BAYESIAN APPROACH APPLIED TO PREDICTION OF PAVEMENT DISTRESS (ABRIDGMENT)

This paper describes a Bayesian approach to verification of pavement design recommendations (materials, construction, and properties) that is to serve as a simplified and less expensive substitute for the generally accepted "test-section approach" and its alternative which involves verification of research results by using case-history studies. The procedure of the Bayesian method used in this study contains the following steps: selection of assessors, identification of primary variables, selection of analysis model, and conducting of personal interviews. From the results presented in this study, it can be conclueded that Bayesian methodology can be used to quantify the judgments of experienced individuals in the form of prior estimates for multiple regression models, which provides a valuable resource in addition to that normally available. Bayesian techniques may also be useful in developing predictive models for modes of pavement distress other than those of fatigue cracking described in this report--which include: low temperature cracking, rutting, reflection cracking, stripping, flushing, and ride quality.