Real-time order picking of a robotic put wall: a simulation-based metaheuristic optimisation

A robotic put wall has the potential to significantly enhance picking productivity in the logistics industry. This paper introduces a new computational method for scheduling a robotic put wall system that processes randomly arriving items. The method comprises a simulation-based model and a customised metaheuristic that optimises performance at regular intervals. The simulation model is developed using advanced discrete-event software that can include operational details of the picking process. The genetic algorithm with a new encoding scheme is tailored to solve the combinatorial optimisation problem of determining the appropriate destinations. To evaluate the proposed method, case studies based on real-world applications in a put wall manufacturing company were used. The method outperforms three rule-based real-time scheduling methods, as demonstrated by the results. Moreover, the integrated approach can determine the minimum number of vehicles required.