Software-in-the-Loop Planning of Fast Charging Station based on GNN-Accelerated Ordinal Optimization

The allocation of fast charging stations is vital in promoting the regular travel of electric vehicle. Considering the impact of road network on charging service, we propose a software-in-the-loop planning method to simulate the complex traffic condition. A SUMO-based fast charging station planning model is established, while user experience and charging station's operation economy are both considered. To reduce the optimization time, we apply ordinal optimization theory, and establish a crude model to simplify the calculation. Moreover, based on GNN's capability of learning from networks, we propose a GNN-accelerated ordinal optimization method for fast charging station planning, where GNN is deployed to perform end-to-end fitting for crude evaluation results. Finally, a numeric example is employed to demonstrate the feasibility of proposed method and the effect of GNN.