A Genetic Algorithm Approach to the Balanced Bus Crew Rostering Problem

Abstract

This paper addresses the problem of finding an optimal assignment of duties to bus crews in a given time horizon, in such a way that the total workload should be evenly distributed. This problem is known as the balanced bus crew rostering problem. We first formulate it as an equivalent multi-level balanced assignment problem. Then, a genetic algorithm (GA) approach is proposed to solve this problem. A numerical example is used to illustrate the application of the approach. Finally, we compare the implementation results generated by the GA approach with the results generated by an algorithm proposed by Carraresi and Gallo (1984) and another heuristic algorithm proposed by Ceder (2007). The results indicate that the GA approach has a better computational performance, and can generate more balanced bus crew rosters as well as fewer rosters. This demonstrates that the GA approach is a good alternative to the balanced bus crew rostering problem. To better understand the performance of the algorithm, sensitivity analyses of relevant parameters are presented at the end of the paper. 