Open orbifold Gromov-Witten invariants of \({[\mathbb{C}^3/\mathbb{Z}_n]}\): Localization and mirror symmetry (Q657840)

Compared to the Gromov-Witten invariants of closed holomorphic curves invariants of their open counterparts are poorly understood and a satisfactory general definition of them is still lacking. These invariants are supposed to count holomorphic curves in a variety \(X\) with the boundaries on a Lagrangian submanifold \(L\subset X\). In addition to geometry their values depend on ``framing'' of \(L\), which remains mysterious in general. Most explicit computations rely on physical heuristics and involve mirror symmetry and/or large \(N\) duality. However, in some special cases open invariants can be defined ad hoc and computed mathematically. One of the computable cases, considered by \textit{S. Katz} and \textit{C.-C. M. Liu} [Adv. Theor. Math. Phys. 5, No. 1, 1--49 (2001; Zbl 1026.32028)], is when \(X\) is a Calabi-Yau threefold and \(L\) is a fixed locus of an antiholomorphic involution with a suitable torus action. Framing can then be identified with a choice of the action. Katz and Liu postulated a perfect obstruction theory for this case, analogous to the closed case, and computed the open invariants by virtual localization for the resolved conifold \(X=\mathcal{O}(-1)\oplus\mathcal{O}(-1)\) with a natural Lagrangian in it. The paper under review presents a similar computation for the Calabi-Yau orbifolds \(\mathbb{C}^3/\mathbb{Z}_n\) with suitable Lagrangians. The fixed loci of the torus action comprise a compact curve with twisted marked points and attached orbi-discs. It is the disc contributions, reflecting the \(\mathbb{Z}_n\) action on \(\mathbb{C}^3\), that are essentially new compared to the Katz-Liu case. All open invariants are reduced recursively to the degree \(0\) closed Gromov-Witten invariants with descendants, and a special function that the authors name the disc function. Although the approach works for any \(n\), explicit computations are performed for \(n=3,4\) with two different choices for the Lagrangian when \(n=4\), symmetric and asymmetric. Large part of the paper compares the results to mirror symmetry predictions based on \(B\)-model computations. The authors perform them within the paper, and one of the sections is an exposition of the \(B\)-model aimed at mathematicians. In particular, the mirror pair \((\widehat{X},\widehat{L})\) is constructed explicitly following Aganagic-Vafa, and Eynard-Orantin recursions for higher genus \(B\)-model invariants are reviewed in detail. For \(\mathbb{C}^3/\mathbb{Z}_3\) and the asymmetric Lagrangian in \(\mathbb{C}^3/\mathbb{Z}_4\) both the disc and the annulus potentials agree with the mirror computations. In the latter case this confirms that the annulus potential is a quasi-modular form of \(\Gamma(2)\subset \mathrm{SL}(2,\mathbb{Z})\). However, for the symmetric Lagrangian in \(\mathbb{C}^3/\mathbb{Z}_4\) there is a phase discrepancy. The authors speculate that \(\mathbb{Z}_2\)-isotropy on the fixed circle of the Lagrangian produces a non-trivial normalizing factor.