McKay correspondence for symplectic quotient singularities. (Q1608569)

Let \(V\) be a finite dimensional complex vector space and \(G\subset \text{SL}(V)\) a finite subgroup. Let \(X:=V/G\). Then two natural questions arise: 1) when does \(X\) admit a crepant resolution of singularities \(f\colon Y\to X\), and 2) if such a resolution exists, what can be said about the homology \(H_*(Y,\mathbb Q)\)? In dimension \(2\) J. McKay proved that there always exists a crepant resolution \(f\colon Y\to X\) such that the fiber \(f^{-1}(0)\) over the singularity of \(X\) is a rational curve whose components are numbered by the conjugacy classes of \(G\); moreover, the homology classes of these components freely generate \(H_2(X,\mathbb Q)\) and \(H_i(X,\mathbb Q)=0\) for \(i>0\) and \(i\neq 2\). This is the so-called McKay correspondence [cf. \textit{J. McKay}, Finite groups, Santa Cruz Conf. 1979, Proc. Symp. Pure Math. 37, 183--186 (1980; Zbl 0451.05026)]. If \(\text{dim}(V)=3\) the first question was solved affirmatively by several people independently, while the second question was solved by \textit{Y. Ito} and \textit{M. Reid} who proved that the same result holds true as in dimension \(2\) [in: Higher dimensional complex varieties. Proceedings of the International Conference, Trento, Italy, June 15--24, 1994, 221--240 (1996; Zbl 0894.14024)]. In the paper under review the author imposes an additional assumption of the pair \((V,G)\), namely he assumes that \(V\) has a nondegenerate symplectic form and the inclusions \(G\subset \text{Sp}(V)\subset \text{SL}(V)\) preserve not only the volume form in \(V\) but also the symplectic form. Under these hypotheses he proves a higher-dimensional analogue of the McKay correspondence.