Advances and Applications in Statistics
Volume 31, Issue 1, Pages 1 - 19
(November 2012)
|
|
DURATION MODEL ESTIMATION FOR PAVEMENT REHABILITATION AND SERVICE LIFE
Atef Badr and Aristides G. Karlaftis
|
Abstract: Modeling and predicting pavement distress and deterioration have attracted considerable attention in the literature. Despite this interest, it has largely ignored the effects of rehabilitation treatments on extending service life, an area of obvious importance for both research and practice. In this paper, the problem of extending pavement life due to different rehabilitation treatment applications is addressed. This is done on the basis of data collected from the Specific Pavement Study 5 of the Long Term Pavement Performance Program (LTPP). Bayesian stochastic duration models are developed that account for both parameter uncertainty and model specification uncertainty. Results indicate that, as expected, traffic and the type of treatment significantly affect pavement condition following rehabilitation. Further, the log-logistic functional form estimated via the Bayesian Inference describes distress initiation better than existing approaches. |
Keywords and phrases: |
|
Number of Downloads: 373 | Number of Views: 1069 |
|