摘要
基于梯度下降的BP人工神经网络应用广泛,但网络目标函数误差曲面极其复杂,网络初始值的选取对网络训练结果影响很大,导致收敛速度慢,容易陷入局部极小等问题.基于粒子群算法(pso)的训练方法能够摆脱陷入局部最优的困扰,但粒子群算法局部搜索能力不够,影响网络的训练效果,在充分研究两种算法特点的基础上,提出一种新的组合训练方法,建立了PSO-BP组合人工神经网络模型.
BP artificial neural network based on gradient algorithm method is widely applied,but because the error surface of object function is extramly complex and the choose of initial value effects network training results,convergence rate is slow and local minimum is likely to fall into.Paticle swarm optimization algorithm has better global searching ability to get rid the puzzles of falling into local minimum.By adequately studying on the two algorithms＇ characteristics,a new type of combined artificial neural network training method is put forward,and PSO-BP ann model is successfully built.
出处
《山西大同大学学报（自然科学版）》
2010年第3期66-69,共4页
Journal of Shanxi Datong University(Natural Science Edition)
关键词
组合人工神经网络
BP算法
PSO算法
算法设计
combined artificial neural network
bp algorithm
pso algorithm
algorithm design
