Simulation-based Optimization for Integrating Storage Location Assignment and Manual Order Picking in Warehousing

We propose a simulation-based optimization (SO) framework that integrates a novel mathematical formulation for the storage location assignment problem (SLAP) featuring vertical replenishment and a simulation-guided heuristic for manual order picking. Due to the highly-combinatorial nature of the problem, to solve realistic instances, we design a customized iterated local search (ILS) metaheuristic. Specifically, it determines the "top m" promising moves at each iteration of the ILS and then evaluates the corresponding solutions by discrete-event simulation. This allows measuring the impact of location assignments on warehouse productivity, while also accounting for mutual interferences and waiting phenomena arising during the operations carried out by multiple order pickers. The solution bearing the best (estimated) throughput maximization-based objective function of the order picking process is used to update the global optimum. Hence, simulation guides the search and change of the current feasible SLAP solution. Numerical results are presented for a real distribution center under a picker-to-products rack system following an S-shape policy with a skip-and-go rule to deal with lacking items. After a proper tuning of the ILS parameters and once the ILS has assigned the higher-rotating items as close as possible to the loading gate, the SO framework allows to achieve significant improvements on warehouse productivity.