摘要
针对RBF(RadialBasisFunction)网特点 ,提出基于两层编码的进化算法来同时确定网络结构和参数。文中分别采用两种不同的遗传算法GA (GeneticAlgorithm )和进化规划EP (EvolutionaryProgramming)训练网络。实验结果表明 ,基于这三种不同进化算法的RBF网络都能有效地控制混沌系统 。
Evolutionary algorithms based on two layer encoding for RBF neural networks are proposed to acquire both the topology and parameters of network during training.Two different genetic algorithms(GA) and evolutionaray programming have been respectively used to train network in this paper.Experimental results show that RBF neutral networks based on the three different algorithms effectively control chaos system,where the training and controlling results of EP are optimization.
出处
《华东船舶工业学院学报》
EI
2001年第6期26-29,共4页
Journal of East China Shipbuilding Institute(Natural Science Edition)
关键词
两层编码
进化算法
RBF神经网络
混沌
控制
evolutionary algorithms
radial basis function neural networks
chaos
control
