摘要
本文利用某些技巧和Lyapunov方法,得到了Hopfield连续联想记忆模式的吸引域及其中每一点趋向记忆模式的指数收敛速度的一些全新的估计结果。这些结果可用于评价Hopfield连续反馈联想记忆网络的容错能力,且可用于综合连续反馈联想记忆网络。
Some new results about estimation of attraction domain of memory patterns and exponential convergence rate of the network trajectories to memory patterns for Hopfield continuous associative memory are obtained by using of some new technics and Lyapunov method. These results can be applied to evaluate the error-correction capability of Hopfield continuous feedback associative memory and to synthesze the procedures of Hopfield continuous associative memory neural networks.
基金
国家自然科学基金
云南省自然科学基金
关键词
LYAPUNOV方法
神经网络
吸引域
指数收敛速度
Lyapunov method, Neural network, Attraction domain, Exponential convergence rate, Eigenvalue
