Stability analysis of discrete hopfield neural networks with weight function matrix

Abstract

Most matrixes of Discrete Hopfield neural networks(DHNNs) and DHNNs with delay are constant matrixes. However, most weight matrixes of DHNNses are variable in many realistic systems. So, the weight matrix and the threshold vector with time factor are considered, and DHNNs with weight function matrix (DHNNWFM) is described. Moreover, the result that if weight function matrix and threshold function vector respectively converge to a constant matrix and a constant vector that the corresponding DHNNs is stable or the weight matrix function is a symmetric function matrix, then DHNNWFM is stable, is obtained by matrix analysis. © 2008 Springer Berlin Heidelberg.