Cohen-Grossberg神经网络的全局指数稳定性分析(英文)  

Global Exponential Stability Analysis of Cohen-Grossberg Neural Networks

作　　者：万安华[1] 王绵森[2] 彭济根[2] 毛卫华[3] 

机构地区：[1]中国科学院自动化研究所,北京100080 [2]西安交通大学理学院,陕西西安710049 [3]华南农业大学应用数学系,广东广州510642

出　　处：《应用数学》2006年第2期381-387,共7页Mathematica Applicata

基　　金：SupportedbyNNSFofChina(10371097)andPrincipalFoundationsofSouthChinaAgriculturalUniversity(2004K055and2005K023)

摘　　要：本文研究了CohenGrossberg神经网络模型的指数稳定性.为避免构造Lyapunov函数的困难,我们采用广义相对Dalquist数方法来分析神经网络的稳定性.借助这一方法,我们不但得到了CohenGrossberg神经网络模型平衡解的存在性、唯一性和全局指数稳定性的新的充分条件,而且给出了神经网络的指数衰减估计.所获结论改进了已有文献的相关结果.In this paper, the exponential stahility of Cohen-Grossherg neural network model is considered, To avoid the difficulty of constructing a proper Lyapunov function, the generalized relative Dahlquist constant approach is employed to analyze the stahility of neural networks, and sufficient conditions for the existence of a unique equilihrium and the glohal exponential stahility of Cohen- Grossherg neural networks are presented. Moreover,the exponential convergence rate of the neural networks to stahle equilihrium point is estimated. Our results improve the existing ones.

关 键 词：指数稳定性Cohen—Grossberg神经网络 广义相对Dalquist数 指数衰减估计 

分 类 号：O175.13[理学—数学] TP183[理学—基础数学]