Summary
International Symposium on Nonlinear Theory and its Applications
2009
Session Number:B4L-A
Session:
Number:B4L-A3
Effect of Pattern Overlap in Delay Feedback Control Method for Chaotic Neural Network
Kouta Teramoto, Jousuke Kuroiwa, Hisakazu Ogura, Tomohiro Odaka, Haruhiko Shirai,
pp.-
Publication Date:2009/10/18
Online ISSN:2188-5079
DOI:10.34385/proc.43.B4L-A3
PDF download (453.7KB)
Summary:
In this paper, we investigate an effect of pattern overlap in a delay feedback control method for a chaotic neural network (CNN) model with chaotic associative memories. In the memory search of recalling process, it is important to investigate the control method which could change the chaotic wandering state to a certain periodic state gradually and softly. As one of control methods, we have proposed a delay feedback control method for a chaotic neural network model. From computer experiments, we succeeded to stabilize chaotic orbits which wander among memory patterns by applying the delay feed-back control. Furthermore, it is shown that a stabilized state depends on the overlap between itself and a state of the CNN model at the moment when the control signal was injected.