时滞 Hopfield 神经网络的稳定动力学行为分析

STABLE DYNAMICAL ANALYSES FOR NEURAL NETWORK WITH DELAYS

神经网络的硬件实现常伴随反馈信号的延时，网络的应用需要对具有时滞反馈网络稳定性的充分研究。用Ｒａｚｕｍｉｋｈｉｎ型李雅普诺夫泛函方法，对非线性时滞Ｈｏｐｆｉｅｌｄ神经网络的稳定动力学行为进行了分析，利用泛函滞后微分方程的稳定性定理，导得时滞反馈连续Ｈｏｐｆｉｅｌｄ神经网络一致稳定和一致渐近稳定的充分判据。

The physical implementation of neural network follows delays of feedback signal and the application of network requires the study of the stability.In the paper,by Liapnuov functional methods of Razumikhin type,the stability for nonlinear continuous Hopfield neural network with delays is studied,and several new important theorems guaranteeing the uniform stability and uniform asymptotic stability of the network are derived.