Econometric methods for estimating infrastructure deterioration models with discrete condition data and for computing transition probabilities
Abstract
Markovian transition probabilities have been used extensively in the field of infrastructure management, to provide forecasts of facility conditions. Specifically, Markovian transition probabilities have been used as inputs into discrete dynamic programming algorithm which outputs are in the form of optimal allocation of limited funds among a network of facilities which subsequently will maximize the performance of the facilities. However, existing approaches used to estimate these transition probabilities from inspection data are mostly ad hoc and suffer form several statistical limitations resulting in the optimal solutions of questionable equality; thus, affecting the success of infrastructure management system. This thesis presents new methodologies based on econometric methods for the estimation of infrastructure deterioration models and associated transition probabilities from condition rating data. The first method is based on the Poisson regression model which recognizes the discreteness of condition ratings and, as opposed to state-of-the-art methods, explicitly links deterioration to relevant explanatory variables. The second method is based on the Negative Binomial model, a generalization of the Poisson regression model which relaxes the assumption of equality of mean and variance. The third method is a more rigorous econometric method based on the ordered logit/probit techniques. This technique, as in the case of the Poisson regression and the Negative Binomial models, explicitly links deterioration to relevant explanatory variables. However, in contrary to the Poisson regression and the Negative Binomial models, the ordered logit/probit model recognizes not only the discreteness but also the ordinal nature of condition ratings. An empirical case study, using a bridge inspection data set from Indiana, demonstrates the capabilities of the proposed methodologies.
Degree
Ph.D.
Advisors
Madanat, Purdue University.
Subject Area
Civil engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.