Date on Master's Thesis/Doctoral Dissertation
Industrial Engineering, MS
Pavement deterioration is one of the most important driver for prioritizing pavement management and preservation (PMP) projects. The primary goal of this thesis is to provide reasonable predictive functions from multiple linear regression (MLR) models and artificial neural networks (ANN) that can be adopted by Kentucky Transportation Cabinet (KYTC). Furthermore, we use analytic hierarchy process (AHP) to design a composite pavement distress index in order to help Kentucky Transportation Cabinet (KYTC) prioritizing PMP projects based on 11 different distress indices. Numerical results show that the MLR models provide relatively high R square values of approximately 0.8. Both MLR and ANN models have small average squared errors (ASE). Finally, for all nine distress indices studied in this thesis, MRL models are recommended to KYTC due to their simplicity, interpretability along with robust performance that is comparable to the ANN model. Finally, a priority rating method is developed using analytical hierarchy process and it integrates 11 pavement distress indices into one priority score. A case study shows that the propose AHP-based rating method overcomes the drawback of KYTC’s current rating system for overemphasizing the international roughness index (IRI) among all distress indices.
Luo, Chenglong, "Pavement deterioration modeling and design of a composite pavement distress index for Kentucky interstate highways and parkways." (2014). Electronic Theses and Dissertations. Paper 868.