含时延的双向联想记忆神经网络的指数吸引性分析  

Analysis of Exponential Attraction of Bidirectional Associate Memory Neural Networks with Delays

作　　者：周小平[1] 

机构地区：[1]四川师范大学数学与软件科学学院,四川成都610066

出　　处：《四川师范大学学报（自然科学版）》2005年第4期386-390,共5页Journal of Sichuan Normal University（Natural Science）

基　　金：四川省应用基础研究;四川省学术与技术带头人培养资金资助项目

摘　　要：针对含时延的双向联想记忆神经网络模型具有多个对应于不同记忆模式的平衡点的特点,采用不等式技巧和非负矩阵性质,建立了平衡点指数吸引域及其指数收敛速度估计的方法,所得结论对含时延的双向联想记忆神经网的设计和容错性能的评价都很有意义.In general, bidirectional associate memory neural networks with delay （BAMNND） have more equilibria corresponding different memory mode. In this paper, the asymptotic stability is studied and some estimation results about the exponential attraction domain and the exponential convergent rate for BAMNN with delays are established using the inequality techniques and the properties of nonnegative matrices. The obtained results have important significanee in the design and the evaluation of error-correction capability.

关 键 词：时延 双向联想记忆神经网络 指数吸引域 指数收敛速度 

分 类 号：O175.21[理学—数学]