Research On The Dynamic Taxipool Matching Problem

With economy’s rapid development in China, the speeding up of urbanization, vehicle ownership increases rapidly, the urban traffic congestion, environmental pollution and energy consumption problem is increasingly serious. Ride-sharing has generated much interest, as it can improve road carrying capacity and reduce traffic congestion.And, taxipooling is about efficiently allocating and using of existing resources, so as to improve the taxi capacity, relieve the contradictions of "difficult to take a taxi". Meanwhile, the benefit of passengers and drivers should be considered. Both good transportation services for passengers and considerable income of the driver should be guaranteed. Therefore, this thesis studies the dynamic taxipool matching problem based on 973 Program (2012CB725402).Based on the analysis of modalities and characteristics of ridesharing system and the studies of intelligent dispatching system, this dissertation develops a dynamic taxipooling matching model and insert algorithm that is aimed at identifying suitable matches between passengers requesting taxipool services with appropriate drivers. This model minimizes the overall travel time and taxi fees of the passengers subject to driver and passengers time window, taxi capacity and taxi availability constrains, considering the the benefit of passengers and drivers.The insert algorithm successfully works for inserting the real-time requests into vehicle routings, and computing the taxi price rate of every passenger. Then the dynamic taxipool matching system is established to analyze the proposed model and algorithm behaviors on a virtual road network by using matlab software. Base on those simulation results, multivariate analysis and multi-parameter combination analysis is conducted on several parameters and variables to evaluate the influence on the performance of the taxipool system operation. The study shows that arrival rate of taxipooling requests, passenger’s tolerance for detour, ordering lead time, vehicle speed and the number of vehicles influences the rate of success and average saving time to varying degrees; passengers’discount rate and drivers’ profit coefficient has the significant impact on the benefit of driver and passengers.The research findings in this paper can be directly used for theoretical research on the dynamic taxipool matching problem.Also, they can provide the references and advices in the construction, operation and management of practical dynamic taxipooling system.It could not be better if the findings have a positive impact on alleviating traffic problems.