A Hybrid Algorithm for Computing Tours in a Spare Parts Warehouse

We consider a real-world problem arising in a warehouse for spare parts. Items ordered by customers shall be collected and for this purpose our task is to determine efficient pickup tours within the warehouse. The algorithm we propose embeds a dynamic programming algorithm for computing individual optimal walks through the warehouse in a general variable neighborhood search (VNS) scheme. To enhance the performance of our approach we introduce a new self-adaptive variable neighborhood descent used as local improvement procedure within VNS. Experimental results indicate that our method provides valuable pickup plans, whereas the computation times are kept low and several constraints typically stated by spare parts suppliers are fulfilled.

• Type: Conference
• Year: 2009
• Where: EVOW
• Authors: Matthias Prandtstetter, Günther R. Raidl, Thomas Misar
• Keywords: Artificial Intelligence, EVOW 2009, Self-adaptive Variable Neighborhood, Spare Parts, Variable Neighborhood