IMPROVING ORDER PICKING LABOR UTILIZATION IN A MANUFACTURING WAREHOUSE USING A HYBRID SIMULATION–OPTIMIZATION APPROACH

Abstract

Labour utilization is a critical indicator in labour-intensive order picking systems, where excessive walking and non-value-added activities reduce operational efficiency. While prior studies apply simulation or optimization techniques, few systematically integrate experimental design with multi-algorithm benchmarking to evaluate interaction effects on labour utilization. This study proposes a hybrid framework combining discrete-event simulation, Design of Experiments (DOE), and four metaheuristic algorithms—Hill Climbing, Tabu Search, Simulated Annealing, and Particle Swarm Optimization—to analyze the effects of order size, collecting point location, and picking sequence on labour utilization. Both main and interaction effects are examined under identical warehouse conditions. Results show that all three factors significantly affect labour utilization, with notable two-way and three-way interactions. The optimal configuration—middle collecting point with PSO-based routing—improves labour utilization from approximately 55% to 77%. The study’s novelty lies in its interaction-aware benchmarking framework, which integrates simulation, factorial DOE, and comparative metaheuristic evaluation within a real warehouse setting. The findings provide practical guidance for improving labour efficiency without major capital investment.