On the relationship between the Bruno function and the breakdown of invariant tori (Q1972885)

Let a Hamiltonian system depending on a parameter \(\varepsilon\) have an invariant torus carrying quasiperiodic orbits with frequency vector \(\omega\). Then, when varying the parameter, the torus can be destroyed. Till its destruction the rotation vector is preserved. Suppose one deals with two degrees of freedom system or symplectic maps, then the rotation vector is in fact the rotation number of the torus. Denote by \(\varepsilon_c(\omega)\) the critical value of the parameter at the destruction moment. The authors study numerically the relation between the so-called Bruno function \(B(\omega)\) of the number \(\omega\) defined on the set of irrational numbers and the function \(\varepsilon_c (\omega)\), more precisely, if \(\varepsilon_c (\omega)= C(\omega)\exp(-\eta B(\omega))\) is valid with a continuous \(C(\omega)\) and some \(\eta\) (hypothesis by Marmi and Stark). If so, it would allow one to give some relation between these two quantities, one of which is purely arithmetic. Simulations with the standard map showed that the hypothesis is not true (on the level of rather careful numerical work).