“货到人”拣选系统中电商订单分批优化研究

A Study on E-commerce Order Batching Optimizationof “Cargo-to-Person” Picking System

本文在电商行业迅速发展及人工智能日趋成熟的背景下,研究以自动导引车(AGV)为搬运工具的“货到人”拣选系统订单分批问题。首先以最小化AGV搬运货架次数为目标建立订单分批模型,然后提出了基于货架相似度的两阶段订单分批算法,在第一阶段分为新批次创建及订单加入批次两个步骤得到初始解;在第二阶段采用局部搜索改进初始解。在算法中构造订单选取货架规则,定义货架相似度函数并设计两种方法创建新批次,同时考虑周转箱数量限制求解订单分批方案。最后通过实验测试验证了模型和算法的有效性,分析了两种批次创建方法的适用性,并通过灵敏度分析给出了合理的周转箱数量配置建议。本研究可为采用“货到人”拣选系统的企业通过订单分批优化进一步提高AGV拣选效率提供理论依据和实践指导。

With the rapid development of E-commerce and artificial intelligence technology, the paper aims to address the order batching problem of “cargo-to-person” picking system which uses the automated guided vehicle(AGV)as a transport tool. First, we establish an order batching mathematical model to minimize the number of times that AGVs transport shelves. Then, this paper presents the shelf similarity-based two-phase order batching algorithm. In the first stage, it is divided into two steps of new batch creation and order addition. Then, we us local search to improve the initial solution. Inthe algorithm, we construct rule of order selecting selves, define shelf similarity functionanddesign two methods to create a new batch.Besides, the authors consider the quantity limit of the turnover box in order batch scheme. Finally, the experiment verifies the effectiveness of the model, and the algorithm proposed in this paper and we analyze the applicability of two batch creation methods. This paper provides suggestions of the optimal configuration of turnover box quantity through sensitivity analysis. The study can provide theoretical basis and practical guidance for enterprises by order batch optimization to improve the AGVs picking efficiency in “cargo-to-person” picking system.