This paper presents an approach to electric city bus scheduling optimization, which results in a Pareto frontier in two conflicting criteria being minimized: the number of buses required to serve predetermined routes and the excess of distance travelled (so-called deadhead distance). These criteria reflect the city bus fleet investment and operational costs, respectively. The sequential optimization strategy is executed in two phases: 1) finding the minimal number of buses, and 2) gradually incrementing the number of buses from the minimal one and minimizing the deadhead distance. Two optimization methods are proposed: mixed integer linear programming and genetic algorithm, where the former provides the optimal solution but it is limited to small-scale problems (fleets), while the latter can deal with large fleets but generally results in a nearly optimal solution. The optimization approach is demonstrated on a custom-generated dataset reflecting characteristics of real-world city bus transport systems.

[OPEN ACCESS ]