Research On The Matching Model Of Ridesharing And Subway Intermodal Transit Towards Taxi Users

Building a public transport system with subway as the backbone,regular bus as the basis and taxi as the supplement is an important measure to alleviate traffic pollution and congestion in big cities,and developing multimodal transit service is an effective means to improve the operation efficiency and overall efficiency of public transport system.With the application of mobile Internet and the popularity of sharing economy,ridesharing services have attracted wide attention and it also provides a new service mode for the traditional public transport multimodal transport.Ridesharing and subway intermodal transit is also considered to be a greener and more efficient travel mode,whose advantages of sustainability and large-scale development are more significant,especially for car users.However,the research and application of this mode is still in the exploratory stage,and there are still many limitations in data mining,mode definition,operation management and other aspects.Moreover,a mature and effective passenger matching mechanism for this mode has not been established.In order to solve above problems,through mining the travel big data of cruising taxi and online ride-hailing,this paper analyzed the co-opetition relationship between taxi and subway and the passenger travel time value.Then we defined the travel service modes of the ridesplitting and subway intermodal as well as the vanpooling and subway intermodal transit for different types of taxi passengers,respectively.According to the service characteristics of these two travel modes including on-demand response and integrated intermodal etc.,the optimization models of passenger matching were constructed respectively.Finally,the case application analysis on these two intermodal transit service modes were conducted for Beijing taxi users.From methods and technologies to practical application,this paper conducted a comprehensive and in-depth study on the ridesharing and subway intermodal transit mode,which could provide beneficial support for the scheme design and promotion of this mode.The main work are as follows:(1)We made a quantitative analysis on the co-opetition relationship between taxi and subway from both macro and micro levels based on the large-scale actual travel data.From the macro level,this paper first compared the travel statistics characteristics of taxi and subway passengers and described the spatio-temporal travel characteristics of competitive and cooperative taxi passengers to understand their performance of competition and cooperation relationship.From the micro level,the cooperation intensity index of subway station level and the competition intensity index of OD level between subway stations were constructed to quantify the degree of competition and cooperation.(2)The temporal and spatial travel demand matrix of taxi passengers was established,and the basic travel patterns of competitive and cooperative taxi passengers was mined based on the Nonnegative Matrix Factorization(NMF)to estimate the travel time value of passengers.Based on stochastic utility theory,combined with NMF and Logit model,the NMF-Logit algorithm was constructed to calibrate the equation of passengers’ travel choice between taxi and subway,and then calculate their value of travel time.Taking Beijing as an example,the travel time value parameters of cruising taxi passengers and online ride-hailing passengers were estimated respectively,laying a foundation for subsequent research.(3)For competitive taxi passengers,a door-to-door travel service model of ridesplitting and subway intermodal transit was proposed,and a two-stage optimization matching model based on passengers’ choices was constructed.In the first stage,according to the integrated intermodal characteristics,the generalized cost of ridesplitting passengers is established by comprehensively considering the travel time value,the subway congestion penalty,the order of shared passengers and the other factors for trip stages of subway,ridesharing and the intermodal transit.Based on the principle of maximum utility and the core idea of substituting route segment by segment,this paper proposes a search algorithm to determine the transfer station of ridesplitting intermodal transit.In the second stage of the model,the feasible matching set was established based on the principle of trip similarity and trip cost saving,and the matching problem of ridesplitting and subway intermodal transit was formulated as a 0-1 integer programming equation.Taking Beijing as a case study,the optimal matching scheme of ridesplitting and subway intermodal transit was analyzed for the competition taxi passengers during the morning peak hours.Lastly,sensitivity experiments were designed to analyze the overall performance of matching model under the influence of different policies.(4)For cooperative taxi passengers,the travel service mode of direct online vanpooling and subway intermodal transit was proposed,and the matching optimization model based on the interaction between passenger choice and vehicle operation was constructed.Firstly,based on the characteristics of demand response and direct line,a feasible matching search method based on the interaction mechanism between passenger choice and vehicle operation is proposed.The feasible passenger matching set and bus operation information consistent with the optimization objective were determined by the interactive iteration among the generalized passenger travel cost,the minimum capacity of online canpooling,the bus stop location and departure time.On this basis,method to merge bus routes for passenger twice ridesharing was proposed based on the principle of trip similarity and generalized cost saving.Then the 0-1 integer programming model was used to express the matching problem of online vanpooling and subway intermodal transit based on the feasible matching set.A k Higher Passenger Capacity Groups(k-HPCG)heuristic algorithm was established to solve the problem of large-scale combinatorial matching.Taking Beijing as an example,the optimal matching scheme was analyzed for cooperative taxi passengers during the morning peak hours,and the sensitivity experiments were designed from the perspectives of passenger attitudes and bus operation strategies to analyze the overall matching performance of this intermodal transit mode.Based on large-scale actual travel data and starting from the competition and cooperation relationship between taxi and subway,this paper established a complete research chain for the novel travel service mode of ridesharing and subway intermodal transit,which includes the user feature mining,travel service mode definition,matching model algorithm construction,and practical case application.The proposed intermodal integration mechanism and a series of innovative matching technologies and methods have certain theoretical value and practical significance for promoting the integrated services of ridesharing and public transport.