Ant colony optimization for railway driver crew scheduling: from modeling to implementation

Abstract

This study addresses the crew-scheduling problems for railway drivers’ duty trips on a railway timetable represented as a time–space diagram. Based on the diagram, the railway driver-scheduling problem is then transformed into an arc routing problem (ARP). Because of the special properties and features of the problem, the ARP can be treated as a typical vehicle node routing problem. The Ant Colony Optimization algorithm is employed to solve the transformed problem. Real data from the Taiwan Railways Administration are used to test the proposed models and algorithm. The results showed that the dead-heading-allowed approach is able to obtain a better solution in terms of fewer drivers and shorter idle time.