Floer-Novikov cohomology and the flux conjecture (Q866872)

Using the Floer-Novikov cohomology, \textit{H.-V. Lê} and \textit{K. Ono} [Topology 34, 155--176 (1995; Zbl 0822.58019)] proved the flux conjecture for general closed symplectic manifolds. For an symplectic isotopy \(\{\varphi_t\}\) of a closed symplectic manifold \((M, \omega)\), its flux is defined by \[ \int^1_0\bigl[ i(X_t)\omega\bigr]\,dt\in H^1(M;\mathbb{R}),\quad\frac{d}{dt}\varphi_t=X_t\circ \varphi_t. \] It gives a homomorphism \[ \widetilde{\text{Flux}}:\widetilde {\text{Symp}}_0 (M,\omega)\to H^1(M;\mathbb{R}), \] where \(\widetilde{\text{Symp}}_0 (M, \omega)\) is the universal covering group of the identity component subgroup of the group of symplectomorphisms of \((M,\omega)\). The flux group \(\Gamma_\omega\) is \(\widetilde{\text{Flux}}(\pi_1(\text{Symp}_0(M,\omega)))\). The flux conjecture states that \(\Gamma_\omega\) is discrete in \(H^1(M;\mathbb{R})\). Flux induces a homomorphism \[ \text{Flux\,:\,Symp}_0(M,\omega)\to H^1(m;\mathbb{R})/ \Gamma_\omega. \] Its kernel is the group of Hamiltonian diffeomorphisms Ham\((M,\omega)\) of \((M,\omega)\) [\textit{A. Banyaga}, Comment. Math. Helv. 53, 174--227 (1978; Zbl 0393.58007)]. So, the flux conjecture is equivalent to the assertion that Ham\((M, \omega)\) is closed in \(\text{Symp}_0(M,\omega)\) with respect to the \(C^1\)-topology (\S 2). To prove the flux conjecture, the construction of Novikov-Floer cohomology in [1] is extended to symplectic isotopies on general closed symplectic manifolds in \S 3, by using technologies such as Kuranishi structures and multi-valued perturbation [cf. \textit{K. Fukaya} and \textit{K. Ono}, Topology 38, No. 5, 933--1048 (1999; Zbl 0946.53047), hereafter referred to as [2]; \textit{G. Liu} and \textit{G. Tian}, Floer homology and Arnold conjecture, J. Differ. Geom. 49, No. 1, 1--74 (1998; Zbl 0917.58009)]. A review of Novikov-Floer cohomology adopting cohomological convention, i.e., describing the gradient flow lines as solutions of a perturbed Cauchy-Riemann equation, is also given in this section. The flux conjecture is proved to show the existence of a neighborhood \(U\) of the origin of \(H^1(M;\mathbb{R})\) such that \(\Gamma_\omega\cap U=\{0\}\) (Theorem 4.1). This is done estimating the rank of the Novikov-Floer cohomology by using the rank of the Novikov cohomology, when the flux is small. Necessary discussions on the orientation of moduli space are done in \S 5 (Theorem 5.5. cf. \S 21 of [2]). In \S 6, Floer-Novikov cohomology with coefficients in a local system is introduced. Then, comparing Floer cohomologies of Hamiltonian isotopies and non-Hamiltonian symplectic isotopies with small non-zero flux and observing their different behavior under twisting with the local system, an alternative proof of Theorem 4.1 is given. The author says that, in the context of mirror symmetry, considering Floer cohomology for pairs of Lagrangian submanifolds equipped with flat bundles, constructed in the same manner. The flux conjecture implies that the quotient space Symp\(_0(M,\omega)/\text{Ham}(M,\omega)\) is Hausdorff. As for a similar problem that asks whether the quotient of the space of Lagrangian submanifolds by the Hamiltonian deformation equivalence relation is Hausdorff, the author remarks the argument of this paper is not directly applied [cf. \textit{K. Fukaya, Y.-G. Oh, H. Ohta} and \textit{K. Ono} Lagrangean intersection Floer theory -- obstruction and anomaly, preprint (2000)].