摘要
神经网络和以遗传算法为代表的进化算法都是仿效生物处理模式来获得智能信息处理功能的理论。其中，神经网络已被广泛应用于智能控制系统优化、信号及信息处理、模式识别等领域。而遗传算法则是模拟生物的进化现象（自然淘汰、交叉、变异等），来表现复杂现象的一种概率搜索方法，以达到快速有效地解决各种困难问题。但神经网络和遗传算法目标相近而方法各异。因此，将这两种方法相互结合，必能达到取长补短的作用。近年来，在这方面已经取得了不少研究成果，形成了以遗传算法与神经网络相结合的进化神经网络（ＥＮＮ）。本文以综述的形式总结了遗传算法在神经网络训练中的应用情况。
Neural networks（NN） and the genetic algorithm (GA) are both theories for abtaining basic function of intelligent information by simulating a model of life which focues on information processing. Neural networks, which are inspired by a model of brain activity, have been widely applied in many fields such as optimization of intelligen t controlling system, signal and information , pattern recognition etc .. The genetic algorithm, which is regarded as a typical algorithm for evolution, is a search routine based on probability. This method repr esents complicated phenomenon by simulating evolutionary processing of life, such as natural selection, crossover, mutation. It is capable o f solving a lot of difficult problems quickly and efficiently. NN and GA take different paths to arrive at a similar goal. However, the two methods can be combined, the strong points of one offsetting the weak nesses of the other. Recently, in this aspect, much has been accomplis hed. An evolutionary neural network (ENN) has been formed, which is co mbined neural networks genetic algorithm. This paper summarizes the ap plication of GA in neural network training.
出处
《燕山大学学报》
CAS
2001年第3期234-238,共5页
Journal of Yanshan University
关键词
神经网络
遗传算法
进化神经网络
优化
Neural network, genetic algorithm, evolutionary neural n etwork (ENN).
