Travel time and travel time reliability are important attributes of a trip. Travel time reliability does not only rely on variability but also on the stability of travel times. This paper clarifies the attributes of reliability and proposes a new analytical formula to express travel time unreliability in terms of these elements, in which the travel time (un)reliability is computed as the sum over the products of the consequences (variability or uncertainty) and corresponding probabilities of traffic breakdown (instability). The travel time reliability model is considered as a function of a variety of factors. In essence, these factors are conditionals, that is, the function expresses travel time reliability for a certain inflow level, given certain circumstances. These circumstances include road characteristics, and all other relevant factors like traffic control measures, the prevailing traffic state (congested or not), and possibly external factors like weather, luminance, etc. The model is validated and calibrated on the basis of the empirical data collected from Regiolab-Delft traffic monitoring system. It is found that both the probability of traffic breakdown and travel time unreliability increase with the increasing inflows. To conclude, the travel time reliability model developed in this dissertation yields a tractable approach that captures the main features of travel times and travel time (un)reliability. It provides a new, useful framework for describing and understanding travel time unreliability and to find ways to improve reliability and analyse the impact of unreliability. Such a reliability model can be easily embedded in a DTA model to estimate or predict traffic conditions. In that sense, the findings and results from this dissertation thesis pave the way for quantitative and both theoretically and empirically underpinned research into the role of travel time reliability in route and departure time choice. (Author/publisher)
Samenvatting