    Example 1 Let \(X\) be the subspace \([0,1]\cup[2,3]\) of \(\mathbb{R}\), and let \(Y\) be the subsp ...

随机推荐

  1. DIV正确打开方式 ~~~~哈哈哈

    <div style='margin-left:25px;margin-right:25px;margin-top:10px;height:350px;min-height:50px;backg ...

  2. Flask 模板系统

    模板 基本数据类型 可以执行python语法,如:dict.get(), list['xx'] 比django 更加亲近于 python 传入函数 - django,自动执行 - flask,不自动执 ...

  3. vim命令编辑Host文件

    用vim命令打开Host文件 [root@localhost /]# vim /etc/hosts 按"i"键,进入编辑模式按"Esc"键,退出编辑模式 按&q ...

  4. ajax实现长连接

    项目需求:需要实时的读取日志文件里的数据,并且使用Echart实时更新折线图. 使用ajax实现客户端与服务器端的数据传输. 目的:我想通过ajax与服务器建立一个长连接,服务器会不断的传输数据给前台 ...

  5. sublime中编译的sass如何改变css输出风格?【这里有答案】

    由于在网上找了一遍没找到如果在sublime中将sass编译的css转换成为自己喜欢的风格,所以换了一种思路搜索到了答案,这里我将讲述如果更改. 首先sass总共有四种编译风格 (1) nested( ...

  6. JavaScript数据类型 Boolean布尔类型

    前言 布尔值Boolean类型可能是三种包装对象Number.String和Boolean中最简单的一种.Number和String对象拥有大量的实例属性和方法,Boolean却很少.从某种意义上说, ...

  7. SpringBoot系列: 制作Docker镜像的全过程

    本文主要参考了 https://segmentfault.com/a/1190000016449865 , 感谢作者的付出. 另外,  在本文中, 演示了Windows+Maven+Docker To ...

  8. sqlserver 获取所有表的字段类型等信息

    USE [MultipleAnalysisDataFY] GO /****** Object: View [dbo].[selectfieldtype] Script Date: 2018/11/7 ...

  9. nginx禁止ip登录,只允许域名访问

    公司要求,线上solr.kibana要求只能通过域名进行访问,禁止用ip+端口进行登录访问,那么,下面介绍下我是如何实现的 1.禁止ip,允许域名访问 如下图,默认安装好nginx,不让ip方式访问, ...

  10. js中精度问题以及解决方案

    js中的数字按照IEEE 754的标准,使用64位双精度浮点型来表示.其中符号位S,指数位E,尾数位M分别占了1,11,52位,并且在ES5规范中指出了指数位E的取值范围是[-1074, 971]. ...