Research The Optimization Of Intercity Ridesourcing Scheduling Based On The Prediction Of Travel Demand

With the popularity of smart phones and the development of China ’ s urban agglomeration,residents’ intercity travel demand has increased rapidly,and intercity ridesourcing services have gradually risen.On the one hand,the intercity ridesourcing service including ridesplitting provides residents with more convenient and affordable intercity travel choices;on the other hand,the increasing intercity travel demand and long-distance driving distance have exacerbated the imbalance between supply and demand of intercity ridesourcing,affected the waiting time of passengers,and caused the decline in the efficiency of ridesourcing service.In this context,this paper comprehensively uses data analysis technology,and on the basis of travel demand prediction by optimization method,studies the scheduling optimization of inter-city online ride-hailing.The specific research contents and results are as follows.(1)Firstly,the characteristics,behavior patterns and influencing factors of intercity ridesourcing are explored.The actual observation data of intercity ridesourcing is preprocessed and described,and the spatial and temporal characteristics of intercity ridesourcing are analyzed by using the actual observation data.It is found that the characteristics of intercity ridesourcing are obviously different from those in the city.Then,based on the geographically weighted regression model,the time,space and other factors that affect the intercity ridesourcing travel are studied.(2)Secondly,forecast the demand of intercity network ridesourcing.After data preprocessing and feature construction,a ST-transformer(Spatiotemporal Transformer)prediction model with rich self-attention mechanism was established.The model stacked the time module and space module in a framework to extract spatio-temporal features simultaneously.Finally,the prediction effect of the model is verified on the actual data set,and compared with five classical models,the method of intercity online car hailing demand is optimized.(3)Thirdly,the influencing factors and prediction of intercity ridesplitting are studied.The ridesplitting recognition algorithm is used to identify the online ridesplitting trips in the actual data set,and then combined with the local built environment data,the potential influencing factors of spatio-temporal characteristics of intercity ridesplitting demand are analyzed in depth.Based on the potential factors,the binary logistic regression model is used to study and determine the key factors affecting the intercity ridesplitting.On this basis,a spatio-temporal geographically weighted regression prediction model is constructed to explore the spatiotemporal nonstationarity of intercity ridesplitting,and the travel demand of intercity ridesplitting is predicted with the key influencing factors as input.(4)Finally,based on demand forecast,the scheduling optimization of intercity ridesourcing is carried out.Firstly,the topological graph of the road network structure in the study area is constructed.Then,based on the results of the intercity car-hailing demand forecasting model,an intercity ridesourcing scheduling optimization model is proposed with the minimum passenger waiting time,the shortest driver scheduling distance and the highest driver income as the objectives.On the basis of this model,the matching process of ridesplitting is added to construct the optimization model of ridesplitting scheduling in intercity Internet.Finally,the model is used to optimize the vehicle scheduling.The results show that the optimization model reduces the waiting time of passengers,mproves the platform revenue and the success rate of ridesplitting matching,and verifies the effectiveness of the method to improve the operation efficiency of intercity ridesourcing.