人工神经网络与遗传算法的结合：进展及展望

近来，人工神经网络（ＡＮＮ）与遗传算法（ＧＡ）结合的研究引起了人们的极大注意。本文综述了这一领域的发展现状，并试图从两者的结合上分为辅助式及合作式两种结合方式，对于辅助式结合，ＧＡ不仅能为ＡＮＮ选择训练数据，也可用来选择网络的学习参数或学习规则，而且可利用ＧＡ解释或分析ＡＮＮ的结果；对于合作式结合，主要论述了ＧＡ作为ＡＮＮ的一种权重训练方法的不足和长处，同时评述了它与ＢＰ训练算法的优缺点，最后着重综述了利用ＧＡ自动优选ＡＮＮ的各种方法。

Research involving some sort of combination of Genetic Algorithms (GAs) and Artificial Neural Networks (ANNs) has attracted a lot of attention recently. Some current work combining GAs and ANNs can be divided into two broad categories: supportive combinations, which typically involve using a genetic algorithm to select training data and to interpret the output behavior of neural networks; collaborative combinations, which typically involve using the GA to determine the neural network weights or the network topology or both. This article presents a review of the state of art and research prospects in this area.