基于MEA-GA-BP耦合模型的城市需水量预测  

Urban Water Demand Prediction Based on MEA-GA-BP Coupling Model

作　　者：苏志伟 吴元梅 张丽娟[1] 杨旭亮 SU Zhi-wei;WU Yuan-mei;ZHANG Li-juan;YANG Xu-liang(School of Civil and Transportation Engineering,Qinghai Minzu University,Xining 810007,China;Qinghai Water Conservancy and Hydropower Research Institute Co.,Ltd.,Xining 8100000,China)

机构地区：[1]青海民族大学土木与交通工程学院,青海西宁810007 [2]青海省水利水电科学研究院有限公司,青海西宁810000

出　　处：《水电能源科学》2022年第11期31-34,共4页Water Resources and Power

基　　金：青海省应用基础研究项目(2020-ZJ-756);青海民族大学校级教学改革项目(2021-JYQN-007)。

摘　　要：为精确预测城市需水量,针对BP神经网络及其优化方法收敛速度慢、易陷入局部最优等缺点,提出一种思维进化算法与遗传算法联合优化的BP神经网络,构建MEA-GA-BP耦合需水预测模型,同时建立GA-BP预测模型作为对比模型。实例分析表明,MEA-GA-BP耦合模型具有更高的预测精度和预测速度,有效解决了BP神经网络存在的缺点,可作为一种有效的需水预测模型。In order to accurately predict urban water demand, aiming at the shortcomings of BP neural network and its optimization method, such as slow convergence speed and easy to fall into local optimum, a BP neural network optimized by mind evolutionary algorithm and genetic algorithm was proposed. The MEA-GA-BP coupled water demand prediction model was constructed, and the GA-BP prediction model was established as the comparison model. The example analysis shows that the MEA-GA-BP coupling model has higher prediction accuracy and prediction speed, and effectively solves the shortcomings of BP neural network, which can be used as an effective water demand prediction model.

关 键 词：需水预测 BP神经网络 思维进化算法 遗传算法 

分 类 号：TV213.4[水利工程—水文学及水资源]