Skip to Content
Math4202Topology II (Lecture 6)

Math4202 Topology II (Lecture 6)

Manifolds

Imbedding of Manifolds

Definition for partition of unity

Let {Ui}i=1n\{U_i\}_{i=1}^n be a finite open cover of topological space XX. An indexed family of continuous function ϕi:X[0,1]\phi_i:X\to[0,1] for i=1,...,ni=1,...,n is said to be a partition of unity dominated by {Ui}i=1n\{U_i\}_{i=1}^n if

  1. supp(ϕi)={xX:ϕi(x)0}Ui\operatorname{supp}(\phi_i)=\overline{\{x\in X: \phi_i(x)\neq 0\}}\subseteq U_i (the closure of points where ϕi(x)0\phi_i(x)\neq 0 is in UiU_i) for all i=1,...,ni=1,...,n
  2. i=1nϕi(x)=1\sum_{i=1}^n \phi_i(x)=1 for all xXx\in X (partition of function to 11)

Existence of finite partition of unity

Let {Ui}i=1n\{U_i\}_{i=1}^n be a finite open cover of a normal space XX (Every pair of closed sets in XX can be separated by two open sets in XX).

Then there exists a partition of unity dominated by {Ui}i=1n\{U_i\}_{i=1}^n.

A more generalized version, If the space is paracompact, then there exists a partition of unity dominated by {Ui}iI\{U_i\}_{i\in I} with locally finite. (Theorem 41.7)

We will prove for the finite partition of unity.

Proof for finite partition of unity

Some intuitions:

By definition for partition of unity, consider the sets Wi,ViW_i,V_i defined as

Wi=fi1((12n,1])fi1([12n,1])Vi=fi1((0,1])supp(fi)UiW_i=f^{-1}_i((\frac{1}{2n},1])\subseteq f^{-1}_i([\frac{1}{2n},1])\subseteq V_i=f^{-1}_i((0,1])\subseteq \operatorname{supp}(f_i)\subseteq U_i V1V1U1V_1\subseteq \overline{V_1}\subseteq U_1

Note that ViV_i is open and ViUi\overline {V_i}\subseteq U_i.

And i=1nVi=X\bigcup_{i=1}^n V_i=X.

and WiW_i is open and WiVi\overline{W_i}\subseteq V_i.

And i=1nWi=X\bigcup_{i=1}^n W_i=X.


Step 1:

\exists V_iopesubsetsope subsetsi=1,\dots,nsuchthatsuch that\overline{V_i}\subseteq U_i,and, and \bigcup_{i=1}^n V_i=X$.

For i=1i=1, consider A1=X(U2U3Un)A_1=X-(U_2\cup U_3\cup \dots \cup U_n). Therefore A1A_1 is closed, and A1U1=XA_1\cup U_1=X.

So A1U1A_1\subseteq U_1.

Note that A1A_1 and XU1X-U_1 are disjoint closed subsets of XX.

Since XX is normal, we can separate disjoint closed subsets A1A_1 and XU1X-U_1.

So we have A1V1V1U1A_1\subset V_1\subseteq \overline{V_1}\subseteq U_1 (by normal space proposition ).

For i=2i=2, note that V1(i=2nUi)=XV_1\cup\left( \bigcup_{i=2}^n U_i\right)=X,

Take A2=X(V1(i=3nUi))A_2=X-\left(V_1\cup\left( \bigcup_{i=3}^n U_i\right)\right) (skipping U2U_2).

Then we have V2V2U2V_2\subseteq \overline{V_2}\subseteq U_2.

For i=ji=j, we have

Aj=X((i=1j1Vi)(i=j+1nUi))A_j=X-\left(\left(\bigcup_{i=1}^{j-1}V_i\right)\cup \left(\bigcup_{i=j+1}^n U_i\right)\right)

and i=1nVi=X\bigcup_{i=1}^n V_i=X.

Repeat the above construction for {Vi}i=1n\{V_i\}_{i=1}^n.

Then we have {Wi}i=1n\{W_i\}_{i=1}^n open and WiWiViViUiW_i\subseteq \overline{W_i}\subseteq V_i\subseteq \overline{V_i}\subseteq U_i.

