Finding stable attractors of a class of dissipative dynamical systems by numerical parameter switching

摘要

In this article we prove numerically, via computer graphic simulations and specific examples, that switching the control parameter of a dynamical system belonging to a class of dissipative continuous dynamical systems, one can obtain a stable attractor. In this purpose, while a fixed step-size numerical method approximates the solution of the mathematical model, the parameter control is switched every few integration steps, the switching scheme being time periodic. The switch occurs within a considered set of admissible parameter values. Moreover, we show via numerical experiments that the obtained synthesized attractor belongs to the class of all admissible attractors for the considered system and matches to the averaged attractor obtained with the control parameter replaced with the averaged switched parameter values. This switched strategy may force the system to evolve along on a stable attractor whatever the parameter values and introduces a convex structure inside of the attractor set via a bijection between the set of parameter control values and the attractors set. The algorithm besides its utility in systems stabilization, when some desired parameter control cannot be directly accessed, may serve as a model for the dynamics encountered in reality or in experiments, e.g. three species food chain models, electronic circuits, etc. This method, compared, for example, to the OGY algorithm where only small perturbations of parameter control can be issued, allows relatively large parameter perturbations. Also, it does not allow to stabilize an unstable orbit but, using an appropriate parameter switching algorithm, it allows to reach an already existing attractor. The present work extends the results we obtained previously and is applied to Lorenz, Roessler and Chen systems.