Research On Spatiotemporal Characteristics And Order Dispatch Optimization Of Ridesplitting Service

As an innovative transportation mode,the ridesplitting service,i.e,shared ride-hailing service,has deeply practiced the concept of sharing economy,it is in line with the macro trend of low carbon and carbon reduction,and it is also an effective way to mitigate the potential negative impact of the solo ride-hailing service.Encouraging shared over solo ride-hailing trips could offer a win-win-win proposition for transportation network companies(TNCs),ride-hailing riders,and cities.For example,increasing the order response rate and ridesplitting matching success rate during rush hours,improving the operation efficiency of the platform,and alleviating various urban traffic and environmental pollution problems,etc.However,the development of the ridesplitting service is facing following problems,such as the low ridesplitting adoption rate(RAR,the proportion of ride-hailing trips with shared trip authorized,which reflects the willingness of ride-hailing riders to share rides),the poor travel experience of passengers,the lack of application scenarios,and the need to further improve and optimize the order dispatch algorithm.In order to give full play to the role of the ridesplitting system,this dissertation focuses on the research on spatiotemporal characteristics and order dispatch optimization of the ridesplitting service.The research is carried out in two aspects.First,the impact mechanism of RAR of the ridesplitting service is explored,which is the primary core issue for the successful operation of this service.Second,order dispatch algorithms in different scenarios are optimized for this service,which is the key technical support to promote the innovation and upgrading and improve the efficiency of operation management of this service.Specifically,the research contents include the following aspects:(1)Analyze the development history and operation characteristics of the ridesplitting serviceThe development histories of the ridesplitting service and the ride-hailing service at home and abroad are summaried,the characteristics of different ridesplitting modes are elaborated,and the operation processes of real-time ridesplitting system and reservation ridesplitting system are descirbed.Based on these,the necessity,the feasibility,and the problems and challenges when developing the ridesplitting are pointed out.(2)Explore the spatial and temporal variations of RAR and its influential factorsCombined with the trip data of TNCs,various built environment and socio-economic data,multicollinearity analysis and ordinary least square regression(OLS)model are carried out and established to preliminarily filter variables,which provides the basis for variable statistical test for the establishment of influential factor model.Based on these,geographically weighted regression(GWR)models are established to examine the factors that affect the spatial variations of RAR during different time periods,such as weekday,weekend,weekday morning peak and evening peak.The research evaluates the advantages of GWR over OLS in model fitting,and analyzes the model parameters to reveal the spatial heterogeneity of the impacts of various factors on the RAR.Based on the research results,the refined policy suggestions are proposed to promote shared ride-hailing trips from the perspectives of policy makers and TNCs.(3)Optimize ridesplitting order dispatch problem with midway stopsFirstly,the ridesplitting order dispatch problem with midway stops and the ridesplitting scenes are analyzed.Then,considering two coexisting matching forms: pre-pool-matching and en-route matching,this problem is formulated as an integer programming.Specifically,the constraints of order allocation,vehicle matching,vehicle capacity,waiting time,distance savings,detour time,cumulative boarding and alighting times are considered,and the objectives are to maximize the distance savings of ridesplitting travel,and minimize the waiting time of passengers and the number of unmatched passengers.An exact algorithm based on bipartite graph,and a two-stage method(order grouping and order dispatching)based on Kuhn-Munkres algorithm are designed.Finally,based on the actual ride-hailing trip data,the performances of these two methods are compared and analyzed,and the benefits of ridesplitting matching compared with solo matching,and the impacts of the characteristics of midway stops on the order dispatch results are discussed.(4)Optimize the order dispatch problem of the long-distance and reservation ridesplittingConsidering the characteristics of the long-distance and reservation ridesplitting(also called intercity or interregion ridesplitting),for example,the starting and ending positions are located at two operation areas,passengers only have the departure time window restrict,and drivers have the maximum driving duration limit,the corresponding integer linear programming model is established.According to the problem characteristics,the corresponding variable reduction strategies are proposed to reduce the feasible region of the problem.A solution framework which combines Lagrangian relaxation(LR)and alternating direction method of multipliers(ADMM)is constructed.Specifically,the LR problem is constructed by relaxing the coupling constraint of each passenger is visited once in the original problem.Based on this,the ALR problem is constructed by adding a quadratic penalty term.The ALR problem is decomposed into a series of interdependent vehicle path subproblems by using linearization technique and block coordinate descent method.The upper bound of the model is generated by the solution of ALR problem and the greedy algorithm.The lower bound is given by solving LR problem.The solution quality is evaluated by the upper and lower bounds,and the Lagrange multiplier is updated iteratively.Finally,the feasibility and effectiveness of the proposed algorithm are verified by experiments.The research on spatiotemporal characteristics and order dispatch optimization of the ridesplitting service is an important means to improve the attraction and the operation efficiency of the ridesplitting system.The research results can not only help urban planners and TNCs to formulate refined incentive strategies and measures to promote shared ride-hailing trips,they can also provide theoretical support for the development of new business type and models of the ride-hailing service,promote the innovation and upgrading of ridesplitting products,accelerate the promotion and implementation of new ridesplitting scenarios,and promote the sustainable and healthy development of the ride-hailing industry.