4
$\begingroup$

Let $\mathcal{A}_g$ be the moduli space (stack?) of $g$-dimensional principally polarized abelian varieties.

We have the universal family of abelian varieties $\chi_g\rightarrow \mathcal{A}_g$, where the fiber over any $(A,\theta)$ is the corresponding principally polarized abelian variety. Let $L$ be the universal line bundle corresponding to the universal theta divisor.

If $(A,\theta)\in\mathcal{A}_g$. Let $V$ be a rank $r$ vector bundle on $A$ with chern classes $c_i$.

1) I was told that there is a family of coarse moduli spaces $\mathcal{M}_{\mathcal{A}_{g}}(r,c_i)$ parametrizing basically pairs $(A,F)$ where $A\in \mathcal{A}_g$ and $F$ is a coherent $\mu_{L|_{A}}$ -semistable sheaf on $A$ with rank $r$ and chern classes $c_i$. That is $F\in\mathcal{M}_A(r,c_i)$. Are there any conditions for this space to exist?

2) Consider the fiber product of $\mathcal{M}_{\mathcal{A}_{g}}(r,c_i)\times_{\mathcal{A}_{g}}\chi_g$. Can we find a universal sheaf (or a quasi-universal sheaf) $\mathcal{E}$ on this product, which when restricted to the fiber over any $(A,F)\in\mathcal{M}_{\mathcal{A}_{g}}(r,c_i)$ gives $F$? When can we find such a sheaf?

How do I make these notions precise?

It would be great if someone can also direct me to references of the above especially of (1) and (2). Thanks in advance!

Edits: I have made edits based on Prof. Jason Starr's comment

$\endgroup$
4
  • $\begingroup$ Even for a single $(A,\theta)$, there is not a moduli scheme $\mathcal{M}(r,c_i)$ for that $A$. There is an Artin stack. If you specify a stability condition, there is frequently a projective coarse moduli space of semistable objects. Also, if memory serves, the Brauer group of $\mathcal{A}_g$ equals $\mathbb{Q}/\mathbb{Z}$ (I think I proved that here some years ago). Since this is nontrivial, likely there do not exist universal sheaves over those coarse moduli spaces (I am not sure why you refer to such a universal sheaf as a Poincare "line bundle). $\endgroup$ Commented Dec 10, 2015 at 13:18
  • $\begingroup$ I am sorry, I meant the universal sheaf only! $\endgroup$ Commented Dec 10, 2015 at 13:44
  • $\begingroup$ Prof. @JasonStarr, and when you said there is a projective coarse moduli space of semistable objects, is that for a single $(A,\theta)$ or is it over $\mathcal{A}_g$? $\endgroup$ Commented Dec 10, 2015 at 14:29
  • $\begingroup$ Have you had a look at "Geometry of the moduli spaces of sheaves" by Huybrechts and Lehn? They construct these coarse moduli spaces for polarized projective schemes. They also construct a relative version: a coarse moduli scheme for sheaves on the fibers of a projective morphisms with a relative ample line bundle. Maybe this is a good starting point, see ncatlab.org/nlab/files/HuybrechtsLehn.pdf $\endgroup$
    – Bernie
    Commented Dec 14, 2015 at 12:19

0

You must log in to answer this question.

Browse other questions tagged .