And i=1nWi=X\bigcup_{i=1}^n W_i=X.

Step 2:

Using Urysohn’s lemma . To construct the partition of unity ϕi\phi_i.

Note

Suppose

  • XX be a normal space
  • Z1,Z2XZ_1,Z_2\subseteq X are closed
  • Z1Z_1 and Z2Z_2 are disjoint

Then:

There exists f:X[0,1]f:X\to[0,1] such that

  • f(Z1)={0}f(Z_1)=\{0\} and f(Z2)={1}f(Z_2)=\{1\}
  • ff is continuous.

Since W1W1V1V1U1W_1\subseteq \overline{W_1}\subseteq V_1\subseteq \overline{V_1}\subseteq U_1,

Note that W1\overline{W_1} and XV1X-V_1 are two disjoint closed subsets of normal space XX

Then we can have f1:X[0,1]f_1:X\to[0,1] such that f1(W1)={0}f_1(\overline{W_1})=\{0\} and f1(XV1)={1}f_1(X-V_1)=\{1\}.

Then we have the remaining list of function f2,,fnf_2,\dots,f_n.

Recall the definition for support of functions supp(fi)={xX:fi(x)>0}\operatorname{supp}(f_i)=\overline{\{x\in X: f_i(x)>0\}}. Since fi(x)=0f_i(x)=0 for xXVix\in X-V_i, we have supp(fi)Vi\operatorname{supp}(f_i)\subseteq \overline{V_i}

Next we need to check i=1nfi(x)=1\sum_{i=1}^n f_i(x)=1 for all xXx\in X.

Note that xX\forall x\in X, since i=1nWi=X\bigcup_{i=1}^n W_i=X, then there exists ii such that xWix\in W_i, thus fi(x)=1f_i(x)=1.

And i=1nfi(x)1\sum _{i=1}^n f_i(x)\geq 1.

Then we do normalization for our value. Set F(x)=i=1nfi(x)F(x)=\sum_{i=1}^n f_i(x).

Since F(x)F(x) is sum of continuous functions, FF is continuous.

Then we define ϕi=fi/F(x)\phi_i=f_i/F(x), since F(x)1F(x)\geq 1, we are safe to divide by F(x)F(x) and ϕi(x)\phi_i(x) is continuous.

And supp(ϕi)=supp(fi)ViUi\operatorname{supp}(\phi_i)=\operatorname{supp}(f_i)\subseteq \overline{V_i}\subseteq U_i.

And i=1nϕi(x)=i=1nfi(x)F(x)=F(x)F(x)=1\sum_{i=1}^n \phi_i(x)=\frac{\sum_{i=1}^n f_i(x)}{F(x)}=\frac{F(x)}{F(x)}=1 for all xXx\in X.

Some Extension

Definition of paracompact space

Locally finite: xX\forall x\in X, \exists open xUx\in U such that UU only intersects finitely many open sets in B\mathcal{B}.

A space XX is paracompact if every open cover AA of XX has a locally finite refinement B\mathcal{B} of AA that covers XX.

Algebraic Topology

Homeomorphism: A topological space XX is homeomorphic to a topological space YY if there exists a homeomorphism f:XYf:X\to Y

  • ff is continuous
  • f1f^{-1} is continuous
  • ff is bijective

Equivalence relation: If \sim satisfies the following:

  • \sim is reflexive xX,xx\forall x\in X, x\sim x
  • \sim is symmetric x,yX,xy    yx\forall x,y\in X, x\sim y\implies y\sim x
  • \sim is transitive x,y,zX,xy,yz    xz\forall x,y,z\in X, x\sim y, y\sim z\implies x\sim z

Homeomorphism is an equivalence relation.

  • Reflexive: identity map
  • Symmetric: inverse map is also homeomorphism
  • Transitive: composition of homeomorphism is also homeomorphism

Main Question: classify topological space up to homeomorphism.

Invariant in Mathematics

Quantities associated with topological spaces that don’t change under homeomorphism.

We want to use some algebraic tools to classify topological spaces.

Last updated on