Lagrangian Floer theory over integers: spherically positive symplectic manifolds (Q2637468): Difference between revisions

Lagrangian Floer theory is a very hot topic in symplectic topology. In particular, it leads to the notion of Fukaya category of a symplectic manifold, which plays an important role in Kontsevich's famous homological mirror symmetry conjecture. The main issue to develop the Lagrangian Floer theory is to construct some geometric structures on the moduli space of stable pseudo-holomorphic discs and then to capture the homological information of the moduli space. More precisely, we need to construct a virtual fundamental chain (or cycle). In general, the moduli space is not a smooth manifold nor even an orbifold. It admits a Kuranishi structure. Assume that \(X\) is a Kuranishi space, the local model of \(X\) can be described as follows. For any \(p\in X\) there exists a quintet \((V_{p},E_{p},\Gamma_{p},\psi_{p},s_{p})\) satisfying: (1) \(\Gamma_{p}\) is a finite group; (2) \(V_{p}\) is a smooth manifold with finite dimension on which \(\Gamma_{p}\) acts smoothly and effectively; (3) \(E_{p}\) is a finite-dimensional vector space over the real numbers and \(\Gamma_{p}\) has a linear representation on \(E_{p}\); (4) \(s_{p}:V_{p}\rightarrow E_{p}\) is a \(\Gamma_{p}\)-equivariant smooth map; (5) \(\psi_{p}\) is a homeomorphism from \(s^{-1}_{p}(0)/\Gamma_{p}\) to a neighborhood of \(p\) in \(X\). For simplicity, \(s_{p}\) can be considered as a section of the \(\Gamma_{p}\)-equivariant trivial vector bundle \(E_{p}\times V_{p}\). Intuitively, a Kuranishi space is an object constructed by gluing together the local model \(s^{-1}_{p}(0)/\Gamma_{p}\) compatibly. According to the result of differential topology we can make a slight perturbation of \(s_{p}\) such that the perturbed section, denoted by \(s^{\prime}_{p}\), is transversal to the zero section. This implies that the zero locus of \(s^{\prime}_{p}\) is a smooth submanifold in \(V_{p}\), however, it is not \(\Gamma_{p}\)-invariant since we can not guarantee the existence of a \(\Gamma_{p}\)-equivariant smooth perturbed section which is transversal to the zero section. In general, it is unrealistic to require a smooth section to be equivariant and transversal, and therefore, the multi-valued perturbation is necessary to obtain the transversality. This technique has been used by the authors to develop the Lagrangian Floer theory over \(\mathbb{Q}\) in their research monographs [Lagrangian intersection Floer theory. Anomaly and obstruction. I and II. Providence, RI: American Mathematical Society (AMS); Somerville, MA: International Press (2009; Zbl 1181.53002 and Zbl 1181.53003)]. The paper under review is a continuation of [loc. cit.]. The primary objects of the study of this paper are the following: (1) From Section 3 to Section 10 the authors construct a single-valued perturbation for a Kuranishi space by using some results of analysis and homological algebra [Theorem 3.1]. (2) For a compact spherically positive symplectic manifold \((M,\omega,J)\) with a given Lagrangian submanifold \(L\subset M\) and a homology class \(\beta\in H_{2}(M,L;\mathbb{Z})\), they study the moduli space of stable \((k+1)\)-marked pseudo-holomorphic discs with respect to \(L\) and \(\beta\). In particular, they give an estimation of the virtual dimension of the singular locus of the moduli space \(\mathcal{M}^{\mathrm{main}}_{k+1}(\beta;P_{1},\dots,P_{k})\) [Proposition 12.1]. From this estimation it follows that the single-valued perturbation of \(\mathcal{M}^{\mathrm{main}}_{k+1}(\beta;P_{1},\dots,P_{k})\) may give rise to a virtual fundamental class over \(\mathbb{Z}\) rather than \(\mathbb{Q}\). (3) For a given Lagrangian submanifold (respectively, a relatively spin Lagrangian) \(L\) the authors construct a filtered \(A_{\infty}\) algebra structure on \(H^{*}(L;\wedge^{\mathbb{Z}_{2}}_{0,\mathrm{nov}})\) (respectively \(H^{*}(L;\wedge^{\mathbb{Z}}_{0,\mathrm{nov}})\)) [Theorem 1.1]. (4) Considering the pair of Lagrangian submanifolds (respectively, the relatively spin pair) \((L_{0},L_{1})\) of \((M,\omega)\), the authors prove that there exists a unital filtered \(A_{\infty}\) bimodule structure on the Lagrangian Floer complex \(C(L_{1},L_{0};\wedge^{\mathbb{Z}_{2}}_{0,\mathrm{nov}})\) (respectively \(C(L_{1},L_{0};\wedge^{\mathbb{Z}}_{0,\mathrm{nov}})\)) [Theorem 1.2]. This is an interesting and meaningful work.