引用本文:余昭旭, 吴惕华.一类时滞神经网络系统的指数稳定性[J].控制理论与应用,2005,22(2):321~324.[点击复制]
YU Zhao-xu, WU Ti-hua.On the exponential stability of neural networks systems with time delays[J].Control Theory and Technology,2005,22(2):321~324.[点击复制]
一类时滞神经网络系统的指数稳定性
On the exponential stability of neural networks systems with time delays
摘要点击 1359  全文点击 1034  投稿时间：2003-07-04  修订日期：2004-03-09
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/j.issn.1000-8152.2005.2.030
  2005,22(2):321-324
中文关键词  神经网络  时滞  指数稳定性  范数  矩阵测度
英文关键词  neural networks  time delays  exponential stability  norm  matrix measur
基金项目  河北省自然科学基金资助项目(602623).
作者单位
余昭旭, 吴惕华 华东理工大学 自动化系,上海 200237
上海交通大学 自动化系,上海 200030
河北省科学院自动化研究所,河北 石家庄 050081 
中文摘要
      利用矩阵测度研究了一类时滞神经网络系统的指数稳定性,给出保证神经网络系统指数稳定的充分条件.输出函数不需要满足Lipschitz条件,且也不要求它们可微或严格单调递增.在关联矩阵不对称的情况下,所得到的结论仍然成立.最后一个数值例子验证了判据的有效性.
英文摘要
      The exponential stability of Hopfield-type neural networks with time delays is analyzed by using the method of matrix measure,and sufficient conditions are obtained for general exponential stabilities.The system admits a unique equilibrium in which the output functions do not satisfy the Lipschitz conditions and neither requires them to be differential or strictly monotonously increasing.All the results still hold without assuming any symmetry of the connection matrix.Finally a numeric example is presented to verify the validity of these criteria.