一类具有变时滞的Hopfield型神经网络的全局指数稳定性  

Global Exponential Stability for Hopfield Neural Networks with Varying Delays

作　　者：孙建枝[1] 武怀勤[1] 单彩虹[1] 

机构地区：[1]燕山大学理学院,河北秦皇岛066004

出　　处：《计算技术与自动化》2007年第3期9-12,共4页Computing Technology and Automation

基　　金：国家自然科学基金资助项目(10571035)

摘　　要：通过构造适当的Lyapunov函数,利用Halanay不等式和Young不等式,讨论一类具有变时滞的Hopfield型神经网络的全局指数稳定性。在对网络施加两个不同的神经元激励函数的条件下,导出网络全局指数稳定的一个充分条件,得到的充分条件在实际应用中易于验证,且有较小的保守性,因而对网络的应用和设计具有重要意义。最后,一个数值实例进一步验证结果的正确性。In this paper, global exponential stability of Hopfield neural networks with time - varying delays is investigated by constructing suitable Lyapunov function and using Young inequality and Halanay inequality in connection with inequality analysis techniques , A new sufficient condition is obtained to guarantee global exponential stability of networks, The condition are easy to check in practice and impose less conservation. Thus, the obtained results possess importance significance in the application and design of networks, At last, an example is given to illustrate the application of our results.

关 键 词：神经网络 Halanay不等式 全局指数稳定性 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]