Skip to Content
Math4201Topology I (Lecture 27)

Math4201 Topology I (Lecture 27)

Continue on compact spaces

Compact spaces

Heine-Borel theorem

A subset KRnK\subseteq \mathbb{R}^n is compact if and only if it is closed and bounded with respect to the standard metric on Rn\mathbb{R}^n.

Definition of bounded

ARnA\subseteq \mathbb{R}^n is bounded if there exists cR>0c\in \mathbb{R}^{>0} such that d(x,y)<cd(x,y)<c for all x,yAx,y\in A.

Proof for Heine-Borel theorem

Suppose kRnk\subseteq \mathbb{R}^n is compact.

Since Rn\mathbb{R}^n is Hausdorff, KRnK\subseteq \mathbb{R}^n is compact, so KK is closed subspace of Rn\mathbb{R}^n. by Proposition of compact subspaces with Hausdorff property.

To show that KK is bounded, consider the open cover with the following balls:

B1(0),B2(0),...,Bn(0),...B_1(0), B_2(0), ..., B_n(0), ...

Since KK is compact, there are n1,...,nkNn_1, ..., n_k\in \mathbb{N} such that Ki=1kBni(0)K\subseteq \bigcup_{i=1}^k B_{n_i}(0). Note that Bni(0)B_{n_i}(0) is bounded, so KK is bounded. x,yBni(0)\forall x,y\in B_{n_i}(0), d(x,y)<2nid(x,y)<2n_i. So KK is bounded.


Suppose KRnK\subseteq \mathbb{R}^n is closed and bounded.

First let M=[a1,b1]×[a2,b2]××[an,bn]M=[a_1,b_1]\times [a_2,b_2]\times \cdots \times [a_n,b_n].

This is compact because it is a product of compact spaces.

Since KK is bounded, we can find [ai,bi][a_i,b_i]s such that KMK\subseteq M.

Since KK is closed subspace of Rn\mathbb{R}^n, KK is closed in MM.

Since any closed subspace of a compact space is compact, KK is compact.

Warning

This theorem is not true for general topological spaces.

For example, take X=B1(0)X=B_1(0) with the standard topology on Rn\mathbb{R}^n.

Take K=B1(0)K=B_1(0), this is not compact because it is not closed in Rn\mathbb{R}^n.

Extreme Value Theorem

If f:XRf:X\to \mathbb{R} is continuous map with XX being compact. Then ff attains its minimum and maximum.

Proof

Let M=sup{f(x)xX}M=\sup\{f(x)\mid x\in X\} and m=inf{f(x)xX}m=\inf\{f(x)\mid x\in X\}.

We want to show that there are xm,xMXx_m,x_M\in X such that f(xm)=mf(x_m)=m and f(xM)=Mf(x_M)=M.

Consider the open covering of XX given as

{Uαf1((,α))}αR\{U_\alpha\coloneqq f^{-1}((-\infty, \alpha))\}_{\alpha\in \mathbb{R}}

If XX doesn’t attain its maximum, then this is an open covering of XX:

  1. UαU_\alpha is open because ff is continuous and (,α)(-\infty, \alpha) is open in R\mathbb{R}.
  2. αRUα=X\bigcup_{\alpha\in \mathbb{R}} U_\alpha = X because for any xXx\in X, by the assumption there is xXx'\in X with f(x)<f(x)f(x)<f(x') (otherwise f(x)f(x) is the maximum value). Then xUf(x)x\in U_{f(x')}.

So there is an open covering of XX and hence it’s got a finite subcover {Uαi}i=1n\{U_{\alpha_i}\}_{i=1}^n.

X=i=1nUαi=i=1nf1((,αi))=f1(,αk)X=\bigcup_{i=1}^n U_{\alpha_i}=\bigcup_{i=1}^n f^{-1}((-\infty, \alpha_i))=f^{-1}(-\infty, \alpha_k)

and α1α2αn\alpha_1\leq \alpha_2\leq \cdots \leq \alpha_n. There is xix_i such that αi=f(xi)\alpha_i=f(x_i).

