The reliability of travel times is a critical factor contributing to the efficiency and service quality of a transportation system. While in the past traffic network performance has been assessed mainly in terms of the overall congestion level measured by average travel time, the probability of unexpected delays or variability of travel times becomes a growing concern for individual travelers and shippers whose trips are under tight time constraints. This has led travel time reliability to be regarded as a key performance indicator of service quality of a given traffic network. As such, traffic analysts and policy makers increasingly recognize the importance of incorporating travel time reliability in network planning and traffic operations as an essential evaluation criterion. Given this, the objective of this dissertation is to develop a framework and methodologies for modeling travel time reliability that provide a systematic approach to evaluating a broad range of transportation policies and projects. Towards achieving this goal, the study addresses two main aspects of travel time reliability modeling and evaluation: (i) to quantify and characterize travel time variability given travel time observations to allow accurate evaluation of reliability performance and (ii) to obtain travel time distributions that reflect real-world variability and uncertainty under a particular system to be evaluated.;For the former, the study presents a novel approach to characterizing different types of travel time variability reflected in detailed travel time data collected across different vehicles and days. In particular, the study proposes a compound Gamma distribution that captures both vehicle-to-vehicle and day-to-day variability in modeling travel time reliability. The proposed model, termed Gamma-Gamma distribution, is analytically derived in a straightforward manner by using a specific property of observed travel time data, namely, a linear relation between the standard deviation and mean of distance-normalized travel time. The main advantage of the Gamma-Gamma model is its ability to recognize different variability dimensions reflected in travel time data, and clear physical meanings of its parameters in connection with vehicle-to-vehicle and day-to-day variability. As such, the model provides a systematic means of quantifying, comparing and assessing different types of variability, which is important in understanding travel time characteristics and evaluating various transportation interventions that affect reliability.;For the latter, the study proposes a comprehensive framework that features scenario-based simulation approaches. The approach aims to capture exogenous sources of unreliability such as weather, incidents and special events through a set of explicit roadway disruption scenarios and analyze their impacts on travel time outcomes using simulation-based dynamic traffic assignment (DTA) models; this then provides the ability to observe potential variations in travel time and obtain accurate estimates of population travel time distributions. After presenting a conceptual framework, the study addresses practical issues that arise in implementing the proposed approach. Two major issues include: (a) the generation of scenarios considering complex interactions among different uncertainty sources and (b) the selection of critical scenarios that effectively represent the entire scenario space.
展开▼
