Adaptive Algorithm for Storage Location Assignment Problem under the Mixed Shelves Strategy in the Warehouse

Abstract

This paper addresses an algorithm for optimizing the storage location assignment problem in a warehouse operating under a mixed-shelves strategy. Under this strategy, each SKU (Stock Keeping Unit) can be assigned to multiple storage locations, allowing various SKUs to be positioned close to one another and thus improving order-picking efficiency. The mixed-shelves strategy is particularly effective for e-commerce retailers where many kinds of products are ordered in small quantities. However, the number of storage positions to be considered increases substantially, making the problem more complex.

To address this issue, this paper proposes a simulated annealing-based algorithm (SA) for simultaneously solving the storage location assignment problem and the order picker routing problem. The proposed SA employs a neighborhood structure based on a normal distribution that enables the algorithm to consider SKU demand efficiently; in this mechanism, high-demand SKUs are assigned to shelves near the depot with higher probability. In addition, a correlation-based neighborhood is introduced to account for the frequency with which SKUs are ordered together. In this neighborhood, pairs of SKUs with high correlation are assigned to the same shelf with higher probability. Moreover, to accommodate various types of orders, both neighborhoods are selected probabilistically and adaptively during the search process.

The effectiveness of the proposed algorithm was validated using real-sized datasets. In numerical experiments, compared with existing methods that do not consider SKU demand or SKU correlations, the proposed algorithm significantly reduced total travel distance. These findings highlight the importance of simultaneously considering both demand and correlation in the storage location assignment problem.