1
$\begingroup$

For $F \subset G$ two algebraic groups, consider a homogeneous space $H$ of the form $G/F$. Now every vector bundle over $H$ is a coherent sheaf, but the converse is not true. What happens in the equivariant setting? Do there exist equivariant sheafs that are not equivariant vector bundles?

$\endgroup$
4
$\begingroup$

I assume that the algebraic group $G$ is smooth and connected, and that you are asking about equivariance for the natural action of $G$ on $H$. There is a quotient morphism $q:G\to H$ that is $G$-equivariant and faithfully flat. Thus, to prove that a $G$-equivariant coherent sheaf on $H$ is locally free, it suffices to check that the pullback $G$-equivariant coherent sheaf on $G$ is locally free. Since $G$ is smooth, for every coherent sheaf $\mathcal{F}$ on $G$, there exists a maximal open subscheme $U$ of $G$ on which $\mathcal{F}$ is locally free (possibly zero), and $U$ is dense in $G$. If $\mathcal{F}$ is $G$-equivariant, then $U$ is $G$-invariant. The only dense $G$-invariant open subset of $G$ is all of $G$. Thus every $G$-equivariant coherent sheaf on $G$ is locally free. Therefore every $G$-equivariant coherent sheaf on $H$ is locally free.

Edit. As Sasha points out, it is unnecessary to pullback to $G$ (it just makes it easier for me to think about).

$\endgroup$
  • 2
    $\begingroup$ Jason, what was the reason of considering the pullback of $\mathcal{F}$ to $G$? You could use the same argument directly on $H$ --- since the action of $G$ on $H$ is transitive the only $G$-invariant open subset of $H$ is all $H$. $\endgroup$ – Sasha Dec 2 '13 at 17:36
  • $\begingroup$ . . . just a thought though - does this mean that the category of equivariant vector bundles is an abelian category? $\endgroup$ – Christian Fischmann Dec 2 '13 at 17:52
  • 2
    $\begingroup$ Yes, and moreover it is equivalent to representations of $F$ (the equivalence is given by taking a fiber of a vector bundle at some chosen point of $H$). $\endgroup$ – Sasha Dec 2 '13 at 18:11

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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