SPAP: Simultaneous Demand Prediction and Planning for Electric Vehicle Chargers in a New City

For a new city that is committed to promoting Electric Vehicles (EVs), it significant plan the public charging infrastructure where demands are high. However, difficult predict before actual deployment of EV chargers for lack operational data, resulting in deadlock. A direct idea leverage urban transfer learning paradigm learn knowledge from source city, then exploit demands, and meanwhile determine locations amounts slow/fast stations target city. demand prediction charger planning depend on each other, required re-train model eliminate negative between cities varied plan, leading unacceptable time complexity. To this end, we design an effective solution S imultaneous Demand P rediction nd lanning ( SPAP ): discriminative features extracted multi-source fed into Attention-based Spatial-Temporal City Domain Adaptation Network AST-CDAN ) cross-city prediction; novel Transfer Iterative Optimization TIO algorithm designed by iteratively utilizing fine-tuning algorithm. Extensive experiments real-world datasets collected three China validate effectiveness efficiency . Specially, improves at most 72.5% revenue compared with deployment.