A Mixed-Integer Program (MIP) for One-Way Multiple-Type Shared Electric Vehicles Allocation With Uncertain Demand

This paper proposes a mixed-integer program (MIP) to address a challenge issue of vehicle upgrade policy in current electric car rental market, i.e., idle luxury and high-end vehicles can be used as ordinary vehicles when the demand for ordinary vehicles is high. This model essentially is to maximize the total profit through balancing the demand and supply with a comprehensive consideration of revenue of rental operation, cost of potential demand loss, dispatching cost, energy consumption per kilometer, mileage limitation of electric vehicles (EVs), and the uncertainty of user demand for multi-type EVs. The proposed MIP model can be solved by CPLEX or Gurobi to search for the global optimal solution. Finally, real data from an electric carsharing system (ECS) with three types of EVs and 20 carsharing stations was used to validate the model. As a result, the daily increase of profit could reach 11.09% with an average of 8.08%.

I. Introduction

The electric carsharing system (ECS) has been used in many megacities as an efficient transportation mode to relieve traffic congestion, improve efficiency, and reduce emissions (e.g. [1], [2]). Many enterprises are offering carsharing service, such as Car2Go, EVCARD, UrCar and ZipCar. Based on vehicle return rules, carsharing can be either one-way or round-trip [3]. For the round-trip, the shared vehicle must be given back to the station where it was rented, which brings inconvenience to users [4]. On the other hand, one-way system has advantages in both customers travelling and corporate revenue [5]. However, the one-way carsharing system encounters the imbalance problem between the vehicle supply and customer demand among stations (i.e., over-saturation and depletion), which can lead to profit loss.