Global Exponential Stability of Recurrent Neural Networks with Time-Delays
By using generalized Halanay inequality, and M-matrix, the problem of global exponential stability is discussed for a class of recurrent neural networks with time-varying delays. Without assuming Lipschitz conditions on the activation functions, several new and simple sufficient conditions are obtained to ensure existence uniqueness and global exponential stability at the equilibrium point of the neural networks. It is shown that the estimates obtained by the generalized Halanay inequalities improve the estimates obtained by the Lyapunov methods. The neural network model considered in this paper includes the delayed Hopfield neural networks and delayed cellular neural networks as its special cases. In addition, these criteria can be easily checked in practice.
Xinnian Chen Yiping Luo
Computer Science and Technology Department Hunan Institute of Engineering Xiangtan, China 411104 Mathematics Physics Department Hunan Institute of Engineering Xiangtan, China 411104
国际会议
南宁
英文
2007-07-20（万方平台首次上网日期，不代表论文的发表时间）