A KAM theorem for finitely differentiable Hamiltonian systems

General theory of finite-dimensional Hamiltonian and Lagrangian systems, Hamiltonian and Lagrangian structures, symmetries, invariants (37J06) Stability problems for finite-dimensional Hamiltonian and Lagrangian systems (37J25) Perturbations of finite-dimensional Hamiltonian systems, normal forms, small divisors, KAM theory, Arnol'd diffusion (37J40) Nearly integrable Hamiltonian systems, KAM theory (70H08)

Abstract: Given

l > 2 u > 2 d g e q 4

, we prove the persistence of a Cantor--family of KAM tori of measure

O (v a r e p s i l o n^{1 / 2 - u / l})

for any non--degenerate nearly integrable Hamiltonian system of class

C^{l} (m a t h s c r D i m e s m a t h b b T^{d})

, where

m a t h s c r D s u b s e t m a t h b b R^{d}

is a bounded domain, provided that the size

v a r e p s i l o n

of the perturbation is sufficiently small. This extends a result by D. Salamon in cite{salamon2004kolmogorov} according to which we do have the persistence of a single KAM torus in the same framework. Moreover, it is well--known that, for the persistence of a single torus, the regularity assumption can not be improved.

Review text

The following KAM theorem is proved for a Hamiltonian of the form \(H(y,x)=K(y)+P(y,x)\), with \(K,P\in C^l(D\times\mathbb{T}^d)\) and \(D\) being a bounded domain in \(\mathbb{R}^d\). Denote by \(\varepsilon\) the \(C^l\)-norm of the perturbation \(P\). If \(K\) is nondegenerate and \(l>2\nu>2d\geq 4\) then all the KAM tori of the integrable system \(K\) with frequency \((\alpha,\tau)\)-Diophantine, with \(\alpha\sim\varepsilon^{1/2-\nu/l}\) and \(\tau=\nu-1\), do survive being slightly deformed. Moreover, the corresponding family of KAM tori fills the phase space up to a set of Lebesgue measure of \(O(\varepsilon^{1/2-\nu/l})\). \textit{J. Pöschel} [Commun. Pure Appl. Math. 35, 653--696 (1982; Zbl 0542.58015)] proved that a Cantor family of positive measure of KAM tori survives from a small perturbation \(P\) of class \(C^k\), with \(k>3d-1\), of an analytic integrable Hamiltonian. \textit{D. A. Salamon} [Math. Phys. Electron. J. 10, Paper No. 3, 37 p. (2004; Zbl 1136.37348)] showed that for the persistence of a single torus it is enough that both the integrable and perturbed part are of class \(C^l\) with \(l>2d\). Moreover, \textit{A. Bounemoura} [J. Éc. Polytech., Math. 7, 1113--1132 (2020; Zbl 1473.37074)] proved that if \(K\in C^{l+2}\) and \(P\in C^l\) with \(l>2d\), then a family of KAM tori persists filling the phase space up to a set of measure \(O(\sqrt{\varepsilon})\). The present work is a generalization of all these results. The proof relies on a KAM scheme à la Arnold. First a quantitative KAM step is proved for the real-analytic case. Then one approximates the \(C^l\) functions \(K,P\) by a sequence of real-analytic functions getting a sequence of real analytic Hamiltonian to which the previous KAM step is applied. This approximation is quantitative and performed in a refined manner, not just truncating the Fourier expansion.

Recommendations

Cites work

Cited in

(17)

This page was built for publication: A KAM theorem for finitely differentiable Hamiltonian systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2184699)