Record:   Prev Next
Author Tong, Chee Chung
Title A study of dynamic departure time and route choice behavior of urban commuters
Descript 141 p
Note Source: Dissertation Abstracts International, Volume: 52-01, Section: B, page: 0411
Supervisor: Hani S. Mahmassani
Thesis (Ph.D.)--The University of Texas at Austin, 1990
Traffic congestion is widely recognized as a problem plaguing the growth of modern cities, particularly during the peak hours when most work commuting trips are made. Because it is not generally feasible to physically expand the existing roads, much effort has turned to developing improved control strategies to relieve the congestion in the existing facilities. Such effort requires understanding the fundamentals of traffic build-up and dissipation which result from the aggregation of individual commuters' travel decisions during a specific period of time in a particular location. This leads to the need for studying the commuters' behavior with regard to how these travel decisions are made
The focus of this dissertation is on the day-to-day dynamic decisions of route and departure time by urban commuters. In order to investigate the underlying behavioral mechanisms for such decisions, a decision process framework is developed integrating concepts from psychology, behavioral science, and marketing science. Within this framework, two alternative models predicated on two distinct decision behavior theories are developed: (1) a "switching model" based on bounded-rationality and the "satisficing rule" and (2) a "best-alternative" model based on the random utility maximization. Formulations and specifications of both models are explored, including the econometric issues associated with specifying the random error structure for parameter estimation purposes. A general probit model form is used for the dynamic switching model, allowing the inclusion of state dependence and serial correlation in the model specification. The nested logit model form is used for the "best alternative" random utility model of joint departure time and route choice
Observations were obtained through a laboratory-like experiment where real commuters interact with a special purpose traffic simulation system. The experiment was conducted over a period of approximate 30 working days in order to provide the data base for studying the consecutive day-to-day decisions and estimating model parameters. Participants were divided into two information availability groups to examine the effect of exogenous information on daily commuting decisions and overall system performance
School code: 0227
Host Item Dissertation Abstracts International 52-01B
Subject Engineering, Civil
Alt Author The University of Texas at Austin
Record:   Prev Next