摘要
本文运用GCM混沌神经网络对Hopfield神经网络在求解优化方面的问题进行了改进。通过混沌遍历 ,可使Hopfield网络在整个相空间进行搜索 ,从而避免网络在运行过程中陷入局部极小值。通过对一个对弈的实例进行实验 ,结果显示Hopfield网络的寻优特性获得了较大改进。
A GCM chaotic neural network is introduced to improve on Hopfield neural network on optimization. Hopfield network can searching for entire phase space by chaotic travelling. So local minimum can be avoided. An example is experimented and the result showing that big improvement exit in optimization. [
出处
《微机发展》
2000年第6期5-8,共4页
Microcomputer Development
关键词
优化问题
神经网络
混沌
解
Hopfield Neural Network
Chaotic Neutral Network
Optimization
