no. 13 · Wednesday 11 November 2026
Herman's converse KAM mechanism revisited
Abstract
In his counterexample to the KAM theorem, Herman used a hyperbolic perturbation together with a bump and asked whether the latter was necessary \cite[p.~78, Remark~4.7.3]{H1}. Motivated by this question, we compare destruction and persistence for a sine perturbation, with and without a Gevrey bump. The combined perturbation destroys all invariant circles with sufficiently small nonzero rotation numbers, whereas the sine perturbation alone preserves a sequence of analytic invariant circles in the same frequency range.
- fields
- Dynamical Systems