A Two-Stage Algorithm for Origin-Destination Matrices Estimation Considering Dynamic Dispersion Parameter for Route Choice

• DOI: 10.1371/journal.pone.0146850
• Publication Date: 2016-01-13T18:41:37Z
• Authors: Yong Wang; Xiaolei Ma; Yong Liu; Ke Gong; Kristian C. Henri...; Maozeng Xu; Yinhai Wang
• Keywords: Washington; Travel; Time Factors; Science; Q; 05 social sciences; R; Numerical Analysis, Computer-Assisted; Models, Theoretical; Choice Behavior; 0502 economics and business; Medicine; Algorithms; Research Article

This paper proposes a two-stage algorithm to simultaneously estimate origin-destination (OD) matrix, link choice proportion, and dispersion parameter using partial traffic counts in congested network. A non-linear optimization model is developed which incorporates dynamic parameter, followed by Generalized Least Squares (GLS) estimation Stochastic User Equilibrium (SUE) assignment are iteratively applied until the convergence reached. To evaluate performance of algorithm, proposed approach implemented hypothetical network input data with high error, tested under range variation coefficients. The root mean squared error (RMSE) estimated OD demand flows used results. results indicate that theta insensitive shown outperform two established methods produce estimates close ground truth. In addition, an empirical Seattle, WA validate robustness practicality this methodology. summary, study evaluates innovative computational accurately matrices link-level flow data, provides useful insight for optimal selection modeling travelers' route behavior.