Macroscopic Dispatching Of Taxi In Multi-region Network Based On MFD

With the recent proliferation of taxi hailing app,the traditional taxi industry is facing severe challenges and it is urgent for them to search for new dispatching method to solve such problems,unreasonable resources distribution,high vacant rate and long waiting time for passengers.The popularization of acquisition equipment of traffic data and the development of traffic big data technology provide new ideas for taxi dispatching.With the perspective of traffic big data,this paper aim to present a network-scale taxi dispatch model based on MFD while considering the interrelated impact of normal traffic flows and taxi dynamics.The proposed model considers dividing city into multiple regions and a MPC approach is devised to control the movement of vacant taxi between the regions to achieve such goals,reduce vacant rate of taxi and waiting time of passengers in city.In this paper,a real urban network is taken as the research region.first,the research region is supposed to be partition into multiple sub-regions each with similar traffic conditions and fusing taxi GPS and intersection detector data to construct MFD model of each sub-region.then,building boarding model of each sub-region and estimating passenger demand by using boarding passenger and vacant taxi density count by taxi data.predicting passenger demand of taxi based on several days` data;finally,finishing building the model of taxi dispatch and numerical simulation based on the dispatching model.then,test the reliability of the model and the validity of the strategy based on MPC by using actual data of each sub-region(passenger demand,well-defined MFD etc.)as simulation`s input.The test results proved that: taxi dispatch model based on MFD can well reflect actual state and taxi dispatching strategy based on MPC can reduce vacant rate of taxi and waiting time of passengers effectively.