| The combination of independent individuals' travel choices within the current transportation system results in complicated travel patterns, involving many variables such as personal characteristics, traffic conditions, incidents, accidents, construction, and weather. Stochastic Dynamic Transportation Network Modeling (SDTNM) is to combine extensive knowledge of travel demand estimation, traveler behavior analysis, traffic flow theory, and socio-economic science to generate a good approximation of the actual traffic condition within a certain error range. This study concentrates on enhancing the SDTNM by using actual and perceived travel costs, the principles of Stochastic Dynamic User Optimal, and corresponding solution algorithms towards real-time applications.; This study synthesizes the Maximum Likelihood Estimation and Expectation Maximization methods to calibrate and validate SDTNM with detection errors and incomplete real-time data. The enhanced model provides the estimation and prediction of time-varying traffic volumes on the transportation network, as well as dynamic route choices based on maximum likelihood estimation.; This study extends the disutility theory to represent diverse route choice behavior and stochastic dynamic traffic condition. The risk-taking human factors and stochastic dynamic travel costs are modeled using multiple classifications of dynamic route choice models. The enhanced model provides the calibration capability for multi-class travelers such as different risk-taking types of travelers, Single-Occupancy Vehicles, and High-Occupancy Vehicles.; The Dynamic System Optimal principle is used in the dynamic value pricing in order to enhance the efficiency of network utilization. A practical process to obtain optimal dynamic value pricing strategy is also developed. Furthermore, this study provides a decision support tool to predict, evaluate and implement Intelligent Transportation Systems applications such as dynamic value pricing strategies by implementing the enhanced SDTNM to evaluate dynamic value pricing strategies on the I-394 corridor network in Minneapolis. The enhanced model provides quantitative support to make decisions on the optimal price and location for toll collection. Except for the simulation on the I-394 corridor network, the methodology and the solution process are also demonstrated on simple networks. In the future, more testing should be conducted on large networks to find out potential drawbacks or limitations. |
