Research On Shared Vehicle Relocation Based On Travel Trajectory Of Users

With the acceleration of urbanization and the improvement of people's living standard,the number of cars has increased dramatically.That leads to various problems such as shortage of parking spaces,traffic congestion and low travel efficiency in major cities.For this reason,the private sector restricted purchase policy has been promulgated in various areas,which has eased the current situation,but worsened the contradiction between supply and demand,resulting in the dissatisfaction of urban car demand.In this case,the sharing vehicle came into being.As a powerful supplement to public transportation and a representative product of sharing economy,car sharing has characteristics of convenient and efficient,environmental protection,green travel,and alleviate the contradiction between supply and demand in a certain extent,solving the vehicle idle for a long time,the road overload problem.But there is still a range of problems in this area,including imbalanced distribution and coexist phenomenon of car shortage and seats overage.In order to solve the existing problems and achieve the full sharing of shared cars,this paper studies on a shared vehicle scheduling problem based on users trajectory,the major work is as following:(1)The mode of sharing automobile service.Proposed set of "ride" and "relay car" or "carpool" as one of the user-participated car sharing service scheduling model for users on travel routes between various relations,designed an operation mode including achievements and reward system according to users' psychology.(2)Sharing car demand.On the basis of user travel path,combined with different travel paths city prior knowledge,using BP neural network recognition of car travel trajectory,dig out the potential demand for car sharing,clustering sharing car demands based on trajectory spatiotemporal similarity,in order to obtain temporal and spatial distribution of demand.(3)Sharing vehicle scheduling.In user type car sharing service mode,based on the clustering results of users travel trajectory recognition and car sharing demand,considering the user satisfaction and aiming to minimize the total cost of the platform,set the user scheduling priority scheduling as the main constraints and establish the user-participated sharing vehicle scheduling model,then use the improved genetic algorithm to solve the model.(4)Empirical analysis.Using GeoLife travel trajectory of Microsoft Asia Research Institute open dataset and tensorflow Python language learning framework to realize the depth of user travel trajectory recognition based on MATLAB software genetic algorithm code,for the purpose of solving the model of user-participated sharing vehicle scheduling to acquire the optimal vehicle routing and the total cost,proof the validity of this model and algorithm through the method of comparison of experimental and independent scheduling.Through the study of user-participated car sharing service mode based on vehicle scheduling method of sharing user travel path,the problems of empty seat,the high cost of the vehicle scheduling platform,imbalance of supply and demand can be solved to a certain extent,the empirical results show that the proposed user-participated car sharing scheduling method can effectively reduce the total cost of the platform,improve vehicle utilization rate and provide a new idea for the study of automobile sharing service model and vehicle scheduling method.