A vehicle scheduler for on-demand bus systems based on a heuristic cost estimation

Abstract

This paper proposes a scheduling scheme for solving the minimum cost single vehicle routing problem with designated deadlines, based on the dynamic programming and heuristic cost estimation. We conducted several experiments to evaluate the efficiency of the proposed scheme, and found that in more than 76% of randomly generated instances, it outputs an optimal solution within few seconds provided that the number of passengers assigned to the vehicle is at most 10. In addition, even for larger instances, it can output a better quasi-optimal solution much more quickly compared with conventional schemes.