The \(y\)-genus of the moduli space of \(\mathrm{PGL}_n\)-Higgs bundles on a curve (for degree coprime to \(n\)) (Q2016421): Difference between revisions

The article under review explains a method to compute the cohomology of the moduli space of semistable Higgs bundles of coprime rank and degree on a smooth projective curve, using wall-crossing arguments. This work is based on the strategy announced in a previous work of the authors together with \textit{A. Schmitt} [``On the motives of moduli of chains and Higgs bundles'', to appear in J. Eur. Math. Soc., preprint, \url{arXiv:1104.5558}], giving a complete answer to the computation of the cohomology of such moduli spaces. The method relies on the classical argument of variating the stability parameter to study the moduli spaces of chains, from the stack-theoretic point of view. The regions of the space of stability parameters where the notion of stability changes are called walls. Those semistable chains with respect to a stability condition which are not with respect to another one, at the crossing of a wall, can be described as a union of Harder-Narasimhan strata, meaning chains whose Harder-Narasimhan filtration shares the same numerical invariants. This idea is used to compute the cohomology of the moduli space for different stability parameters by recursion. By studying the necessary conditions for the non emptyness of the moduli space of semistable chains (c.~f. Section 4), it turns out that there is a finite number of these Harder-Narasimhan strata (c.~f. Corollary 6). In section 5, there is a proof of Theorem B, asserting that we can compute recursively the cohomology ring of the moduli space of semistable chains in terms of symmetric powers of the cohomology class of the curve. The main application is the computation of the \(y\)-genus (c.~f. Theorem A), defined for varieties \(X\) over \(\mathbb{C}\) as \[ H_{y}(X,y):=\sum_{p,q\geq 0}(-1)^{k}h_{c}^{k;p,q}(X)y^{q}\; , \] of the space of \(\mathrm{PGL}_{n}\)-Higgs bundles for any rank \(n\). This proves a conjecture of Hausel.