Title :
Exponential convergence estimates for analog neural networks with multiple delays
Author :
Chu, Tianguang ; Wang, Zhaolin ; Wang, Long
Author_Institution :
Center for Syst. & Control, Peking Univ., Beijing, China
Abstract :
A component-wise estimate of exponential convergence is obtained for a class of neural networks with multiple delays by using a method based on a comparison principle of delay differential systems. The method is simple and straightforward in analysis, without resorting to any Lyapunov functionals. The results explicitly show the effect of time delays on the exponential decay rate of the networks and is of practical significance for designing fast and stable neural networks. Some existing results for the Hopfield model and the delayed cellular neural network (DCNN) model via the Lyapunov functional method are found to be special cases of the present result
Keywords :
Hopfield neural nets; Lyapunov methods; analogue processing circuits; cellular neural nets; circuit stability; convergence; delay-differential systems; delays; estimation theory; network synthesis; neural chips; Hopfield model; Lyapunov functional method; analog neural networks; comparison principle; componentwise estimate; delay differential systems; delayed cellular neural network model; exponential convergence estimates; exponential decay rate; fast stable neural network design; multiple delays; Circuit stability; Control systems; Convergence; Delay effects; Delay estimation; Delay systems; Image converters; Lyapunov method; Neural networks; Neurons;
Conference_Titel :
Decision and Control, 2001. Proceedings of the 40th IEEE Conference on
Conference_Location :
Orlando, FL
Print_ISBN :
0-7803-7061-9
DOI :
10.1109/.2001.980682
