引用本文:崔东文,包艳飞.基于人工生态系统优化算法的组合生长需水预测模型[J].水资源保护,2020,36(6):122-130DOI:10.3880/j.issn.1004-6933.2020.06.020
CUI Dongwen,BAO Yanfei.Combined growth water demand forecasting model based on artificial ecosystem optimization algorithm[J].Water Resources Protection,2020,36(6):122-130.
【打印本页】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 207次   下载 187 本文二维码信息
码上扫一扫!
分享到: 微信 更多
基于人工生态系统优化算法的组合生长需水预测模型
崔东文1, 包艳飞2
1.云南省文山州水务局, 云南 文山 663000;2.云南省水文水资源局曲靖分局, 云南 曲靖 655000
摘要:为提高需水预测精度,拓展生长模型在需水预测中的应用,提出基于人工生态系统优化(AEO)算法的组合生长需水预测模型。结合实例,选取6个标准测试函数在不同维度条件下对AEO算法进行仿真验证,并与鲸鱼优化算法(WOA)、灰狼优化(GWO)算法、教学优化(TLBO)算法和传统粒子群优化(PSO)算法的仿真结果进行比较。基于Weibull、Richards、Usher 3种单一生长模型构建Weibull-Richards-Usher、Weibull-Richards、Weibull-Usher、Richards-Usher 4种组合生长模型,利用AEO算法同时对组合模型参数和权重系数进行优化,提出AEO-Weibull-Richards-Usher、AEO-Weibull-Richards、AEO-Weibull-Usher、AEO-Richards-Usher需水预测模型,并构建AEO-Weibull、AEO-Richards、AEO-Usher、AEO-SVM、AEO-BP模型作对比,以上海市需水预测为例进行实例验证,利用实例前30组和后8组统计资料对各组合模型进行训练和预测。结果表明,在不同维度条件下,AEO算法寻优精度优于WOA、GWO、TLBO、PSO算法,具有较好的寻优精度和全局搜索能力。4种组合模型对实例预测的平均相对误差绝对值、平均绝对误差分别在0.94%~1.17%、0.30亿~0.37亿m3之间,预测精度优于AEO-Weibull等其他5种模型。4种组合模型均具有较好的预测精度和泛化能力,表明AEO算法能同时有效优化组合生长模型参数和权重系数,基于AEO算法的组合生长模型用于需水预测是可行和有效的。
关键词:  需水预测  人工生态系统优化算法  组合生长模型  Weibull模型  Richards模型  Usher模型  参数优化
DOI:10.3880/j.issn.1004-6933.2020.06.020
基金项目:
Combined growth water demand forecasting model based on artificial ecosystem optimization algorithm
CUI Dongwen1, BAO Yanfei2
1.Wenshan WaterAffairs Bureau of Yunnan Province, Wenshan 663000, China;2.Qujing Branch of Yunnan Hydrology and Water Resources Bureau, Qujing 655000, China
Abstract:In order to improve the accuracy of water demand forecasting and expand the application of growth model in water demand forecasting, a combined growth water demand prediction model based on artificial ecosystem optimization(AEO)algorithm was proposed. Combined with an example, six standard test functions were selected to simulate AEO algorithm in different dimensions, and the simulation results were compared with those of Whale optimization algorithm(WOA), gray wolf optimization(GWO), teaching optimization(TLBO)algorithm and traditional particle swarm optimization(PSO). Based on the combination of three single growth models(Weibull, Richards, and Usher), Weibull-Richards-Usher, Weibull-Richards, Weibull-Usher and Richards-Usher were constructed. The AEO algorithm was used to optimize the parameters and weight coefficients of the four combined growth models. The AEO-Weibull-Richards-Usher, AEO-Weibull-Richards, AEO-Weibull-Usher, AEO-Richards-Usher water demand forecasting models were proposed and AEO-Weibull, AEO-Richards, AEO-Usher, AEO-SVM, AEO-BP models were constructed for comparison. Taking the water demand forecast of Shanghai as an example, the combined models were trained and predicted by using the statistical data of the first 30 groups and the last 8 groups. The results show that the optimization accuracy of AEO algorithm is better than that of WOA, GWO, TLBO and PSO algorithms in different dimensions, and has better optimization accuracy and global search ability. The average absolute relative error and the average absolute error of the four combined models are 0.94%-1.17% and 30 million -37 million m3, respectively. The prediction accuracy is better than the other five models such as AEO-Weibull. The results show that the AEO algorithm can effectively optimize the parameters and weight coefficients of the combined growth model, and the combined growth model based on AEO algorithm is feasible and effective for water demand prediction.
Key words:  water demand forecast  artificial ecosystem optimization algorithm  combined growth model  Weibull model  Richards model  Usher model  sparameter optimization

手机扫一扫看
分享按钮