具有时变时滞的Hopfield神经网络的全局指数渐近稳定性

摘要

By using the usual Hopfield neural network models, some delay-independent stability criterion for neural dynamics with time-varying delays are derived. This paper extends previously known resurlts obtained by the other authors to the time-varying case. Both the differentiability and boundness conditions for the input-output functions are removed. Our suffcient conditions here not only guarantee the existence of the equilibrium for the delayed neural network, but also assure its global exponential asymptotic stability.%对具有时变时滞的Hopfield神经网络模型,给出了时滞无关的全局指数稳定判据.去掉了输入-输出函数的可微性和有界性条件,推广了其他作者基于常数时滞的有关结果,所给的充分条件不但能保证该时滞神经网络平衡点的存在性,而且能使其全局指数渐近稳定.