Data-driven planning for wireless charging lanes

Abstract

The widespread adoption of electric vehicles (EVs) is significantly hindered by the long charging time and range anxiety resulting from slow charging speed and limited battery capacity. Meanwhile, wireless charging lanes (WCLs) and solarpowered EVs (SEVs) offer a promising solution by providing wireless and solar charging power while driving. Under this circumstance, addressing the optimal planning of WCLs while considering SEV operations is crucial for facilitating the widespread adoption of EVs. Considering the uncertain solar power harvesting of SEVs, we propose a data-driven twostage distributionally robust optimization (DRO) model for this integrated planning and operation problem. In the first stage, we optimize the deployment of WCLs with budget constraints, and the second stage determines the optimal operation schedules of SEVs under uncertain solar charging power characterized by a moment-based ambiguity set. To address the computational challenges (due to the discrete variables in both stages and the infinite-dimensional optimization in the second stage), we develop two approximation models and an integrated distributed method. Finally, extensive numerical experiments with synthetic and real transportation networks are conducted to demonstrate the effectiveness and scalability of our proposed models and algorithms. Specifically, the proposed DRO model achieves a 1.17 lower total cost in out-of-sample tests than the sample average approximation method, and with higher wireless charging power rates and increased battery capacities, we can build fewer WCLs.