Adaptive intelligence learning for nonlinear chaotic systems

In this paper, a reinforcement learning algorithm is proposed for a class of nonlinear differential chaotic systems. The nonlinear function of the chaotic systems is assumed to be bounded but the bounds are unknown. The unknown bounds need to be on-line adjusted. An adaptive optimal (or near optimal) control input with the reinforcement signal can be obtained compared with the current adaptive control for chaotic systems. The reinforcement signal is approximated by the neural networks. Based on Lyapunov analysis theory and by using Young’s inequalities, the closed-loop system is guaranteed to be stable. Finally, the simulation results are given to illustrate the effectiveness of the approach.