Abstract
Decision reliability is becoming an important issue for both researchers and transportation service providers; different models and methods have been proposed to improve decision reliability. A method is proposed for improving decision reliability by analyzing the moments and central moments of historical travel time data. Because of the existence of traffic congestion—especially nonrecurrent traffic congestion—travel time is generally a random variable with an asymmetric historical distribution. Statistical methods are applied to obtain the higher-order moments and central moments of historical travel time, which provide quantitative information about the variability, asymmetry, and leptokurtosis of travel time. This information is useful in making reliable decisions and avoiding scenarios that can potentially lead to extremely high costs. The method is based on advanced traveler information systems, which provide useful historical travel time information for traffic networks. The new method and other methods are compared with two testing experiments.
Get full access to this article
View all access options for this article.