基于粒子群优化算法的BP网络模型在径流预测中的应用  

Runoff prediction by BP networks model based on PSO

作　　者：杨道辉[1] 马光文[1] 刘起方[1] 陶春华[1] 过夏明[1] 

机构地区：[1]四川大学水电学院,成都610065

出　　处：《水力发电学报》2006年第2期65-68,共4页Journal of Hydroelectric Engineering

基　　金：国家自然科学基金重点项目(50539140)

摘　　要：本文介绍了一种全局随机优化算法一粒子群优化算法,并将其用于BP网络模型,对模型参数进行优化来提高模型的收敛速度和精度。最后利用参数优化前后的BP模型进行径流预测,结果表明模型收敛速度和精度明显提高。Particle swarm optimizer（PSO） algorithm with the global stochastic optimization technique is introduced and applied to BP neural networks to optimize the parameters of BP neural networks so as to improve the convergence speed and precision of BP neural networks. The runoff predicting results are obtained from two different kinds of BP neural networks, it is shown that convergence speed and precision of BP neural networks are improved.

关 键 词：水资源 径流预测 粒子群算法 人工神经网络 

分 类 号：TV212[水利工程—水文学及水资源]