Controlling Chaos by PSO-Based Neural Network

In this paper, a method of neural network control system which is based on Particle Swarm Optimization (PSO) learning algorithm is proposed to control chaotic dynamical systems. Under the condition of chaotic model unknown, the control system may stabilize a chaotic orbit into an unstable fixed point based on the technique of small perturbations. Structurally, the control system is composed of two integrated feed-forward neural networks: a predict neural network which help to adjust a control neural networks. Both the predict network and the control network use the PSO to adapt itself according to the fitness defined by predict and previous control. The proposed method is an adaptive search for the optimum control technique. Simulations show that, by defining proper fitness function, the control system is suitable for chaos stabilization.