Point-like bounding chains in open Gromov-Witten theory

TL;DR

This paper develops a new method to define genus zero open Gromov-Witten invariants for Lagrangian submanifolds of any dimension, overcoming bubbling issues and generalizing previous results.

Contribution

It introduces a bounding chains technique to cancel disk bubbling and defines invariants in arbitrary dimensions, extending prior work by Welschinger and others.

Findings

Invariants are defined for broader classes of constraints and Lagrangians.

The method generalizes known invariants in low dimensions.

A canonical family of point-like bounding chains is identified.

Abstract

We present a solution to the problem of defining genus zero open Gromov-Witten invariants with boundary constraints for a Lagrangian submanifold of arbitrary dimension. Previously, such invariants were known only in dimensions $2$ and $3$ from the work of Welschinger. Our approach does not require the Lagrangian to be fixed by an anti-symplectic involution, but can use such an involution, if present, to obtain stronger results. Also, non-trivial invariants are defined for broader classes of interior constraints and Lagrangian submanifolds than previously possible even in the presence of an anti-symplectic involution. The invariants of the present work specialize to invariants of Welschinger, Fukaya, and Georgieva in many instances. The main obstacle to defining open Gromov-Witten invariants with boundary constraints in arbitrary dimension is the bubbling of $J$-holomorphic disks. Unlike in low dimensions or for interior constraints, disk bubbles do not cancel in pairs by anti-symplectic involution symmetry. Rather, we use the technique of bounding chains introduced in Fukaya-Oh-Ohta-Ono's work on Lagrangian Floer theory to cancel disk bubbling. At the same time and independently, gauge equivalence classes of bounding chains play the role of boundary constraints, in place of the cohomology classes that usually serve as constraints in Gromov-Witten theory. A crucial step in our construction is to identify a canonical up to gauge equivalence family of "point-like" bounding chains, which specialize in dimensions $2$ and $3$ to the point constraints considered by Welschinger.