Integrated Optimization of Electric Vehicles Charging Location and Allocation for Valet Charging Service
Xiaoxiao Shen,Jun Lv,Shichang Du,Yafei Deng,Molin Liu,Yulu Zhou
DOI: https://doi.org/10.1007/s10696-023-09508-8
IF: 2.2092
2024-01-01
Flexible Services and Manufacturing Journal
Abstract:Since electric vehicles (EVs) have definite benefits over gasoline vehicles, the vehicle market could be dominated by EVs in the future. This paper focuses on the new valet charging service to send staffs to replace users for charging their EVs, which can largely reduce charging anxiety. In this study, the location of charging stations and the allocation of charging demands to charging stations are optimized simultaneously due to the interaction of these decisions. The queueing behavior at the charging station is incorporated into the model, and the average charging waiting time is derived. We construct a mixed integer nonlinear optimization model based on the characteristics of valet charge service and an infinite-source queuing model. The objective is to minimize a total cost of the construction of charging facilities and valet charge service launching (i.e., charging staffs’ road service, round-trip time, and the charging waiting time). The planning problem for valet charging service in this paper contributes to the existing literature on self-charging way where the users of EVs drive at charging stations to recharge their EVs by themselves. An improved genetic algorithm is developed to obtain deployment and operation plans for large-scale instances constructed based a real case in Shanghai. The improved genetic algorithm shows high performance in convergence and solution quality, which provide the service providers an efficient decision support tool. Meaningful managerial insights are also provided, which can help the service provider make better cost-effective design of charging location and allocation plans. For example, the charging station location decisions are not as much as sensitive to critical variables (such as demand level, charging capacities, and the value of time) than the overall cost to those. This means that partial location decisions remain unchanged when the key parameters vary.