A proof of the Faber intersection number conjecture (Q847622): Difference between revisions

Faber's intersection number conjecture concerns tautological integrals over the moduli space of genus g curves. One possible formulation gives the values \[ \int_{\overline M_{g,n}} \psi_1^{d_1} \dots \psi_n^{d_n} \lambda_g\lambda_{g-1}=\frac{(2g-3+n)!|B_{2g}|}{2^{2g-1}(2g)!\prod_{j=1}^{n} (2d_j-1)!!}, \] where \(B_{2g}\) is the Bernouli number. The classes \(\psi_i\) are defined as the Chern classes of the line bundle \(L_i\) over \(\overline M_{g,n}\) whose fiber over a curve \((C, x_1, \dots, x_n) \) equals the cotangent line \(T_{x_i}^{\star}C\), while the \(\lambda_i\)'s are the Chern classes of the Hodge bundle, fiberwise described as \(H^0(C, \omega_C)\). The first proof of the conjecture came in two steps. It was shown by \textit{E. Getzler} and \textit{R. Pandharipande} that the statement follows from the Virasoro conjecture for \(\mathbb P^2\); in fact only the degree \(0\) part is necessary [J. Nucl. Phys., B 530, No. 3, 701--714 (1998; Zbl 0957.14038)]. The Virasoro constraints for \(\mathbb P^2\) were later established by Givental. The current paper gives a more direct proof of the conjecture. Mumford's relations are used to remove the lambda insertions, thus reducing the problem to identities between integrals involving only \(\psi\)-classes; these are the subject of the Kontsevich-Witten theorem. The proof then relies on skilfull combinatorial manipulations. Finally, another proof of the conjecture was recently given by \textit{A. Buryak} and \textit{S. Shadrin} [``A new proof of Faber's intersection number conjecture'', \url{arXiv:0912.5115}]. See also the approach of \textit{I. P. Goulden, D. M. Jackson} and \textit{R. Vakil} [``The moduli space of curves, double Hurwitz numbers, and Faber's intersection number conjecture'', \url{arXiv:math/0611659}] which establishes the formula in special cases.