关键词: 无人仓, 货到人, 订单分批, 分类型数据, 取大运算, K-max聚类算法, 加权距离

Abstract: To improve the efficiency and reduce the cost of order picking,the order batching problem of unmanned warehouse system based on Automatic Guided Vehicle (AGV) was investigated.Two main factors affecting the cost and efficiency of order picking under “parts-to-picker” mode were considered：the number of item categories being picked up from the shelves,and the number of shelves being transported to working station.An integer programming model to minimize the total cost of order picking was established.Based on the items category in a batch of orders and the shelves to be transported when picking the batch of orders,two types of clustering centers related to the cost of order picking were defined using “max” operator respectively,the first clustering center was based on item categories to be picked and the second clustering center was based on the shelves to be transported.Two types of distances from an order to two clustering centers were defined respectively.Furthermore,the weighted distance from an order to a batch (clustering center) was formulated,and the weight coefficients were the unit cost of worker picking up one type of items and the unit cost of AGV moving one shelf.Based on the idea of K-means clustering algorithm,a new clustering algorithm named K-max clustering was designed to solve the order batching problem based on the categorical data.A numerical example was used to analyze the effective of K-max clustering algorithm.The sensitivity analysis of the weighted coefficients was given,and the effectiveness of K-max clustering algorithm for solving the order batching problem was verified.

Key words: unmanned warehouse, parts to picker, order batching, categorical data, max operator, K-max clustering algorithm, weighted distance