Event Details

PhD student Seyma Bekli will present a seminar to the Management Science Department. This is the first of a two seminar session.

Abstract: Car-sharing is an urban transportation concept that allows users to rent vehicles for short periods. Vehicles in a designated area or multiple parking spots are reserved and used by the users of the system. In one-way car-sharing systems, the users do not necessarily return the vehicle to the origin station at the end of their trips. Due to the environmental benefits, electric vehicle fleets are gaining more attention among car-sharing companies.

The two most important challenges that electric one-way car-sharing systems (EOWCS) experience are the imbalanced distribution of the vehicles and the occasional need for charging. The former problem is addressed by relocations. A fleet of personnel relocate vehicles across stations to serve as many users as possible. As it comes to charging, it takes more than six hours to fully charge the vehicles with the chargers mainly used by EOWCS companies. With the new faster chargers, this duration could be decreased down to 30 minutes.

In this study, we introduce an integer programming (IP) model for optimal implementation of chargers with different charging speeds while considering the relocation activities.Unfortunately, the exact model becomes intractable for large instances. To address this issue, we use heuristic algorithms that are based on relocation reduction and grouping approaches. The former approach removes the unlikely relocation options from the model whereas the latter treats the stations as groups and make charger acquisition decision in two stages. Both approaches decrease the number of variables and eventually make the models tractable without compromising the quality of the solution.

We tested the proposed models and algorithms on real-life instances taken from an operating EOWCS company based in Nice, France. The preliminary results suggest that heuristic algorithms give near-optimal solutions in reasonable computational time.