Based on a large number of charging behavior data, a charging interest model of electric vehicle users is established. The best charging options which users are interested in but have not found are recommended and charging behavior is orderly guided. In this paper, a recommended model based on a collaborative filtering algorithm is proposed for electric vehicle charging, and the best model parameter index is obtained through test and evaluation. New users with 10 times or less may use the user-based collaborative filtering algorithm while old users who have charged more than 10 times may adopt the collaborative filtering algorithm based on items. The optimal neighbor and recommended list length is 3 for the user-based collaborative filtering algorithm, and 4 for the algorithm based on items. The paper points out that the load aggregator can re-optimize the recommendation list in combination with the demand response plan and filter out the recommended information that conflicts with the demand response to realize the orderly charging control.

ordered charging  /  electric vehicles  /  collaborative filtering  /  intelligent recommend