Presents a method of adaptive learning to control chaos. It is a composite of artificial neural networks and the approach of Ott, Grebogi and Yorke (OGY) (1990) to control unstable periodic orbits in deterministic chaotic systems. The authors implement an OGY type control using a simple linear feedforward network or perceptron and present a method of learning to continuously update the control strategy. To realize supervised learning, the authors least square fit the weights of the perceptron according to the behavior of the system and its response to the control signals. The logistic map and a three dimensional model of an electrochemical system are used as examples.
展开▼