摘要
利用细胞神经网络激励函数的特点,对连接权矩阵进行适当分块,结合线性矩阵不等式分析技巧,得到了指数稳定及周期解存在的新判据,得到的新判据具有更弱的保守性。仿真结果表明,新判据是有效的。
Based on the feature of activation function of cellular neural networks, by using linear matrix inequality and partitioning of matrix technique, obtained two novel criteria ensuring exponential stability and periodicity. Compared with previous results, the novel criteria of this paper is less conservative. Simulation example shows that the new criteria is valid.
出处
《计算机应用研究》
CSCD
北大核心
2009年第8期2879-2880,2911,共3页
Application Research of Computers
基金
国家教育部新世纪人才支持计划资助项目(NCET-06-0811)
贵州财经学院博士基金资助项目(200702)
关键词
细胞神经网络
指数稳定
周期解
分块矩阵
cellular neural networks
exponential stability
periodicity solution
partitioned matrices