Metaheuristic algorithms for the bus driver scheduling problem with complex break constraints

Abstract

The Bus Driver Scheduling Problem (BDSP) is a combinatorial optimisation problem that consists of assigning bus drivers to vehicles with predetermined routes. The objective is to optimise the employees' operating costs and work quality based on goals like the number of vehicle changes. This problem is highly constrained due to the complex rules specified by a collective agreement and law. Hence, solving real-life instances with exact methods in a reasonable time is very challenging. In this work, we investigate and compare metaheuristics based on Tabu Search and Iterated Local Search for solving this problem. We analyse the impact of different solution components, including neighbourhoods, acceptance criteria, tabu lists, and perturbation moves. Further, we provide a new set of large real-life-based instances that extends the existing benchmark. We compare our methods with the state-of-the-art approaches on the extended set of instances and show that our algorithms provide very good solutions for large instances.