Chaos control of Bonhoeffer-van der Pol oscillator using neural networks

This paper addresses the control of chaos using a neural network for a cont inuous time dynamical system, The neural network is trained on both the Ott -Grebogi-Yorke (OGY) control algorithm and the Pyragas's delayed feedback c ontrol algorithm. The system considered for this study is a Bonhoeffer-van der Pol (BVP) oscillator. A feed-forward backpropagating neural network is used for the control application. It is Found that the control effected by the neural network trained on the OGY control algorithm results in smaller control transients than when the control is effected directly by the OGY al gorithm itself. The control transients are of the same order in the case of the Pyragas method. (C) 2001 Elsevier Science Ltd. All rights reserved.