基于文化基因算法的多供应商多零售商VMI问题研究  

The research of multi-vendor multi-retailer VMI problem based on memetic algorithm

作　　者：刘琳琳[1] 杨斌[1] 朱小林[2] 张志强[3] 

机构地区：[1]上海海事大学科学研究院物流研究中心,上海201306 [2]上海海事大学文理学院,上海201306 [3]北京科技大学机械工程学院,北京100083

出　　处：《现代制造工程》2017年第2期1-7,13,共8页Modern Manufacturing Engineering

基　　金：国家自然科学基金资助项目(71171129);上海市科委科研计划项目(12510501600;14DZ2280200;14511107402)

摘　　要：针对多供应商-多零售商-单个中心仓库供应链背景下的供应商管理库存问题,考虑供应商、零售商以及中心仓库的订货量和订货频次的约束,构建了供应商管理库存问题总库存成本的数学模型;针对上述模型,采用文化基因算法进行求解,以总的库存成本最低为目标来求解供应商、零售商及中心仓库每年的订货频次和订货量。该算法将个体表示为VMI问题的解,建立问题和算法的映射关系;在全局搜索过程中,采用基于位置的交叉方式吸收最优个体的优秀基因;在局部搜索过程中,采用基于位置的变异方式产生一定规模的邻域基因个体。采用正交试验对算法的重要参数进行了优化配置,通过实例验证了算法的有效性。Under the background of multi-vendor muhi-retailer single central warehouse supply chain, taking into account the con- straints of the order quantities of the supplier, retailer, central warehouse, and the constrains of their order frequency, the total warehouse costs of vendor managed inventory problem mathematical model was established. To solve this model, aiming at the minimum total inventoy cost, memetic algorithm is used to solve the annual ordering frequency of the supplier, retailer, central warehouse, and the order quantity. An encoding way is built to bridge the VMI and the algorithm. In the process of global search, to absorb the genes of the best individual, use the crossover way that is based on the cross-way location. In the process of local search,to produce a certain scale neighborhood gene individuals, use the mutation way that is based on location. To get the best parameters of the algorithm, the orthogonal experimental design is used. The algorithm is used to solve a specific problem. The ex- ample shows the effectiveness of the algorithm.

关 键 词：多供应商 多零售商 供应商管理库存 VMI 文化基因算法 供应链 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]