Risk-Limiting Multi-Station EV Charging Scheduling with Imperfect Prediction

EV charging scheduling is a fundamental problem for charging infrastructure management as EVs impose both challenges and opportunities on the power grid. The primary difficulty for EV charging scheduling comes from the uncertainty associated with the arriving EVs, including their arrival time, departure time, and charging volume. This makes the scheduling in a single EV charging station already challenging. Clearly, the joint scheduling across different charging stations is even more challenging. To this end, in this paper, we first develop a risk-limiting charging scheduling approach based on chance-constrained optimization in the offline setting. To enable online charging scheduling, we propose a model predictive control (MPC) approach, which could effectively reduce the impact of forecasting errors. We numerically demonstrate that our approach can achieve near-optimal performance. Specifically, we empirically evaluate the influence of prediction error on the scheduling cost, which indicates the value of information in EV charging scheduling.