Travel time unreliability is an important characterization of transportation systems. The appropriate modeling and reporting of travel time reliability is crucial for individual travelers as well as management agencies. In this paper, a novel multi-state model is proposed for travel time modeling and reporting. The model advances travel time modeling in two aspects. First, the multi-state model provides much improved model fitting as compared to single-mode models. Second and more importantly, the proposed model provides a connection between travel time distributions and the underlying travel time state. This connection allows for the quantitative evaluation of probability of each travel time state as well as the uncertainty associated with each state; e.g., the probability of encountering congestion on a given time of day and the expected travel time if congestion were experienced. A simulation study was conducted to demonstrate the performance of the model. The proposed model was applied to field data collected at San Antonio, TX. The variation of the model parameters as a function oftime-of-day was also investigated. The simulation study and field data analysis confirmed the superiority of multi-state model over the state-of-practice single-mode travel time reliability models and ease of interpretation.
Abstract