Mathematics Asked on December 15, 2021
I am new with singular homology
This is a fair but hard exercise for all of my classmate, we all feel dizzy after Imagine what really happen during the procedure of the question and what it really says.
Suppose $X$ is a topological space.
and ${U_{n}mid n geq 0}$ is a open cover of $X$ such that $forall n, U_{n} subseteq U_{n+1}$.
Let $i_{m,n}colon U_{n} rightarrow U_{m}$ and $i_{n}colon U_{n} rightarrow X $ denote the inclusion maps.
The problem is to show that
1) $H_{p} (X) = bigcup^{infty} _{n=0} Im((i_{n_*})_{p})$, and
2) $ker((i_{n_*})_{p}) = bigcup^{infty} _{m=n} ker((i_{m,n*})_{p}) $
Well as far as I knew we have $Delta^{n} = {{ (x_{0}, …, x_{n}) in R^{n+1} | x_{i} geq 0, sum _{i} x_{i}=1 }}$ called n-simplex.
for a singular homology we define a chain complex:
$hspace{3 cm}…rightarrow S_{n+1}(X) rightarrow^{partial _{n+1}} S_{n}(X) rightarrow^{partial _{n}} S_{n-1}(X) rightarrow …$
where:
$hspace{4 cm} S_{n}(X):={{ sigma | sigma: Delta^{n} rightarrow X }}$
and we say $S_{n}(X)$ is a free abelian group based on singular n-simplices.
the chain map is defined as:
$hspace{4 cm} partial_{n}: S_{n}(X)rightarrow S_{n-1}(X)$
$hspace{5 cm}sigma hspace{0.6 cm} rightarrow sum_{i=0}^{n} (-1)^{i}sigma^{(i)}$
where $sigma$ are continouos amd injection maps.
$hspace{4 cm}partial_{n} sigma := sum^{n}_{k=0}(-1)^{k}[p_{0},…,p_{k-1},p_{k+1},..p_{n}]$
on the other hand it means:
$hspace{4 cm}partial_{n}sigma = sigma|_{faces of Delta^{n}}$
As this is a chain complex and by the definition of boundary operation $partial^{n}$ it is can be shown that:
$hspace{4 cm} partial^{n}circ partial^{n+1}=0$
n-th homology group of X is defined as a quotient space:
$hspace{4 cm} H_{n}(x)= frac{kernel partial^{n}}{Image partial^{n+1}}$
and we call $H_{n}(f)$ the induced map, and we usually denote it by $f_{n}*$ or $f_{*}$.
$hspace{4 cm}(i_{n_{*}})_{p}: H_{p}(U_{n})rightarrow H_{p}(X)$
for the first statement we can say, $Delta^{p}$ s are closed and bounded so are compact and since $sigma :Delta^{p} rightarrow X$ are continous maps then It induced a compact space on X so Image of the $sigma$ is also compact.
by e theorem a compact subspace (here $Delta^{p}$) of a space (here X ) that bigger have a covering the subspace have a finite covering {if I am using a right theorem!}.
thus, Image $sigma subseteq bigcup^{m}_{k=0} U_{k}$. meaning that finite number of $U_{k}$s cover the Image $sigma$ since the $U_{n}$s are nested we could find the one k for $U_{k}$ and bring up all $U_{i}$ from 0 to k.
now for $S_{p}(X) = {{sigma | sigma: Delta^{p}rightarrow X }}$ we have:
$hspace{4 cm} S_{p}(X)=S_{p}(bigcup_{0}^{k} U_{i})= S_{p}(U_{k})$
and coresponding to boundary map $partial_{p}: S_{p}(X)rightarrow S_{p-1}(X)$ we can rewrite it to:
$hspace{4 cm}partial_{p}:S_{p}(U_{k})rightarrow S_{p-1}(U_{k})$
remember; $(i_{n*})_p: H_{p}(U_{n}) rightarrow H_{p}(X)$
can say; $bigcup_{0}^{infty} Image(i_{n*})_p = Image bigcup_{0}^{infty} H_{p}(U_{n}) rightarrow H_{p}(X)$
we can show that :
$hspace{4 cm}Image (i_{n*})_{p} subseteq H_{p}(X)$
and so:
$forall nin N;hspace{4 cm} Image (i_{n*})_{p} subseteq H_{p}(U_{k})$
then:
$hspace{4 cm}bigcup ^{infty} _{0} Image (i_{n*})_{p}subseteq H_{p}(X)$
and because:
$hspace{4 cm} H_{p}(U_{k})subseteq bigcup ^{infty} _{0} Image (i_{n*})_{p}$
then:
$hspace{4 cm} H_{p}(X) subseteq bigcup ^{infty} _{0} Image (i_{n*})_{p}$
thus:
$hspace{4 cm} H_{p}(X) = bigcup ^{infty} _{0} Image (i_{n*})_{p}$
Answered by Niloo on December 15, 2021
edit: I think my former answer is clearer, but anyway I try to write the answer as detailed as possible.
Using the compactness of $Delta^p$, the image of any singular simplex $Delta^pto X$ (which is also compact) is contained in some $U_n$. Since any $p$-cycle $sigma$ of $X$ is a finite formal sum $sum_i sigma_i$ of singular simplices $sigma_i: Delta^p to X$, we can find $n$ such that $U_n$ contains the image of all $sigma_i$'s (by taking the maximum of $n_i$'s such that the image of $sigma_i$ is contained in $U_{n_i}$.) This means that there exists $sigma_i': Delta^pto U_n$ such that $i_ncirc sigma_i' = sigma_i$. Define $sigma' = sum_isigma_i'$, so that $(i_n)_ast sigma'=sigma$.
Now let $[sigma]in H_p(X)$ be the homology class represented by the $p$-cycle $sigma$. To prove the first statement, it suffices to prove that $[sigma]$ is in the image of $H_p(U_n)to H_{p}(X)$. This is true because $(i_n)_ast [sigma'] = [sigma]$.
For the second, let $sigma=sum_i{sigma_i}$ (where $sigma_i$'s are singular $p$-simplices) be a $p$-cycle in $U_n$ and suppose $[sigma]in H_p(U_n)$ satisfies $(i_n)_ast [sigma] = 0$. By definition this means that there exists a $(p+1)$-cycle $tau = sum_j tau_j$ in $X$ such that $partial tau =(i_n)_astsigma$ as $p$-cycles of $X$. By exactly the same argument as above, there exists $mgeq n$ such that $U_m$ contains all the images of $tau_j$'s. Now define $(p+1)$-simplices $tau'_j: Delta^{p+1}to U_m$ and a $(p+1)$-cycles $tau' = sum_j tau'_j$ in $U_m$ as above. Obsereve that $(i_m)_astpartial tau' = partial (i_m)_ast tau' = partial tau = (i_n)_ast sigma = (i_m)_ast(i_{m, n})_ast sigma$ as $p$-cycles of $X$. Since $(i_m)_ast: S_p(U_m)to S_p(X)$ is injective, this implies $partialtau' = (i_{m, n})_ast sigma$ as $p$-cycles in $U_m$. This means that, in $H_p(U_m)$, we have $(i_{m, n})_ast[sigma] = [partial tau']=0$, and thus any $[sigma]in ker((i_n)_ast)$ is contained in $ker((i_{m, n})_ast)$ for some $m$.
Answered by Naruki Masuda on December 15, 2021
Get help from others!
Recent Questions
Recent Answers
© 2024 TransWikia.com. All rights reserved. Sites we Love: PCI Database, UKBizDB, Menu Kuliner, Sharing RPP