Abstract
Current practice of measuring travel time reliability is reviewed and a standardized information entropy-based perceived reliability measure is proposed. The methods proposed in the paper utilize existing theories on uncertainty perception such as the categorical perception nature of human minds and the Shannonâ??s entropy concepts to resolve issues resulting from significant correlation between mean travel time and variance so that the measure can be linearly embedded into the existing cost or utility function of the planning and demand-responsive simulation models. By introducing Von Neumannâ??s information entropy, system-wide misperception can be also modeled so that perceived uncertainty can be estimated more realistically. A case study combining path-based traffic assignment algorithm, Bayesian update, and the entropy measure is used to demonstrate the modeling and analysis process. The approach has the potential to be extended to more general transportation problems related to travelersâ??, operatorsâ??, as well as shippersâ?? perception and information availability studies.