Controlling chaos in map models (Q1976700)

This paper deals with a two-dimensional conservative map (Jacobian = +1) on a torus. The first variable \(J\) is a periodic function of the second one (an angle \(q)\). The author indicates that the purpose is a study of regular and chaotic behaviors. The Jacobian being equal to one, the terminology used ``stable (resp. unstable) fixed point'' must be understood as ``center'' (resp. ``saddle'') in the sense of the classical qualitative theory of Nonlinear Dynamics, or as ``elliptic (resp. hyperbolic) fixed point'' in the Birkhoff sense. The ``stability index'' is the sum of the fixed point multipliers. A first part is devoted to the evolution of the chaotic regions of the torus when a parameter \(K\) varies. From numerical simulations the text concludes that an increase of this parameter transforms an ``instability ring'' (terminology not used in the paper) in the Birkhoff sense (cf. fig.1) into what is called a ``chaotic sea'', i.e. the absence of ``stable elliptic'' fixed points. The second part introduces, in the angular variable relation, a perturbation in the form of a second parameter coefficient of a periodic function of the \(J\) image, in order to transform two saddle period one fixed points into centers. The study is limited to numerical experiments related to simulations of the phase space structure, and evaluation of Lyapunov characteristic exponents. Figure 2 corresponds to an heteroclinic connection (called ``junction of two chaotic layers'' here) between the unstable fixed points and the period two ones. Contrarily to what is claimed by the author it is not sure that the whole phase space is chaotic for \(K=2.9\). This would suppose that all the periodic points, whatever be their period, are unstable, which cannot be proved from an argument based on Lyapunov exponent.