Integrated design of electric vehicle sharing networks under uncertain trip duration and energy consumption - Laboratoire Interdisciplinaire des Sciences du Numérique Access content directly
Conference Papers Year : 2023

Integrated design of electric vehicle sharing networks under uncertain trip duration and energy consumption


Carsharing is a concept that allows individuals to borrow cars on a short-term basis from a designated area, for a rental rate charged by the time or the distance driven. Station-based carsharing systems use a network of stations where shared cars should be picked-up and dropped off by customers or remain parked when there is no demand. Several existing carsharing programs have integrated electric vehicles (EVs) to help reduce greenhouse gas emissions and air pollution from road transportation. Building a successful EV sharing program requires an adequate planning of the system design and operations to ensure both customer satisfaction and economic viability of the project. In this work, we focus on the strategic design decisions of EV sharing networks, namely the number of parking slots and chargers to be built in each station, as well as the size of the vehicle fleet to be deployed in the network. To be practically relevant, the problem modelling should take into account the strong interactions between the strategic, tactical and operational planning levels [1]. At the tactical level, fleet deployment in the sharing network at the beginning of the day should be determined to ensure that enough vehicles are placed where demand is likely to arise. At the operational level, vehicle stock imbalance represents one of the main challenges. Another operational challenge is to manage EV recharging whenever the battery state of charge (SOC) becomes low, typically between two customer trips. Since tactical and operational decisions affect the spatiotemporal availability of vehicles and significantly contribute to the overall performance of the system, they should be considered when making strategic decisions [2]. Therefore, we study an integrated optimization model, seeking to simultaneously optimize strategic, tactical, and operational decisions. Since there is only a limited number of papers investigating integrated models for EV sharing networks, this aspect represents a first contribution of our work. Another contribution relates to the incorporation into the mathematical model of uncertainties that may affect the ultimate objective of matching vehicle supply with trip demand. Only a few papers relevant to integrated EV sharing network design have considered the stochasticity of the problem parameters and to the best of our knowledge, all of these models have focused on trip demand uncertainty only [3,4]. However, in practice, the uncertainty of trip duration and energy consumption may affect vehicle availability in the right area and the right time with the right battery SOC. Namely, even if a driver retrieving a shared EV from a station commits to return it at the destination station within a given timeframe, he may be delayed and return it later than promised. Electricity consumption on a trip may also be affected by the driver behaviour during the trip or by traffic conditions. We thus intend in this work to assess the impact of these uncertainties on the design of an EV sharing network, while taking into consideration the operational issues of car relocation and recharging. Using randomly generated instances, we study the sensitivity of the MILP model when trip duration and energy consumption vary according to three different scenarios, optimistic, pessimistic and average ones. Then, we show through the value of stochastic solution (VSS) based on a scenario-based two-stage stochastic program, the significance of incorporating the considered uncertainties into the model. The combination of the integrated and stochastic aspects of the problem results in a model that becomes computationally intractable for realistic, large-size instances, due to the huge number of variables and constraints involved in the formulation. Therefore, one of the main future perspectives of this work consists in investigating more efficient solution approaches, namely decomposition techniques and metaheuristic methods. [1] Golalikhani, M., Oliveira, B. B., Carravilla, M. A., Oliveira, J. F., Antunes, A. P. 2021. Carsharing: A review of academic literature and business practices toward an integrated decision-support framework, Transportation Research Part E, pp 102280. [2]Bekli, S., Boyacı, B., Zografos, K. G. 2021. Enhancing the performance of one-way electric carsharing systems through the optimum deployment of fast chargers, Transportation Research Part B, 152, pp. 118-139. [3]Brandstätter, G., Kahr, M., Leitner, M. 2017. Determining optimal locations for charging stations of electric car-sharing systems under stochastic demand, Transportation Research Part B, 104, pp. 17-35. [4]Boyacı, B., Zografos, K.G., Geroliminis, N. 2015. An optimization framework for the development of efficient one-way car-sharing systems, European Joural Of Operational Research, 240, pp. 718-733, 2015.
No file

Dates and versions

hal-04390779 , version 1 (12-01-2024)


  • HAL Id : hal-04390779 , version 1


Mouna Kchaou Boujelben, Céline Gicquel. Integrated design of electric vehicle sharing networks under uncertain trip duration and energy consumption. CIE50 - 50th International Conference on Computers and Industrial Engineering, Oct 2023, Sharjah, United Arab Emirates. ⟨hal-04390779⟩
32 View
0 Download


Gmail Facebook X LinkedIn More