Stability analysis on a class of nonlinear continuous neural networks
Zhongyu Chen, Zongben Xu · 2002
The global convergence of neural networks is known to be the basis of successful applications of neural networks in various computation and recognition tasks. However, almost all the previous studies on neural networks assumed that the interconnection matrix is symmetric. In this paper, we investigate the sufficient condition to guarantee a class of nonlinear continuous neural networks including the Hopfield model as a special case to be global convergent towards unique stable equilibrium point without the assumption of symmetric interconnection. And we also give the sufficient condition to ensure the global convergence of the networks with symmetric interconnection matrix.>