Near-Optimal Graph-Based Routing for Manual Picker-to-Parts Warehouses: A Case Study of an Apulian Distribution Center

Abstract

Order picking is one of the most costly activities in manual picker-to-parts warehouses, and routing quality has a direct impact on travel effort and operational efficiency. However, commercial warehouse management systems (WMSs) still often rely on simple rule-based policies because exact optimization is difficult to reconcile with real-time execution requirements. This paper presents a WMS-compatible graph-based routing method for manual warehouses based on mission-dependent graph reduction, shortest-path computation, and sequencing optimization over mission-relevant locations. Starting from the physical warehouse graph, the proposed method builds a compact reduced representation that preserves shortest-path distances while significantly decreasing the online computational burden, enabling seamless integration into existing WMSs and real-time operation. The method is validated on a real household-goods distribution center and compared with both practical rule-based routing policies typically adopted in commercial WMSs and exact optimization benchmarks. Results on both synthetic missions and real warehouse orders show substantial travel-distance reductions with runtimes fully compatible with online warehouse operation.