Research on Intelligent Scheduling Algorithm for Electric Vehicle Charging Pile Based on Deep Learning

Along with the rapid popularity of EV, the intelligent dispatching of charging pile is becoming a key problem for improving the charging efficiency and optimizing the distribution of resources. This paper proposes an intelligent scheduling algorithm for electric vehicle charging piles based on RNN, aiming to solve the limitations of traditional scheduling methods in dynamic demand environments. By designing an RNN model containing LSTM units, this paper effectively captures the time dependency of charging demand and optimizes the scheduling strategy of charging piles. Experiments are carried out to compare the performance of RNN and conventional heuristics. Experiments indicate that RNN can significantly decrease the average waiting time (about 28%) and increase the utilization ratio of charging pile (about 17%). In addition, the calculation time of the RNN model meets the real-time scheduling requirements, and the training time is about 15 minutes, which is suitable for the actual needs of large-scale charging pile networks. Through in-depth analysis of the simulation results, this paper finds that the RNN scheduling algorithm shows higher flexibility and robustness in dealing with charging demand fluctuations.