Note that xkUαkx_k\notin U_{\alpha_k} because f(xk)>αkf(x_k)>\alpha_k. So xkXx_k\notin X. This contradicts the assumption that XX doesn’t attain its maximum.

Theorem of uniform continuity

Let f:(X,d)(X,d)f:(X,d)\to (X',d') be a continuous map between two metric spaces. Let XX be compact, then for any ϵ>0\epsilon > 0, there exists δ>0\delta > 0 such that for any x1,x2Xx_1,x_2\in X, if d(x1,x2)<δd(x_1,x_2)<\delta, then d(f(x1),f(x2))<ϵd'(f(x_1),f(x_2))<\epsilon.

Definition of uniform continuous function

ff is uniformly continuous if for any ϵ>0\epsilon > 0, there exists δ>0\delta > 0 such that for any x1,x2Xx_1,x_2\in X, if d(x1,x2)<δd(x_1,x_2)<\delta, then d(f(x1),f(x2))<ϵd'(f(x_1),f(x_2))<\epsilon.

Example of uniform continuous function

Let f(x)=x2f(x)=x^2 on R\mathbb{R}.

This is not uniformly continuous because for fixed ϵ>0\epsilon > 0, the interval δ\delta will converge to zero as x1,x2x_1,x_2 goes to infinity.


However, if we take f[0,1]f\mid_{[0,1]}, this is uniformly continuous because for fixed ϵ>0\epsilon > 0, we can choose δ=ϵ\delta = \epsilon.

Lebesgue number lemma

Let XX be a compact metric space and {Uα}αI\{U_\alpha\}_{\alpha\in I} be an open cover of XX. Then there is δ>0\delta>0 such that for any two points x1,x2Xx_1,x_2\in X with d(x1,x2)<δd(x_1,x_2)<\delta, there is αI\alpha\in I such that x1,x2Uαx_1,x_2\in U_\alpha.

Proof of uniform continuity theorem

Let ϵ>0\epsilon > 0. be given and consider

{f1(Bϵ/2d((x))}xX\{f^{-1}(B_{\epsilon/2}^{d'}((x'))\}_{x'\in X'}

We claim that there is an open covering of XX.

  1. f1(Bϵ/2d((x)))f^{-1}(B_{\epsilon/2}^{d'}((x'))) is open because ff is continuous and Bϵ/2d((x))B_{\epsilon/2}^{d'}((x')) is open in XX'.
  2. X=xXf1(Bϵ/2d((x)))X=\bigcup_{x'\in X'} f^{-1}(B_{\epsilon/2}^{d'}((x'))) because for any xXx\in X, xf1(Bϵ/2d((f(x)))x\in f^{-1}(B_{\epsilon/2}^{d'}((f(x))).

Since XX is compact, there is a finite subcover {f1(Bϵ/2d((x)))}i=1n\{f^{-1}(B_{\epsilon/2}^{d'}((x')))\}_{i=1}^n.

By Lebesgue number lemma, there is δ>0\delta>0 such that for any two points x1,x2Xx_1,x_2\in X with d(x1,x2)<δd(x_1,x_2)<\delta, there is xXx'\in X' such that x1,x2f1(Bϵ/2d((x)))x_1,x_2\in f^{-1}(B_{\epsilon/2}^{d'}((x'))).

So f(x1),f(x2)Bϵ/2d((x))f(x_1),f(x_2)\in B_{\epsilon/2}^{d'}((x')).

Apply the triangle inequality with d(x1,x)d'(x_1,x') and d(x2,x)d'(x_2,x'), we have d(f(x1),f(x2))<2ϵ/2=ϵd'(f(x_1),f(x_2))<2\epsilon/2=\epsilon.

Last updated on