摘要
针对大坝安全监控中小波神经网络模型(WNN)训练时间较长且易陷入局部极小值的缺陷,提出采用粒子群优化算法(PSO)取代传统的梯度下降法对小波神经网络中的各参数进行优化,建立了PSO-WNN模型并用于大坝安全监测的拟合和预报。实例结果表明,PSO-WNN模型收敛速度快、预测精度及稳定性高,为大坝变形监测分析提供了一种有效的新型建模方法。
The wavelet neural network(WNN) models applying for dam safety monitoring have some drawbacks which the training time is long and the calculations easily fall into local minimum.In order to overcome these defects,particle swarm optimization(PSO) algorithm is proposed to replace the traditional gradient-descent algorithm for optimizing parameters of wavelet neural network.The PSO-WNN model is established to apply for fitting and forecasting the monitoring data of dam deformation.The results show that the PSO...
出处
《水电能源科学》
北大核心
2010年第6期61-63,共3页
Water Resources and Power
基金
河海大学水文水资源与水利工程科学国家重点实验室专项基金资助项目(2009586012)
关键词
粒子群优化算法
小波神经网络
大坝变形监测
预报
安全监控模型
particle swarm optimization
wavelet neural network
dam deformation monitoring
forecasting
safety monitoring model
