摘要
介绍了传统的遗传算法与BP算法,并分析了它们的不足。提出了一种将遗传算法和BP算法相结合的遗传BP算法,在遗传算法的群体设计中采用小生境技术,防止优秀个体早期退化。实验证明了此方法的正确性和有效性。
In this paper, the principle and characteristics of genetic algorithm and BP algorithm are (introduced). Their disadvantages are analyzed. A new algorithm combining the genetic algorithm and back propagation algorithm is proposed. In genetic algorithm, the niche technology is applied to avoid early (degeneration) of a excellent network. The experiment illustrates that the new algorithm has higher (efficiency).
出处
《长沙交通学院学报》
2005年第1期53-56,共4页
Journal of Changsha Communications University
关键词
BP算法
遗传算法
小生境技术
神经网络
模式识别
BP algorithm
genetic algorithm
niche technique
neural network
pattern recognition
