The classical KAM theorem for Hamiltonian systems via rational approximations (Q2513966): Difference between revisions

The authors consider small perturbations of integrable Hamiltonian systems, which are defined by a Hamiltonian function of the form \[ H(p,q) = h(p) + \epsilon f(p,q), \quad (p,q)\in \mathbb{R}^n \times \mathbb{T}^n, 0 \leq \epsilon< 1, \] where \(n\geq 2\) is an integer and \(\mathbb{T}^n = \mathbb{R}^n/\mathbb{Z}^n\): the Hamiltonian system associated to this Hamiltonian function is then given by \[ \dot{p}=-\partial_q H(p,q)=-\epsilon\partial_q f(p,q), \] \[ \dot{q}=\partial_p H(p,q)=\nabla h(p)+\epsilon\partial_p f(p,q). \] When \(\epsilon=0\), the system associated to \(H = h\) is trivially integrable: all its solutions are invariant tori on which the dynamics is quasiperiodic with frequency \(\omega_0 = \nabla h(p)\in \mathbb{R}^n\). The authors give a new proof of the classical KAM theorem on the persistence of an invariant quasi-periodic torus, whose frequency vector satisfies the Bruno-Rüssmann condition, in real-analytic non-degenerate Hamiltonian systems close to integrable are given. The proof, which uses rational approximations instead of small divisors estimates, is an adaptation to the Hamiltonian setting of a method which was introduced in [the authors, Math. Z. 275, No. 3--4, 1135--1167 (2013; Zbl 1308.37028)] for perturbations of constant vector fields on the torus.