Simulation-optimization for station capacities, fleet size, and trip pricing of one-way electric carsharing systems

This study presents an event-driven discrete-event simulation (EDDES) approach based on an O2DES framework for a one-way electric carsharing system (OECS). The developed model mainly consists of three modules: station, electric vehicle (EV), and link. The system carefully considers the impact of road congestion on travel speed and designs a detailed charging process for EVs to approximate the real world. Based on the high speed of EDDES, a simulation-optimization framework that jointly determines station capacities, fleet size, and trip pricing to maximize the net revenue of operators is proposed. A simultaneous perturbation stochastic approximation (SPSA) algorithm is adopted to solve this problem. A case of EVCard in Chengdu is conducted to demonstrate the efficiency of the proposed framework. Several control experiments focusing on various pricing schemes are introduced to verify the advantage of optimization results further. Moreover, a one-way gasoline carsharing system (OGCS) is optimized to contrast with optimal OECS. These comparisons reveal some interesting findings: (1) from the perspective of operators, employing dynamic pricing achieves at least a 57.03% increase in net revenue compared to fixed pricing strategies; (2) from service, the optimal configurations and trip pricing of OECS effectively avoids vehicle overstock at stations; and (3) in terms of greener production, the optimal OECS could sustain a larger fleet size and a higher degree of cleaner production, and its profit generated by per CO2 emission is over six times higher than that of the optimal OGCS.