Generation and Optimization of Travel Plan Based on Multi-Mode Travel Time Prediction

To provide a multi-mode collaborative travel plan customized for travelers, a preliminary travel plan is created once the transportation network model is constructed and the traffic data in the road network is incorporated. Then the long-term and short-term memory recurrent neural network model (LSTM) is used for modeling. The travel time of vehicles in various travel plans is predicted, the travel time of the alternatives is compared, and the travel plan is optimized by minimizing the travel time. Using the Beilin District Xi’an as an example, the multi-mode coordinated travel plan with other travel plans is compared. It is concluded that the multi-mode coordinated travel plan not only takes less travel time but also makes better utilization of the resources of transportation system such as car-hailing service and public transportation. Therefore, the multi-mode cooperative travel plan has a positive role in improving traffic utilization and travel time.