作　　者： ; ;

机构地区： 中国人民解放军第一军医大学

出　　处： 《通信学报》 1995年第2期13-19,共7页

摘　　要： 混沌状态下神经网络行为的研究是神经网络理论中的一个新内容，由于它可能揭示脑活动的深层机制而广受重视，但对混沌网络信息处理能力进行大量和深入研究却遇到许多障碍，其问题之一是难以为网络系统选择适当的非线性参数。本文提出一种建立混沌神经网络模型的简单耦合方法，解决了选择参数的困难。计算机模拟结果表明，用该模型研究混沌状态下网络动态行为是有效的。 Chaotic dynamic research of neural networks is a new developing direction in neural network theory.It is widely accepted that chaotic dynamics can give a mathematic explanation of mechanism of brain pattern recognition such as olfactory system.A number of new concept on the mechanism have been put forward,but at the same time,many problems have appeared during the chaotic dynamic research.One of the problems is the selection of nonlinear parameter for chaotic neural networks.In this paper,a simple coupling methord of chaotic neural network modelling was advanced to solve the problem of nonlinear parameter selection.The computer simulation showed that the method is valid.and that the constructed chaotic neural network could realize the associative memory.

关 键 词： 神经网络 混沌 动态特性 联想记忆

领　　域： [自动化与计算机技术] [自动化与计算机技术]