带退货和软时间窗的多仓库选址-路径问题研究  

Research on Multi-Depot Location Routing Problem with Backhauls and Soft Time Windows

作　　者：罗耀波 孙延明[1] 廖鹏[1] 

机构地区：[1]华南理工大学工商管理学院,广东广州510641

出　　处：《运筹与管理》2014年第5期78-85,共8页Operations Research and Management Science

基　　金：国家自然科学基金项目(71071059);国家自然科学基金项目(50675069);国家自然科学基金资助项目(71071057);中央高校基本科研业务费专项资金资助项目(2012ZMO031)

摘　　要：选址-路径问题(location routing problems,LRP)是集成物流网络研究中的难题,也是任何一个大型物流配送企业必须面对的管理决策问题。本文在仓库容量约束和车辆容量约束的基础上,结合送取货一体化的配送模式和客户服务时间要求,建立了带退货和软时间窗的多仓库选址-路径(MDLRP)数学模型。针对MDLRP问题求解的复杂性,引入局部搜索算法和重组策略,设计了自适应混合遗传算法,对模型进行整体求解。最后进行数值实验,表明本文提出的模型和改进算法具有实用性和优越性,可为选址和车辆运输决策提供重要参考依据。Location routing problems is not only of great significance in integrated logistics network planning research , but also an important management decision that every large logistics company has to make .Based on the warehouse capacity and vehicle capacity constraints , the paper proposes a multi-depot location routing prob-lem model（MDLRP）with backhauls and soft time windows .The model takes full consideration of logistics distri-bution mode with the simultaneous delivery and pick-up and the customer service time requirements .Given the complexity of the MDLRP model , the paper proposes an improved hybrid genetic algorithm with iterated local search and recombination strategy to solve the model integrally .The performance of the heuristic is assessed by computational experiments .As can be seen from the solution , the model with its hybrid genetic algorithm is feasible and superior , and it can be provided to be an alternative tool for location and routing decision .

关 键 词：选址-路径问题 集成物流网络 带退货 软时间窗 遗传算法 

分 类 号：F253.4[经济管理—国民经济]