关于随机时滞神经网络稳定性的注记(英文)  

A Note on Stability of Stochastic Delay Neural Networks

作　　者：张子芳[1] 徐道义[2] 

机构地区：[1]淮海工学院数理系,江苏连云港222005 [2]四川大学数学研究所,成都610064

出　　处：《工程数学学报》2010年第4期720-730,共11页Chinese Journal of Engineering Mathematics

基　　金：The National Natural Science Foundation of China(10971240;10371083);the Natural Science Foundation of Huaihai Institute of Technology(KK06004)

摘　　要：利用非负鞅收敛定理、Lyapunov泛函方法和网络自身的特性讨论了变时滞随机递归神经网络的随机指数稳定性,给出了这类神经网络随机指数稳定性的新的代数准则,所给代数准则简单易用。两个应用实例说明即使针对随机Hopfield神经网络所给的代数准则也优于相关的判别准则。The almost exponential stability for a stochastic recurrent neural network with time-varying delays is discussed by means of a nonnegative semi-martingale convergence theorem, the Lyapunov functional method and the characteristics of stochastic delay recurrent neural networks. The new algebraic criteria of the almost exponential stability for the stochastic recurrent neural network with time-varying delays is derived. These algebraic criteria are simple and practical. Two examples show these new algebraic criteria are better than the relative criteria for the stochastic Hopfield neural network.

关 键 词：随机递归神经网络 变时滞 随机指数稳定性 样本Lyapunov指数 

分 类 号：O211.63[理学—概率论与数理统计]