摘要
准确的含沙量预测对于河道的治理以及防洪具有重要意义,本文结合实测数据,在BP神经网络的基础上,针对BP神经网络易陷入局部最小值的缺点,采用人工蜂群算法对权值、阈值进行优化,得到最优的初始权值、阈值,建立了基于ABC-BP的含沙量预测模型。在训练样本中,拟合精度达到了100%;在预测样本中,误差较单纯的BP神经网络有了较大的改善,相对误差最低仅为9.8%,有着较好的预测精度,该方法可为以后定量开展河流含沙量预测研究提供参考。
Accurate prediction of sediment concentration is of great significance for river regulation and flood control.Based on the measured data and BP neural network,a sediment concentration prediction model based on ABC-BP is established,where the artificial bee colony algorithm is used to optimize the weights and threshold values by reducing the impact of local minimum on BP neural network to obtain the optimal initial weights and thresholds.Among the training samples,the accuracy of fitting reaches 100%;among the prediction samples,the BP neural network with simple errors has been significantly improved with the lowest relative error only 9.8%.Underpinned by a better prediction accuracy,this method can provide a reference for the future studies on river sediment concentration prediction.
作者
杨加璐
杨奉广
Yang Jialu;Yang Fengguang
出处
《吉林水利》
2020年第12期5-7,12,共4页
Jilin Water Resources
基金
国家自然科学基金资助项目(51979180)。
关键词
神经网络
人工蜂群算法
ABC-BP
年均含沙量
neural network
artificial bee colony algorithm
ABC-BP
average annual sediment concentration