Session 6: Warehousing – Chair: Jelmer van der Gaast
Masoud Mirzaei (Erasmus University Rotterdam) – Store smarter, retrieve faster: An integrated cluster-based storage assignment policy
Order picking is a labor intensive process in many warehouses. Its efficiency depends largely on the storage assignment policy used. Current storage assignment policies, such as turnover-frequency class-based storage policies ignore information on product affinity, that is the frequency by which products are ordered jointly. We propose an integrated cluster allocation (ICA) storage assignment model that uses both information on product turnover and affinity to assign products to the storage locations. As the problem is NP-hard and the number of products and locations is very large in practice, we use a heuristic to solve it. We generate a greedy high-quality construction heuristic which is used as an initial solution for a solver which yields considerable savings. Compared to class-based and full turnover-based storage policies, the ICA storage policy can save up to 22%, and 30% respectively, in load retrieval time, for low and medium product affinity levels, assuming an automated storage and retrieval system is used. The ICA storage policy appears to be quite robust against new incoming orders. In a real-life test, using order and product data of a large wholesale company with low affinity level, we found that the ICA policy outperforms class-based and full-turnover based storage for an automated storage and retrieval system. However, the ICA storage policy is data intensive and obtained benefit depends largely on the order patterns.
Debjit Roy (Indian Institute of Management Ahmedabad) – Modeling parallel resource movements in intra-logistics systems
Many intra-logistics systems, such as automated container terminals, distribution warehouses, and cross-docks observe parallel process flows, which involve simultaneous (parallel) operations of multiple independent resources while processing a job. Stochastic modeling of such process flows is complex, hence researchers tend to use a `sequential’ modeling approach which assumes serial operations of the resources. To model parallel process flows, we develop an analytical method, the `parallel’ modeling approach, using a stylized closed queuing network with two-phase servers. To analyze the resulting network, we devise solution methods based on the network aggregation dis-aggregation technique. The parallel modeling approach is demonstrated using case studies based on the seaside operations at an automated container terminal with automated guided vehicles (AGVs) and robotized distribution warehousing system.
Jelmer van der Gaast (University of Groningen) – Dynamic batch picking for order picking in warehouses
Dynamic batch picking is characterized by combining product demand from multiple customer orders into one pick tour where new orders are received continuously. Using modern order-picking aids, updated picking instructions can be included in the current pick tours which allows pickers to be re-routed to pick for new orders even when they already started a pick tour. We develop a mathematical model for dynamic batch picking that minimizes the order throughput time of an incoming customer order. Using branch-and-price, we can quickly re-optimize the model in case of a new order arrival and determine new updated pick tours. This allows for short order throughput times and ensures that warehouse companies can set their order cut-off times as late as possible while still guaranteeing that orders can be delivered next day or in some cases even the same day.