Model predictive control of bus platoons in a circular route considering dispatching, holding and boarding limits

Abstract

Public transport forms the backbone of the city's operation. Proper planning and investment of public transport can create additional jobs to revitalize and recover cities from covid-19. In this paper, we propose a combined dispatching-operation bus model predictive control strategy, where a rolling horizon mechanism is adopted to control the bus system in a real-time manner. Either a bus platoon or a single bus is allowed to be dispatched in each trip, and bus re-dispatching is captured in the system to realistically reflect the real-world. Also, the additional bus initial constraints allow control to be applied at any time when buses are either driving on the road or loading at the stop. Model complexity is investigated by solving the optimization problem under various prediction horizons, number of buses and bus stops. Furthermore, the comparison experiment with a high-frequency fixed dispatching method is performed on the Singapore bus line 179A developed in SUMO simulator to illustrate the effectiveness of the proposed method.