摘要
为得出灌区用水动态预测模型和水资源优化配置模型,以甘肃疏勒河流域罐区为例,以极限学习机模型(ELM),随机森林模型(RF)和高斯指数模型(GEM)作为用水预测模型分析,以离散粒子群算法和连续粒子群算法作为水资源优化配置的方法,对灌区不同作物用水量进行了预测和分配,结果表明:ELM模型的RRMSE、R2和Ens分别为5.30%、0.974和0.975,可推荐为灌区用水预测模型;连续粒子群算法对不同作物用水量及种植结构优化配置结果与实际更为相符,效益更高,可为水资源优化配置的标准方法使用。
In order to draw in the irrigation area water dynamic prediction model and optimized allocation of water resources model,this article takes gansu shule river basin in tank farm,for example,by extreme learning machine model( ELM),random forest model( RF) and gaussian index model( GEM) as water prediction model analysis,with discrete particle swarm optimization( pso) algorithm and continuous particle swarm optimization( pso) as a means of optimal allocation of water resources to different crops irrigation area water consumption prediction and allocation,the results show that the ELM model RRMSE R2 and Ens 5. 30 % 、0. 974 and 0. 975,respectively,can be recommended for irrigation water prediction model;The results of continuous particle swarm optimization for the optimal allocation of water consumption and planting structure of different crops are more consistent with the reality and have higher benefits,which can be used as a standard method for the optimal allocation of water resources.
作者
钱瑞秋
QIAN Rui-qiu(shule river resources bureau of gansu province,Lanzhou 730030,China)
出处
《地下水》
2020年第4期83-85,共3页
Ground water
关键词
用水动态预测
水资源优化配置
极限学习机
粒子群算法
Dynamic prediction of water use
optimal allocation of water resources
ultimate learning machine
particle swarm optimization