Counting BPS States via Holomorphic Anomaly Equations
Shinobu Hosono
TL;DR
This work develops a holomorphic anomaly framework for a rational elliptic surface $rac{1}{2}$K3 to compute genus g Gromov-Witten invariants via an affine $E_8$ Weyl symmetry. It constructs a generating function $Z_{g;n}$ organized into Weyl-orbits, derives orbit-based decompositions, and uses vanishing/BPS constraints to fix holomorphic ambiguities, yielding explicit BPS numbers $n_g(eta)$ that are integral and compatible with the Gopakumar–Vafa picture. A key contribution is the explicit link between the local BCOV holomorphic anomaly equation and the HST equation in the local limit, including detailed genus-one and genus-two data for the half-K3 and for $old P^1 imesold P^1$ models. The results provide both new enumerative predictions for rational elliptic surfaces and a clearer bridge between holomorphic anomaly techniques in local mirror symmetry and the counting of BPS states.
Abstract
We study Gromov-Witten invariants of a rational elliptic surface using holomorphic anomaly equation in [HST1](hep-th/9901151). Formulating invariance under the affine $E_8$ Weyl group symmetry, we determine conjectured invariants, the number of BPS states, from Gromov-Witten invariants. We also connect our holomorphic anomaly equation to that found by Bershadsky,Cecotti,Ooguri and Vafa [BCOV1](hep-th/9302103).
