Strengthening the cohomological crepant resolution conjecture for Hilbert-Chow morphisms (Q1952287): Difference between revisions

Ruan's cohomological crepant resolution conjecture asserts that for a smooth Gorenstein orbifold \(\mathcal{X}\) with the coarse moduli space \(X\), the Chen-Ruan cohomology ring of \(\mathcal{X}\) is isomorphic to the quantum corrected cohomology ring of a crepant resolution \(Y\) of \(X\). This conjecture has its origin in physics in which it is believed that the string theories of \(\mathcal{X}\) and \(Y\) are equivalent. A nice class of examples fitting into the assumptions of the conjecture is the case where \(S\) is a smooth complex surface, \(\mathcal{X}=[\text{Sym}^n (S)]\) (and \(X=\text{Sym}^n (S)\)), is its \(n\)-folded symmetric product, and \(Y=\text{Hilb}^n(S)\) is the Hilbert scheme of \(n\) points on \(S\). When \(S\) is projective with trivial canonical class the conjecture is proven independently by \textit{B. Fantechi} and \textit{L. Göttsche} [Duke Math. J. 117, No. 2, 197--227 (2003; Zbl 1086.14046)], \textit{B. Uribe} [Commun. Anal. Geom. 13, No. 1, 113--128 (2005; Zbl 1087.32012)], and \textit{Zh. Qin} and \textit{W. Wang} [in: Orbifolds in mathematics and physics. Proceedings of a conference on mathematical aspects of orbifold string theory, Madison, WI, USA, May 4--8, 2001. Contemp. Math. 310, 233--257 (2002; Zbl 1045.14001)]. \textit{J. Li} and \textit{W.-P. Li} [Math. Ann. 349, No. 4, 839--869 (2011; Zbl 1221.14006)] have partially proven the conjecture for simply connected projective surfaces. The main result of the paper under review is establishing Ruan's conjecture for this class of examples when \(S\) is an arbitrary toric surface. In fact more is proven in this paper: an isometric isomorphism \(L\) is constructed which maps the equivariant Chen-Ruan cohomology of \([\text{Sym}^n (S)]\) onto the equivariant cohomology of \(\text{Hilb}^n(S)\) satisfying \[ \langle \alpha_1,\alpha_2,\alpha_3 \rangle^{[\text{Sym}^n (S)]}(u)=\langle L( \alpha_1),L(\alpha_2),L(\alpha_3) \rangle^{\text{Hilb}^n(S)}(-e^{iu}), \] for any equivariant Chen-Ruan cohomology classes \(\alpha_1,\alpha_2,\alpha_3\). Here \[ \langle -,-,- \rangle^{[\text{Sym}^n (S)]}(u) \in \mathbb{Q}(t_1,t_2)[[u]] \quad \text{and} \quad \langle -,-,- \rangle^{\text{Hilb}^n(S)}(q) \in \mathbb{Q}(t_1,t_2)[[q]] \] are the generating functions of respectively the 3-pointed extended Gromov-Witten invariants of \([\text{Sym}^n (S)]\) and the 3-pointed extermal Gromov-Witten invariants of \(\text{Hilb}^n(S)\). The isomorphism \(L\) is given explicitly. The proof is based on the localization technique and reducing to the case \(S=\mathbb{C}^2\) studied by \textit{J. Bryan} and \textit{T. Graber} [Proc. Symp. Pure Math. 80, Pt. 1, 23--42 (2009; Zbl 1198.14053)] and \textit{A. Okounkov} and \textit{R. Pandharipande} [Invent. Math. 179, No. 3, 523--557 (2010; Zbl 1198.14054)]. The equivariant Chen-Ruan cohomology ring of \(\text{Sym}^n (S)\) and the equivariant quantum corrected cohomology ring of \(\text{Hilb}^n(S)\) are respectively recovered by specializing \(u=0\) and \(q=-1\). Another interesting feature is that the cup product of \(\text{Hilb}^n(S)\) is recovered from the 3-pointed extended Gromov-Witten invariants of \([\text{Sym}^n (S)]\) by specializing \(q=0\).