Equivariant GW theory of stacky curves (Q2447597): Difference between revisions

Let \(\mathcal{C}_{r,s}\) be an orbifold with the coarse moduli space \(\mathbb{P}^1\), and orbifold structures \(\mathbb{Z}_r\) and \(\mathbb{Z}_s\) over 0 and \(\infty\). Let \(\mathcal{X}\) be a \(K\)-gerbe over \(\mathcal{C}_{r,s}\) where \(K\) is a finite abelian group, and let \(R\), \(S\) be the isotropy groups of \(\mathcal{X}\) at 0 and \(\infty\). \(\mathcal{X}\) is a toric DM stack. The fixed point set of \(\mathbb{C}^*\)-action on the inertia stack of \(\mathcal{X}\) consists of \(|R|+|S|\) points labelled by the elements of \(R\) and \(S\), and it gives rise to a basis for the equivariant Chen-Ruan cohomology of \(\mathcal{X}\). The main object of study in the paper under review is the moduli space \(\overline{\mathcal{M}}=\overline{\mathcal{M}}_{g,r,s}(\mathcal{X},d)\) where \(r\) and \(s\) are tuples of elements of \(R\) and \(S\), respectively. This is the moduli space of stable degree \(d\) maps from genus \(g\) nodal orbifold curves to \(\mathcal{X}\). The fixed point loci of the induced \(\mathbb{C}^*\)-action on \(\overline{\mathcal{M}}\) are products of the moduli spaces \(\overline{\mathcal{M}}_{g,n}(BR)\) and \(\overline{\mathcal{M}}_{g,m}(BS)\) which are the moduli spaces of stable maps from genus \(g\) nodal orbifold curves with marked points to the classifying stacks. The component of \(\overline{\mathcal{M}}_{g,n}(BR)\) having monodromy \(r_1,\dots, r_n\) at the marked points is denoted by \(\overline{\mathcal{M}}_{g,r}\), where \(r=(r_1,\dots, r_n)\). Given \(\overline{\mathcal{M}}\) the corresponding descendent Gromov-Witten invariants are denoted by \[ \left\langle \prod_{i=1}^n\tau_{a_i}(0_{r_i})\prod_{j=1}^m\tau_{b_j}(\infty_{s_j})\right\rangle^{\mathcal{X}}_{g,d}. \] They are assembled into the generating \(\tau\)-function \[ \tau_{\mathcal{X}}(x,x^*,u)=\sum_{g\in \mathbb{Z}}\sum_{d\geq 0}u^{2g-2}q^d\left \langle \exp\left(\sum_{i,\rho}x_i(\rho)\tau_i(0_\rho)+ \sum_{j,\sigma}x^*_j(\sigma)\tau_j(\infty_\sigma)\right)\right \rangle ^{\mathcal{X}}_{g,d} \] where \(i,j\) run over nonnegative integers and \(\rho, \sigma\) run over elements of \(R, S\) respectively. The generating function of the Hurwitz-Hodge integrals over \(\overline{\mathcal{M}}_{g,r}\) is denoted by \(H^\bullet_r(z_1,\dots,z_n,u)\) where the bullet indicates that disconnected domain curves are allowed. The main results of the paper under review establish an explicit formulas for \(\tau_{\mathcal{X}}(x,x^*,u)\) and \(H^\bullet_r(z_1,\dots,z_n,u)\) expressed as vacuum expectations on the Fock space associated to the wreath product of \(S_n\) and \(K\). The proof are based on localization, orbifold ELSV formula that expresses \(H^\bullet_r\) in terms of Hurwitz numbers, and the interpretations in terms of the Fock spaces. This part of the paper under review can be viewed as an extension of the Okounkov-Pandharipande work on the equivariant Gromov-Witten theory of \(\mathbb{P}^1\). The first application of the main results proves the decomposition conjecture by Hellerman-Henriques-Pantev-Sharpe. More explicitly, in this case the conjecture takes the simple form (after rescaling variables) \[ \tau_\mathcal{X}=\prod_{\gamma\in K^*}\tau_{\mathcal{C}_{r,s}}. \] The second application proves that \(\tau_{\mathcal{C}_{r,s}}\) is a \(\tau\)-function for the 2-Toda hierarchy.