滴灌轮灌分组优化模型与算法  

Optimization model and algorithm of rotation irrigation group for drip irrigation

作　　者：李伟 陈伟能[1] 田敏[3] 邓红涛 陈红莉 Li Wei;Chen Weineng;Tian Min;Deng Hongtao;Chen Hongli(School of Computer Science and Engineering,South China University of Technology,Guangzhou 510000,China;School of Computer Science and Technology,Shihezi University,Shihezi 832000,China;School of Mechanical Electrical and Engineering,Shihezi University,Shihezi 832000,China;Xinjiang Tianye Group Co.,Ltd.,Shihezi 832000,China)

机构地区：[1]华南理工大学计算机科学与工程学院,广州510000 [2]石河子大学信息科学与技术学院,石河子832000 [3]石河子大学机械电气工程学院,石河子832000 [4]新疆天业集团有限公司,石河子832000

出　　处：《农业工程学报》2021年第10期73-81,共9页Transactions of the Chinese Society of Agricultural Engineering

基　　金：国家自然科学基金资助项目(61962053,61976093)。

摘　　要：传统的轮灌组划分计算方式效率较低且难以获得较好方案。该研究首次采用智能算法来求解轮灌组划分问题,依据《微灌工程技术标准》及轮灌组划分原则,提出了基于流量均衡的数学模型及其约束条件。通过分析支管空间分布,确定了滴灌问题的邻域特征,在半径阈值范围内给出了最大限度查找关键路径的邻域搜索策略和不可行解修复算法,并采用传统遗传算法(Genetic Algorithm,GA)、贪心遗传算法(Greedy-GA)、泰森多边形遗传算法(Voronoi-GA)和网格遗传算法(Grid-GA)算法分别求解模型,探索适应轮灌分组问题的初始化方法。对标准差、组内路程、连通性和运行时间4项指标对比分析,结果表明:Grid-GA算法表现优异,采用的邻域策略可有效避免支管分布过于分散,有利于日常管理与维护。取半径阈值280 m条件下,算法在300代左右达到收敛,最小标准差10.9 m3/h,组内路程8105.2 m,连通性指标25,与一种冒泡+贪心的近似算法相比最小标准差小59.1%。该研究对提高滴灌工程设计效率和促进轮灌工作制度有效运行有着重要研究意义。A rotation irrigation is normally implemented in water-shortage regions,such as Xinjiang area of western China with widespread drip irrigation.Nevertheless,the conventional rotation irrigation group is generally divided into the fixed districts in the manual calculation.A better solution is thus highly demanding in this inefficient calculation.The rotation irrigation group is also representing the prominent regional and hydraulic characteristics in recent years.Therefore,it is necessary to clarify how to divide the rotation irrigation group in practice.Furthermore,the operation of the irrigation system needs to gradually complete the whole plot irrigation under the turn-on and turn-off valve sequence of several branch pipes in a rotation irrigation group.Correspondingly,there is a relatively high labor intensity in the operational mode,but with low investment cost,simple maintenance,and wide popularization,compared with the automatic operation of drip irrigation.In this study,a hybrid mathematical model was proposed to explore a better solution using the flow balance and structural constraints in the technical standards and division principles.Neighborhood characteristics were determined from the spatial distribution of branch pipes in the rotation irrigation group.The neighborhood search strategy and the repair of infeasible solution were given in the radius threshold range,further to find the critical path of the maximum extent.Four algorithms were selected to solve the model separately,including the Genetic Algorithm(GA),Greedy-GA,Tyson polygon-GA(Voronoi-GA),and Grid GA.The GA adopted the branch pipe model of actual number coding.The chromosome coding indicated that the branch pipe was opened in that group.The initial population was constructed using the Random,Greedy,Voronoi polygon,and Grid.The spatial distribution of branch pipe was applied in the neighborhood search strategy,further to serve as the structural feature in the rotation irrigation group.A uniform crossover strategy was adopted for the crossov

关 键 词：模型 算法 网格法 滴灌 轮灌组 遗传算法 智能优化算法 

分 类 号：S274.1[农业科学—农业水土工程]