A multi-agent framework for scheduling fully electrified public transport

This paper describes a framework for scheduling electric buses with the aim of minimizing their overall operational costs as well as the required fleet. The scheduling problem of electric buses comes with several real-time constraints such as the location of the charging stations, number of time-tabled trips, the distance between the depots, etc. The main objective is to minimize the running costs of the electric buses on the road network while satisfying the underlying constraints. The cooperative multi-agent model segregates the scheduling problem into a master agent and a bus agent through asymmetric communication, and this decentralized approach allows the scheduling to happen in a real-time scenario. The proposed scheme is tested on the public transport network of the Maltese Islands.