A Simplified Linear Programming Model for the Assignment of Duplicate Storage Locations

Abstract

Research background: In modern warehouses, there is a strong emphasis on ensuring fast and efficient order-picking. Speed is a critical factor, as customers increasingly demand quick delivery times, especially in the e-commerce sector. Warehouse management optimizes the storage of items and picking routes to minimize delays and maximize productivity, ensuring orders are processed and shipped as quickly as possible. Additional benefits in this field can be achieved by the scattered storage of items. In this paper, we propose a new optimization model that minimizes the weighted sum of shortest distances between picking aisles intended for the storage of correlated items. We assume a decentralized pick-up/drop-off and a random storage of items in the assigned picking aisle. Unlike existing proposals based on exact location assignment, the presented MILP model does not need the distance matrix between storage locations (or picking aisles). Purpose: Optimal storage location assignment for warehouses with scattered storage. Research methodology: We use mixed integer linear programming (MILP) models (for optimal storage location assignment) and simulations (for verification of the obtained results). Results: The adopted simplifications reduce the size of the model by over 99% and allow feasible solutions to be found. We show that for the obtained feasible solutions there is a significant reduction in average order picking times. Additionally, we discuss methods of determining the correlation coefficients between the items. Novelty: Original storage location assignment concept and optimization model.