The time-window strategy in the online order batching problem

AbstractWhen an order arrives at a warehouse it is usually assigned to a batch and a decision is made on how long to wait before assigning the batch to a picker and starting the picking tour. If the idle time of the pickers is minimised, the batch is immediately assigned, and the picking starts. Alternatively, if a time window is introduced, other orders may arrive, and more efficient batches may be formed. The method to decide how long to wait (the time-window strategy) is therefore important but, surprisingly, almost completely overlooked in the literature. In this paper, we demonstrate that this lack of attention is unwarranted, and that the time-window method significantly influences the overall warehouse performance. In the context of the online order batching problem (OOBP), we first demonstrate that the effects of different time-window strategies are independent of the methods used to solve the other subproblems of the OOBP (batching and routing). Second, we propose two new time-window strategies, compare them to existing methods, and prove that our methods outperform those in the literature under various scenarios. Finally, we show how time-window methods influence different objective functions of the OOBP when varying numbers of orders and pickers.Keywords: Online order batching problemtime windowfixed time windowvariable time windoworder pickingwarehousing Disclosure statementNo potential conflict of interest was reported by the author(s).Data availability statementThe authors confirm that the data supporting the findings of this study is freely available upon request and in the Appendix of this paper.Additional informationNotes on contributorsSergio Gil-BorrásSergio Gil-Borrás obtained his Ph.D. in Computer Science from Universidad Politécnica de Madrid in 2022. Additionally, he received his degree in Computer Engineering from the same university and completed a Master's degree in Cybersecurity from Universidad Carlos III de Madrid. He is currently working as a professor at Universidad Politécnica de Madrid and also collaborating with a research group on warehouse process optimisation, particularly the order batching problems, among other issues.Eduardo G. PardoEduardo G. Pardo received his Ph.D. in Computer Science from Universidad Rey Juan Carlos (Spain) in 2011. His research is focused on solving complex optimisation problems using Artificial Intelligence techniques. Among others, he is expert in the development of heuristic and metaheuristic algorithms. Currently, he is professor at the Computer Science School at Universidad Rey Juan Carlos (Spain).Ernesto JiménezErnesto Jiménez graduated in Computer Science from the Universidad Politécnica de Madrid (Spain) and got a Ph.D. in Computer Science from the University Rey Juan Carlos (Spain) in 2004. His research interests include Fault Tolerance in Distributed Systems, Computer Networks and Parallel and Distributed Processing. He is currently an associate professor at the Universidad Politécnica de Madrid.Kenneth SörensenKenneth Sörensen earned his Ph.D. from the University of Antwerp, Belgium, in 2003. He specialises in using Artificial Intelligence to tackle intricate optimisation challenges. His expertise notably lies in crafting heuristic and metaheuristic algorithms. At present, he is a full professor at the University of Antwerp.