摘要
人工并行分区拣选策略是电子商务环境下常用的订单拣选策略,但分区工作量不均衡会从多方面降低拣选效率。本文综合考虑总服务时间和各分区工作量均衡两个因素,研究如何合理进行订单合并优化,以提高服务效率。首先构建以总服务时间最短和分区均衡度最优的双目标订单合并优化模型;采用双目标遗传算法以求解此模型:基于双目标函数的表现矩阵,将染色体对不同目标函数值的优劣进行综合排序,得到染色体适应值。针对三种不同订单环境进行数值实验,研究表明:总服务时间和分区均衡度是两个矛盾的目标;相较不考虑工作量均衡的单目标模型,双目标模型总服务时间有所增加,但增加的幅度远没有平衡度的改善幅度大;考虑分区工作量均衡的并行分区订单合并模型能从总体上提高系统服务效率;并行分区订单合并策略对小批量订单的分区均衡度提高幅度更大。
Retrieving items from warehouse in order to satisfy customer orders is one of the most important parts in e-commerce warehousing. In traditionalwarehouse operation, order picking is the most labor-intensive process that determines warehouse performance. Up to 55% of operational cost in a warehousecan be attributed to order picking. The integration strategies of synchronized zoning and order batching are often used in e-commerce order picking. However,imbalance workload distribution in each zone will reduce the whole efficiency of the picking system, such as reducing equipment utilization rate, delayingclassification and package time, and increasing pickers＇ injustice. Currently, there are few theoretical researches about integration optimization of synchronizedzoning and order batching. In this paper, we evaluate operational efficiency of e-commerce manual synchronized zone order picking system by focusing on theorder batching optimization strategy in the synchronized zone order picking system. Total service time and balance degree in each zone are two critical factorsto measure efficiency of the system. Firstly, bi-objective order batching optimization model is established by minimizing total service time and balancing workload in each zone. Let ussuppose that a picking area is divided into several equal zones and each zone is serviced by only one picker. Total service time means the whole time needed tofinish all the batched orders. Balance degree means the sum value for each batch that uses the maximum service time to minus minimum service time in eachzone. Secondly, as the order batching problem is a NP-hard one, we propose a bi-objective genetic algorithm to tackle the problem. The fitness function ranksthe gene by comparing their performance in each objective function through the sorting matrix created by the objective function. To maintain the diversity ofpopulation and control the population growth towards the direction of the optimal solution, we proposed the crossover, mutation probability function ba
作者
王旭坪
张珺
易彩玉
WANG Xu-ping ZHANG Jun YI Cai-yul(Institute of Systems Engineering, Dalian University of Technology, Dalian 116023, China School of Business, Dalian University of Technology, Panjin 124221, China)
出处
《管理工程学报》
CSSCI
CSCD
北大核心
2017年第2期209-215,共7页
Journal of Industrial Engineering and Engineering Management
基金
国家自然科学基金资助项目(71471025
71171029
71350011)
关键词
电子商务
人工并行分区拣选
工作量均衡
双目标遗传算法
E-commerce
Manual synchronized zone order picking
Workload balance
Bi-objeetive genetic algorithm
