Holomorphic anomaly equations and the Igusa cusp form conjecture (Q1656355)

Let \(S\) be a non-singular projective \(K3\) surface and let \(E\) be A further conjecture is formulated on the quasimodularity and holomorphic anomaly equations for the cycle-valued Gromov-Witten theory of every elliptic fibration with section. The conjecture generalizes the holomorphic anomaly equations for elliptic Calabi-Yau threefolds predicted by Bershadsky, Ceocotti, Oguri, and Vafa [\textit{M. Bershadsky} et al., Commun. Math. Phys. 165, No. 2, 311--427 (1994; Zbl 0815.53082)]. It is shown that a modified conjecture holds numerically for the reduced Gromov-Witten theory of \(K3\) surfaces in primitive classes. an elliptic curve. Let \(X=S\times E\) be the Calabi-Yau threefold. In the 2016 article [\textit{G. Oberdieck} and \textit{R. Pandharipande}, Prog. Math. 315, 245--278 (2016; Zbl 1349.14176)], the authors conjectured the reduced Gromov-Witten theory of \(X\), presenting, in the primitive case, that the topological string partition function of \(X\) is the Igusa cusp forms. This conjecture is called the Igusa csup form conjecture. This article solves the reduced Gromov--Witten theory of \(X\) and establishes the Igusa cusp form conjecture in the primitive case. The main tool used in their proof is the correspondence between Gromov-Witten theory (counting stable maps) and Pandharipande-Thomas theory (counting sheaves). Both sides yield modular constraints and taken together, they determine the partition function form single coefficient. The sheaf theory side gives the elliptic transformation law of Jacobi forms. On the Gromov-Witten side, the authors study the theory for elliptic curves, and \(K3\) surfaces separately, and the following new results are obtained: (1) The generating series of Gromov-Witten classes of an elliptic curve are cycle-values quasimodular forms and satisfies a holomorphic anomaly equation; the quasimodularity generalizes a result of \textit{A. Okounkov} and \textit{R. Pandharipande} [Ann. Math. (2) 163, No. 2, 517--560 (2006; Zbl 1105.14076)], and the holomorphic anomaly equation proves a conjecture of Milanov-Ruan-Shen [\textit{T. Milanov} et al., J. Reine Angew. Math. 735, 287--315 (2018; Zbl 1423.14321)]. Then a conjecture about the quasimodularity and holomorphic anomaly equations is formulated for the cycle-valued Gromov-Witten theory of every elliptic fibration with section. (2) A holomorphic anomaly equation for the numerical reduced Gromov-Witten theory of the \(K3\) surface \(S\) in primitive classes. The Igusa cusp form conjecture, which is formulated as follows, is established in this article: Theorem. Let \(Z(u, q,\tilde{q})\) be the partition function. Then \(Z(u,q,\tilde{q})\) is the Laurent expansion of \(-1/\xi_{10}\) under the variable change \(q=e^{iu}\), \[ Z(u,q,\tilde{q})=-\frac{1}{\xi_{10}(p,q,\tilde{q})}. \] In particular, the theorem determines the Gromov-Witten invariants of the Calabi-Yau threefold \(X=S\times E\) in the primitive case. A further conjecture is formulated on the quasimodularity and holomorphic anomaly equations for the cycle-valued Gromov-Witten theory of every elliptic fibration with section. The conjecture generalizes the holomorphic anomaly equations for elliptic Calabi-Yau threefolds predicted by Bershadsky, Ceocotti, Oguri, and Vafa. It is shown that a modified conjecture holds numerically for the reduced Gromov-Witten theory of \(K3\) surfaces in primitive classes.