变时滞静态神经网络的不变集和周期吸引子  

Invariant set and periodic attractor of static neural network with time-varying delays

作　　者：李翠波[1] 王扬帆[2] 王林山[1] 

机构地区：[1]中国海洋大学数学系,山东青岛266071 [2]中国海洋大学信息工程学院,山东青岛266071

出　　处：《黑龙江大学自然科学学报》2006年第4期553-556,共4页Journal of Natural Science of Heilongjiang University

基　　金：国家自然科学基金资助项目(10171072)

摘　　要：利用M-矩阵的性质及矩阵不等式的分析技巧,研究变时滞静态神经网络的不变集、全局吸引集和全局周期吸引子.去掉了fj在R上的可微性条件以及有界性的限制,给出了较弱的判定变时滞静态神经网络的不变集、全局吸引集和全局周期吸引子的充分条件.特别的,对全局周期吸引子的存在范围进行估计.The invariant set, global attractive set and global attractor of static neural network with time-varying delays are studied by using properties of M-matrix and matrix inequality technique. Without the restrictions of boundedness and differentiability of fjon R, some sufficient criteria to distinguish the invariant set, global attractive set and global periodic attractor of static neural network with time-varying delays are derived. In Particular, an estimate on the existence range of global periodic attractor is provided.

关 键 词：变时滞 静态神经网络 不变集 吸引子 

分 类 号：O175[理学—数学]