Floer homology of open subsets and a relative version of Arnold's conjecture (Q5928277): Difference between revisions

One form of the Arnold conjecture states that, for a Hamiltonian diffeomorphism \(\psi\) on \(T^*M\), \[ \#(\psi(M) \cap M) \geq \text{crit}(M), \] where \(M\) is embedded as the zero section in \(T^*M\). The conjecture has been proved for lower bounds which are bounded above by \(\text{crit}(M)\); for instance, \(\text{cup}(M)+1\) (in the degenerate case). In this paper, the authors prove the Arnold conjecture in the non-degenerate case for open subsets \(U \subseteq M\), where \(M\) is a smooth manifold and \(\partial U\) is smooth. Namely, they show that \[ \#(\psi(M) \cap \nu^*\overline U) \geq \text{rank}(H_*(U:\mathbb Z)). \] Here \(\nu^*\overline U\) is the conormal of \(\overline U\) constructed from the zero section of \(U\) and the negative part of the conormal bundle of \(\partial U\). The authors also mention that they can prove the degenerate case with the cup-length (\(+1\)) of \(H_*(U:\mathbb Z)\) on the right. To prove their result, they create a suitable Floer homology. But, in order to avoid doing analysis on singular Lagrangian submanifolds, they cleverly define an approximation method through smooth Lagrangian submanifolds. In particular, the Floer homology of \(U\) is then a direct limit of the Floer homologies of the approximations.