The bus to depot allocation problem revisited: a genetic algorithm

Transit systems should provide attractive yet low cost services to passengers. However, reduction of operating costs can significantly affect performance of bus systems. Cutbacks in deadhead mileage and depot costs becomes attractive since it leads to reducing operating expenses without interfering with passenger services. We revisit previous work on the bus-to-depot allocation problem; we propose an improved model for assigning buses to depots and employ a novel hybrid genetic algorithm for solving the problem. The model and algorithm are applied to the Athens bus system. Results indicate gains in operating costs compared to the existing allocation and improved computational performance.