Abstract: In response to issues such as whether electric vehicles have the physical conditions and subjective willingness to participate in grid response that are not fully considered in the traditional optimal scheduling process of electric vehicle clusters,a method based on bidirectional gated recurrent unit （Bi-GRU） is proposed to identify the adjustability of electric vehicle charging behavior,and the scheduling is optimized with the objective of maximizing the revenue of aggregators.Firstly,the real data of electric vehicles in the charging station is used to describe the charging time and charging capacity characteristics of electric vehicle users.Starting from the subjective willingness factors that affect users’participation in the optimal scheduling,the charging preferences of users are analyzed and described.Secondly,the physical matrix of the charging process and the user response willingness matrix are constructed,and the Bi-GRU model is used to divide electric vehicles into adjustable electric vehicle clusters and non-adjustable electric vehicle clusters,and the charging capacities are calculated.Finally,from the perspective of maximizing the revenues of electric vehicle aggregators,a model and solution strategy for the participation in power grid optimal scheduling of electric vehicle clusters considering the impact of time of use electricity prices are proposed.The calculation case shows that compared with traditional optimal scheduling methods,the proposed Bi-GRU model can accurately identify whether electric vehicles have adjustability,and the optimal scheduling strategy can effectively suppress load fluctuations and ensure the safe and stable operation of the power grid while ensuring the maximization of aggregator revenues.