时间离散的回归神经网络的定性分析

Qualitative Analysis of Discrete time Recurrent Neural Networks

研究了一类时间离散的回归神经网络的定性问题.首先应用非线性分析中的度理论,建立了这类神经网络平衡点存在唯一的充要条件;其次通过Lyapunov函数,得到了这类神经网络全局指数稳定的充分准则.

In this paper, the qualitative problem of a class of discrete time recurrent neural networks is discussed. First, a sufficient and necessary condition for existence and uniqueness of the equilibrium point of these neural networks is given, where the degree theory of nonlinear analysis is used; second, some sufficient rules for the global exponential stability of these neural networks are obtained by using Lyapunov function.