Energy management modeling method based on bidirectional memory network

The invention discloses an energy management modeling method based on a bidirectional memory network, which predicts the vehicle speed at a future moment through a bidirectional memory network vehicle speed prediction model, and finally performs modeling on energy management in combination with related constraints. The method comprises the following steps of (1) obtaining vehicle speed data through a traffic information open source data set of a specific road condition, and performing normalization processing on the data to obtain a vehicle speed-time sequence; (2) on the basis of a long short-term memory network (LSTM) vehicle speed prediction model, establishing a bidirectional long short-term neural memory network (BILSTM) vehicle speed prediction model; (3) performing energy management modeling based on the predicted vehicle speed of the BILSTM and related constraints of vehicle kinematics; according to the method, vehicle speed prediction is carried out through the bidirectional long-short-term memory ne