中文摘要：

本文研究了Cohen—Grossberg神经网络模型的指数稳定性．为避免构造Lyapunov函数的困难，我们采用广义相对Dalquist数方法来分析神经网络的稳定性．借助这一方法，我们不但得到了Cohen-Grossberg神经网络模型平衡解的存在性、唯一性和全局指数稳定性的新的充分条件，而且给出了神经网络的指数衰减估汁．所获结论改进了已有文献的相关结果．

英文摘要：

In this paper, the exponential stahility of Cohen-Grossherg neural network model is considered, To avoid the difficulty of constructing a proper Lyapunov function, the generalized relative Dahlquist constant approach is employed to analyze the stahility of neural networks, and sufficient conditions for the existence of a unique equilihrium and the glohal exponential stahility of Cohen- Grossherg neural networks are presented. Moreover,the exponential convergence rate of the neural networks to stahle equilihrium point is estimated. Our results improve the existing ones.