Stochastic Traffic Flow Modeling and Optimal Congestion Pricing.

Traffic Flow Modeling;Congestion Pricing;Traffic State Estimation;Intelligent Transportation System;Managed Toll Lane;Industrial and Operations Engineering;Transportation;Engineering;Industrial and Operations Engineering

Yang, LiShen, Siqian May ;

University of Michigan

关键词: Traffic Flow Modeling;    Congestion Pricing;    Traffic State Estimation;    Intelligent Transportation System;    Managed Toll Lane;    Industrial and Operations Engineering;    Transportation;    Engineering;    Industrial and Operations Engineering;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/94054/youngli_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Congestion in surface transportation networks causes serious economical and environmental problems in urban areas of the United States and around the world. Due to the continuous deterioration of urban traffic conditions in recent years, increasing numbers of intelligent transportation system applications have been developed and implemented to help administrators manage the highway and make the usage of the whole traffic network more efficient. The success of most intelligent transportation system applications is dependent on an accurate prediction of the future traffic state, because such accurate prediction can help the decision maker choose the right strategy and/or provide reliable travel information to the drivers. This dissertation explores an innovative stochastic traffic flow model to better predict the future traffic state on the highway, and provides a framework to investigate some potential applications, including dynamic congestion pricing.Part I presents an innovative macroscopic stochastic traffic flow model and the off-line calibration algorithm for this model. It also develops the numerical algorithm for future traffic state prediction based on this model. The model is validated by using real highway data. The empirical results show that this model outperforms, in terms of prediction accuracy, the traditional macroscopic traffic flow model.Part II is devoted to the development of on-line parameter calibration and traffic state estimation algorithms for the proposed innovative stochastic traffic flow model. The algorithms are tested on synthetic data, and the numerical results show that the algorithms are able to capture the change of model parameters and improve the prediction accuracy of the traffic flow model.Part III formulates a mathematical model for the problem of optimal distance-based dynamic congestion pricing. This mathematical model utilizes the proposed stochastic traffic flow model as the underlying dynamics of the traffic flow. It develops the optimal dynamic pricing strategy under a specified objective and a numerical case study is presented to illustrate the solution.