何金涛 杨中华

摘 要：针对“货到人”拣选系统的补货环节，考虑仓库起始状态非空条件下的货位分配问题，将货架现存商品种类及数量信息与订单包含的商品种类及数量信息进行比对，做出商品分配位置以及上架数量决策，以所有货架上的商品相似度总和最大化为目标，构建了整数非线性规划模型，并设计了自适应交叉策略的遗传算法进行求解，以问题实际约束对染色体生成、交叉和变异操作进行设计。通过随机算例来对算法进行测试，结果表明文章设计的算法能够有效解决其实状态非空的货位分配问题。

关键词：“货到人”拣选；货位分配；遗传算法

中图分类号：F252；TP18 文献标志码：A DOI：10.13714/j.cnki.1002-3100.2024.10.003

Abstract： In view of the replenishment process of the "goods to people" picking system，considering the space allocation problem under the condition that the initial state of the warehouse is not empty，the type and quantity information of the existing goods on the shelf is compared with the type and quantity information of the goods contained in the order to make the decision of the distribution location and the number of goods on the shelf. An integer nonlinear programming model was constructed with the goal of maximizing the sum of similarity of goods on all shelves，and an adaptive crossover strategy genetic algorithm was designed to solve the problem. The chromosome generation，crossover and mutation operations were designed according to the practical constraints of the problem. The algorithm is tested by a random example，and the results show that the algorithm designed in this paper can effectively solve the problem of non-empty space allocation.

Key words： "goods to people" picking; space allocation; genetic algorithm

0    引    言

近年来，随着消费者需求的多样化转变，电子商务呈现出高频率、多品种、小批量的特点，对企业的仓储、分拣、订单处理等工作提出了更高的要求。作为一种新兴的拣选处理模式，“货到人”拣选系统（Robotic mobile fulfillment systems，RMFS）采用AGV（Automated Guided Vehicle）、AMR（Autonomous Mobile Robot）、AGC（Automated Guided Cart）等设备[1]将储存货物的货架、托盘等载体搬运至人工拣选站实现“货到人”的拣选。这种“货到人”拣选模式最早于2012年由Amazon应用于仓储分拣系统中，目前国内的“货到人”拣选系统的实际运用已经有阿里菜鸟联盟智能仓、京东天狼货到人系统和快仓等。与传统的“人到货”拣选模式类似，“货到人”系统也需要解决货位分配、订单分批、任务指派和路径规划等[2]问题。其中，作为拣选流程中的先决步骤，仓储系统中的货位分配工作无疑影响着后续工作的组织和效率。……