Highlights Route choice of travelling backward is modelled using a data-driven model. The model comprises three GMM-based hierarchical Bayesian models. No-U-turn sampler (NUTS) is introduced to estimate the model parameters. A case study of Beijing metro network is discussed. The proposed framework can help develop an accurate passenger flow assignment model. The framework has been successfully applied to demand assignment and passenger flow control.
Abstract This study proposes a data-driven approach to understand traveling backward (TB) behavior while making a route choice in congested metros. First, TB behavior during route choice in overcrowded metros is defined and analyzed. Second, a hybrid model comprising three hierarchical Bayesian models is developed to describe the TB behavior. Third, a novel sampling method based on Hamiltonian dynamics is introduced to estimate the model parameters. Finally, a case study of Beijing metro is discussed to illustrate the effectiveness and robustness of the proposed model. The contributions of application of the proposed model to accurate demand assignment and passenger flow control for metro managers are demonstrated.
Data-driven approach for solving the route choice problem with traveling backward behavior in congested metro systems
2020-07-14
Article (Journal)
Electronic Resource
English
Junction design and congested route choice
Tema Archive | 1990
|Information Impacts on Route Choice and Learning Behavior in a Congested Network
Transportation Research Record | 2011
|A topological route choice model for metro
Elsevier | 2010
|