ABSTRACT
Sustainability considerations across the entire pavement life-cycle are essential in the decision-making process under uncertainty to achieve optimal pavement management from the perspectives of economy, environment, and society. This study presents a risk- and sustainability-based management optimization approach for asphalt pavement. First, a deep neural network (DNN) model is developed using data from the Long-Term Pavement Performance (LTPP) database, incorporating various asphalt pavement parameters. Based on multiple time-dependent limit-state functions that account for the uncertainties in these parameters, the DNN model predictions, and International Roughness Index (IRI) measurements, a Monte Carlo simulation is conducted to estimate the system failure probability and risk associated with asphalt pavement. Finally, a genetic algorithm-based tri-objective optimization is employed to identify the optimal maintenance and rehabilitation actions, minimizing economic, environmental, and social impacts throughout the pavement’s life-cycle. The effectiveness of the proposed approach is demonstrated using LTPP asphalt pavement sections in Florida, USA.
