Template-Type: ReDIF-Paper 1.0
Author-Name: Abbink, E.J.W.
Author-Name-Last: Abbink
Author-Name-First: Erwin
Title: Solving large scale crew scheduling problems by using iterative partitioning
Abstract: This paper deals with large-scale crew scheduling problems arising at the Dutch railway operator, Netherlands Railways (NS). NS operates about 30,000 trains a week. All these trains need a driver and a certain number of conductors. No available crew scheduling algorithm can solve such huge instances at once. A common approach to deal with these huge weekly instances, is to split them into several daily instances and solve those separately. However, we found out that this can be rather inefficient.  In this paper, we discuss several methods to partition huge instances into several smaller ones. These smaller instances are then solved with the commercially available crew scheduling algorithm TURNI. We compare these partitioning methods with each other, and we report several results where we applied different partitioning methods after each other. The results show that all methods significantly improve the solution. With the best approach, we were able to cut down crew costs with about 2\\% (about 6 million euro per year).
Creation-Date: 2008-03-18
File-URL: https://repub.eur.nl/pub/11701/ei200803.pdf
File-Format: application/pdf
Series: RePEc:ems:eureir
Number: EI 2008-03
Keywords: crew scheduling, large-scale optimization, partitioning methods
Handle: RePEc:ems:eureir:11701