Deployment of an intelligent transportation systems (ITS) program such as a real-time travel guidance system requires the good understanding of people's travel choice process. The whole travel choice process includes a series of choices including trip choice, destination choice, mode choice, departure time choice and route choice. Traditionally, static travel choice models or transportation network models have been developed to model the travel choice process. However, the static models cannot provide the real time traffic volume and travel time and cannot reflect the time-dependent variation of traffic in a road network. Thus static travel choice models cannot model the dynamic process in travel choice. The dynamic models can provide the real time link and path traffic volume and link and path travel time and can model the dynamic process in travel choice. Dynamic models are also applicable to long-term transportation planning. Unfortunately, the current studies on dynamic travel choice/dynamic transportation network have limitations on either modeling method or solution algorithm, which impede their application in practice.
In this dissertation, I have conducted a comprehensive study on dynamic travel choice problems and have presented a series of variational inequality models and solution algorithms to these problems. The problems that the dissertation addresses include deterministic dynamic user optimal route choice problem (DUO), stochastic dynamic user optimal route choice problem (SDUO), dynamic user optimal simultaneous departure time and route choice problem (DUOSDTRC), combined mode split and dynamic user optimal simultaneous departure time and route choice problem (MS DUOSDTRC), combined trip distribution and dynamic user optimal simultaneous departure time and route choice problem (TD DUOSDTRC), and combined trip distribution mode split and dynamic user optimal simultaneous departure time and route choice problem (TD MS DUOSDTRC). The innovative work is reflective of the successful modeling and development of corresponding algorithms without time-space network expansion. As a result, simplified and potentially efficient solution algorithms to the dynamic travel choice problems over a large-scaled transportation network are developed. All the models and algorithms are validated by numerical examples.