5
$\begingroup$

Can anyone suggest me how to calculate explicitly the Bredon Cohomology of the sphere(at least for 2-dimensional case) with antipodal Z/2-action with constant coefficient system associated to integers?

$\endgroup$
10
$\begingroup$

EDIT: per request.

We can be pretty explicit in this case: the $n$-sphere with antipodal action has a $\Bbb Z/2$-equivariant cell structure whose $k$-skeleton is the $k$-sphere. As a result, we get a cellular chain complex of coefficient systems: $$ 0 \leftarrow C_0 \leftarrow C_2 \leftarrow \cdots \leftarrow C_n $$ where $C_k$ is the coefficient system associated to $(S^k,S^{k-1})$. According to Bredon's definitions, we calculate the value of the coefficient system on the trivial orbit $*$ as $H_k((S^k)^{\Bbb Z/2},(S^{k-1})^{\Bbb Z/2})$, which is trivial; we calculate the value on the orbit $\Bbb Z/2$ as $H_k(S^k,S^{k-1}) = \Bbb Z \oplus \Bbb Z$ where the generator $g$ of $\Bbb Z/2$-action swaps the two factors.

Therefore, in each degree the coefficient group is the same coefficient system $C$. More, one can calculate the cellular chain maps using the same homology calculation (one knows that it has to recover the homology of $S^n$) and find that it is a sequence of maps $$ 0 \leftarrow C \stackrel{1-g}{\leftarrow} C \stackrel{1+g}{\leftarrow} \cdots $$

Now we apply Hom out to any coefficient system $M$. You can show $Hom(C,M)$ is $M(\Bbb Z/2)$. The resulting chain complex is $$ 0 \to M(\Bbb Z/2) \stackrel{1-g}{\to} M(\Bbb Z/2) \stackrel{1+g}{\to} M(\Bbb Z/2) \stackrel{1-g}{\to} \cdots $$ Substituting $M = \Bbb Z$, we get the cochain complex computing the cohomology of $\Bbb{RP}^n$.

This is part of a more general result here when the action is free and properly discontinuous, and the group $\Bbb Z/2$ acts trivially on the value $\Bbb Z$ of the coefficient system on $\Bbb Z/2$. In these cases, the Bredon cohomology coincides with cohomology of the orbit space.

Therefore, we're getting the cohomology of $(S^n) / \Bbb Z/2 = \Bbb{RP}^n$ with coefficients in $\Bbb Z$ (hence $\Bbb Z$ in degree 0, $\Bbb Z/2$ in even positive degrees, and 0 otherwise).

$\endgroup$
1
  • $\begingroup$ :could you please do this problem using the contractions of bredon cohomology made in J.P.May's Equivariant Homotopy and Cohomology Theory Book at pages 20-21? $\endgroup$ Jan 8 '14 at 16:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.