Dynamic dispatch of connected taxis for large-scale urban road networks with stochastic demands: an MFD-enabled hierarchical and cooperative approach

Heng Ding, Jiye Li, Nan Zheng, Xiaoyan Zheng, Wenjuan Huang, Haijian Bai

Research output: Contribution to journalArticleResearchpeer-review

5 Citations (Scopus)


The dynamic dispatch of taxis is an important way to improve the service efficiency of taxis and reduce cruising distance. However, in a complex urban road network with extensive and stochastic roadside service demands, current macroscopic dispatch methods can hardly establish accurate and real-time matching relationships between taxis and passengers, and microscopic dispatch methods have the disadvantage of excessive computational complexity and nonupdatable match relationships. To this end, this paper proposes a hierarchical and cooperative macroscopic and microscopic dynamic dispatching approach for real-time urban network taxis in a connected taxi (CTaxi) information environment. First, network traffic flow dynamic evolution is described based on a macroscopic fundamental diagram (MFD), and the optimal macroscopic dispatching model is proposed to optimize the distribution of CTaxis for large-scale urban road networks with multiple MFD sub-regions. Second, to avoid the phenomenon of CTaxis blindly searching for passengers, a microscopic and relationship-updatable dispatching strategy between CTaxis and passengers is established within each MFD sub-region considering stochastic service demands and the signal timing of intersections. Finally, numerical experiments are conducted in a real road network to compare the performance of four methods: no CTaxi dispatch (NCD), conventional CTaxi dispatch (CCD), MPC dispatching (MPCD) and the dispatching strategy cooperating macro and micro (DSCMM) proposed in this paper. The results show that the DSCMM method can reduce the idle driving distance by 26.8%, 10.8% and 8.3% and the passenger waiting time by 11.0%, 13.4% and 6.4% compared to NCD, CCD and MPCD, respectively. Additionally, the market allocation of CTaxis under four scenarios is discussed, and the results demonstrate that the proposed DSCMM method is effective in improving the efficiency of CTaxis operation and reducing the waste of service resources in urban road networks with different service scales.

Original languageEnglish
Article number103792
Number of pages28
JournalTransportation Research Part C: Emerging Technologies
Publication statusPublished - Sept 2022


  • Connected taxi
  • Dynamic dispatch
  • Large-scale network
  • Macroscopic fundamental diagram (MFD)
  • Stochastic roadside service

Cite this