Clustering techniques for stock location and order-picking in a distribution center

Abstract

In this paper, the stock location and order-picking problems in a distribution center where items are distributed in less-than-case-lot quantities are addressed. Clustering techniques are applied to group items in the slots of gravity-flow racks and to sequence the picking lists by customers. Based on the order-item-quantity rule, two similarity measures are defined for items and customers, respectively. A zero–one integer programming model developed for optimal items- and customers-grouping is presented. An exact primal–dual type algorithm is explored and implemented, where real world data collected from a local distribution center is employed. Simulation results indicate the potential benefits of the proposed techniques. Scope and purpose In this article, issues of stock location and order-picking in a distribution center where the characteristics of dependent customer demands are thoroughly examined. Clustering techniques are applied to extract the correlated information from the customer orders, which is then used for optimizing the stock location and picking process. Instead of applying existing algorithms, a novel primal–dual-type algorithm explored to solve the clustering problem is developed. An example using real world data collected from a local distribution center is presented. Simulation studies using WITNESS simulator reveal the benefits of the stock location achieved by the proposed algorithm.