A KAM theorem for hyperbolic-type degenerate lower dimensional tori in Hamiltonian systems (Q1270164)

KAM theory always is a very technical subject, which often makes it hard to see exactly to what extent new contributions differ from previous ones. The author of the present paper does a very good job in describing in the first section where his work fits into the existing body of literature. Considering a Hamiltonian of the form \[ H = \langle\omega,y\rangle + \tfrac 12\langle A(\omega)z,z\rangle +P,\quad (x,y,z) \in T^n\times{\mathbb R}^n\times{\mathbb R}^{2m}, \] where \(P\) is the perturbation term and the frequencies \(\omega\) belong to an open subset of \({\mathbb R}^n\) with positive Lesbesgue measure, most by now ``classical'' results deal with the hyperbolic case for the unperturbed tori, or the elliptic case where all eigenvalues of \(A(\omega)\) are purely imaginary. The present contribution is about the case where \(A\) has zero eigenvalues, taking for simplicity \(m=1\), and assuming the usual Diophantine condition for \(\omega\). Taking the quadratic terms more specifically to be of the form \({1\over 2} v^2+f(u,v)\) (with \(z = (u,v)\)), normal form theory dictates that \(f\) can be taken to be of the form \(f = u^3G(u)\) for some polynomial \(G\). The case of interest here is when \(G(u) = cu^{2d-3}\) (\(d\geq 2\)). In fact, the main subject of the paper is the case that \(c\) is negative (\(-1\) say), referred to as a hyperbolic type degenerate torus. There also are some indications about the nature of the problems when \(c\) is positive. The main theorem describes that when \(\omega_0\) satisfies the Diophantine condition, there exists a \(\omega^\star\) such that the perturbed Hamiltonian at \(\omega^\star\) possesses an n-dimensional invariant torus carrying rotational flow (frequency \(\omega_0\)), provided that \(P\) is analytic and sufficiently small. The proof makes use of the process of KAM iteration. An outline of the proof is presented in Section 2. One step of the KAM iteration is treated in full detail in Section 3. Convergence of the procedure and the completion of the proof are discussed in Section 4.