Dispatch Optimization of Electric Autonomous Modular Vehicles for Public Transport

Public transit systems based on modular vehicles (MVs) have scalable transportation capacity, thus meeting complex and changeable travel demands. Therefore, MVs have been widely studied in the field of public transport. To understand the effects of capital budget, operation mode, and charging factors on operation cost and service level, a macro-network dispatching model for public transport EVs was developed. Considering the direct transit network to be the modeling object, according to the operation mode of the MVs, the model was divided into two specific mathematical forms: ① The intelligent connected mode, where each MV is equipped with a power system, several MVs form a fleet, and the rear vehicle replicates the behavior of the front vehicle. ② The traction mode, where the fleet in the system includes power and van-type MVs, which should be towed by power MVs. The key point of the model is to use the vehicle-hour conservation constraint to express the mathematical relationship between fleet size, capital budget, and traffic flow. The dispatching models under the two operation modes were formulated as mixed-integer nonlinear programming problems, in which the objective function contains nonlinear terms, and the constraints are linear terms. The nonlinear term in the objective function was transformed into a linear form by introducing auxiliary variables and linear inequalities to improve efficiency. Using the Beijing transit sub-network as an example, numerical experiments under different influencing factors were conducted. The results indicate that the traction mode is more effective than the intelligent network mode in reducing vehicle operation cost, and the intelligent network mode is more effective than the traction mode in improving passenger service level. The running vehicle cost in the intelligent network connection mode is 114% higher than in the traction mode, and the passenger waiting cost in the traction mode is 138% higher than in the intelligent network connection mode. The proposed scheduling model and sensitivity analysis framework are universal and portable, assisting decision makers in better understanding MV-based public transit systems. © 2022, Editorial Department of China Journal of Highway and Transport. All right reserved.