Generic dynamics of 4-dimensional \(C^{2}\) Hamiltonian systems (Q2517924): Difference between revisions

The authors consider compact symplectic manifolds \((M,\omega)\) and Hamiltonian functions on \(M\) of class \(C^s\) where \(s\geq z\). These Hamiltonians are assumed to be constant on each connected component of \(\partial M\). The authors' main theorem is the following. If \((M,\omega)\) is a four-dimensional compact symplectic manifold, then for a general Hamiltonian \(H\in C^2(M,bbfR)\), the union of the regular energy surfaces \({\mathcal E}\) of \(H\) that are either Anosov or have zero Lyapunov exponents almost everywhere (with respect to the measure \(\mu_{\mathcal E}\) on the energy surface) for the Hamiltonian flow forms an open \(\mu-\bmod 0\) and dense subset of \(M\). The authors also prove the existence of a \(C^2\)-dense subset \({\mathcal D}\) of \(C^2(M,\mathbb{R})\) such that, for \(H\in{\mathcal D}\) there is an invariant decomposition \(M=D\cup Z\pmod 0\), where \(D\) is the union of sets with \(m_n\)-dominated splitting of the transverse linear Poincaré flow of \(H\), and \(H\) has zero Lyapunov exponents for \(x\in Z\). Key to the proofs of both theorems is a technical result on perturbations of the Hamiltonian.