一类具有时滞的投影神经网络系统的指数稳定性  

Exponential Stability of Projected Neural Networks System with the Delay

作　　者：许金快[1] 黄元清[2] 钟守铭[1] 

机构地区：[1]电子科技大学应用数学学院,成都610054 [2]四川理工学院计算机学院,四川自贡643000

出　　处：《四川理工学院学报（自然科学版）》2010年第1期14-18,共5页Journal of Sichuan University of Science & Engineering(Natural Science Edition)

基　　金：国家973项目资助(2010CB732501)

摘　　要：文章研究了一类具有时滞的投影神经网络的稳定性问题,根据这类神经网络的特点,将神经网络的状态变量进行分块,通过构造Lyapunov泛函,导出了神经网络系统指数稳定性的充分条件,在适当的初始条件下,给出的稳定性条件与分块矩阵的某些块无关,表明我们所获得的条件的优越性。Stability problem of projected neural networks system with delay is studied. By dividing the projected neural networks system state varied into subgroups according to the characters of the projected neural networks, sufficient conditions for exponential stability is derived from via constructing Lyapunov functional. Those conditions suitable are associated with some initial value and represented by some blocks of the interconnection matrix.

关 键 词：时滞 神经网络 LYAPUNOV泛函 指数稳定性 

分 类 号：O231[理学—运筹学与控制论]