The Conley conjecture for Hamiltonian systems on the cotangent bundle and its analogue for Lagrangian systems (Q1019690)

Dynamics of Hamiltonian systems remains a major inspiration in symplectic geometry and topology. The celebrated Arnold conjectures on fixed points of Hamiltonian diffeomorphisms led Floer to the discovery of a new homology theory that now bears his name. The Conley conjecture concerns the periodic points of such diffeomorphisms, namely it claims (originally for tori) that there are infinitely many of them. Just recently various authors proved the Conley conjecture not only for tori, but also for surfaces of positive genus and symplectically aspherical manifolds. Hamiltonian diffeomorphisms typically arise as time-\(1\) maps of Hamiltonian systems and can be classified in terms of the corresponding Hamiltonians. In the paper under review the conjecture is proved on cotangent bundles over \(C^3\)-smooth manifolds \(M\) for \(C^2\)-smooth strongly convex Hamiltonians \(H\) with quadratic growth along the fibers. If also \(H(-t,q,-p)=H(t,q,p)\), then the time-\(1\) map of the system has infinitely many periodic points already in the zero section . Alternatively, if \(H\) is \(C^4\) and \(M\) is \(C^5\) with dim \(M>1\) then time-\(\tau\) maps have infinitely many periodic points for every \(\tau>0\). Proofs are based on a variational setup applied to the Lagrangian system for the Fenchel transform of \(H\). Spaces of loops in \(M\) are endowed with a Hilbert manifold structure (hence the smoothness assumptions), and infinite-dimensional Morse theory is applied to the Lagrangian. .