Optimizing Order Consolidation with Simulation Optimization
MetadataShow full item record
In this work we study the order consolidation process of an e-retailing warehouse using real data. Order consolidation is a process that follows the picking operation in a warehouse. After stock keeping units (SKUs) of an order are picked into possibly multiple totes, they are sorted and the SKUs are consolidated back into orders to be packed and shipped to customers. There are two main operations in order consolidation: wave sortation and putting. Wave sortation involves tote sequencing which is a major driver of how smooth the system runs. First, totes should be sequenced so that the work load is balanced. Second, totes that carry SKUs of the same order should be close in sequence in order to reduce order processing times, and to reduce resource utilization. At the same time, orders should be assigned to operators for putting in order to balance operator workload and to avoid recirculation. The assignment and sequencing of totes, when assuming known and constant tote induction times and when the goal is total completion time, is a problem similar to parallel machine scheduling. However, the goal from order consolidation is minimizing order, as opposed to tote, consolidation times. Order consolidation also depends on the putting operation. Moreover, processing times are uncertain and may vary because of different sizes and weights of SKUs and because of variations in manual and automatic processes. In order to reflect a realistic system performance accurately, we need to take randomness into account when evaluating measures like order completion time, resource utilization, and congestion. In this thesis, we first build a simulation model for the consolidation process using a given tote assignment and sequencing. We use the empirical distributions derived from the data to run the simulation. Second, we develop a simulated annealing heuristic and apply it to solve the deterministic tote sequencing problem with the goal of minimizing order completion time or order processing time. Third, we develop a stochastic simulated annealing algorithm to optimize the whole consolidation process. The algorithm decides on tote assignment and sequencing and evaluates the sequences using the simulation model. We experimented with optimizing order completion times and SKU wait times for putting. Our approach is then a simulation optimization to optimize the consolidation process using different metrics. We report on numerical tests using real instances.
Cite this version of the work
Daren Zhou (2017). Optimizing Order Consolidation with Simulation Optimization. UWSpace. http://hdl.handle.net/10012/11893