计算机应用与软件2011，Vol.28Issue(1)：249-252,4.
基于改进遗传算法的神经网络优化设计
OPTIMISED DESIGN OF NEURAL NETWORKS BASED ON IMPROVED GENETIC ALGORITHM
胡仁平 1刘刚1
作者信息
- 1. 湖南铁路科技职业技术学院机电工程系,湖南,株洲,412000
- 折叠
摘要
Abstract
In view of the limitations in structure of feed-forward multilayer neural networks and weights design method, an automatic optimised design approach for feed-forward neural networks based on improved genetic algorithm was presented for completing the search of network structure and weights space and improving the neural network convergence speed and the capability of global optimal solution search.Experiments showed that the convergence of the method is fast and the convergent process is stable, and it has good networks generalization ability as well. Therefore this method can play a preferable good role in the design of neural networks.关键词
神经网络/遗传算法/优化设计引用本文复制引用
胡仁平,刘刚..基于改进遗传算法的神经网络优化设计[J].计算机应用与软件,2011,28(1):249-252,4.
