Optimal Planning of Pavement Maintenance and Rehabilitation Considering Pavement Deterioration Uncertainty

Optimization of pavement maintenance and rehabilitation (M R) is one of the most substantial parts of the pavement management system (PMS). Highway agencies should plan M R treatments efficiently. An accurate pavement performance model is required to predict pavements’ future conditions. Thus, an accurate international roughness index (IRI) model was developed to predict IRI. Moreover, M R was scheduled deterministically in many past studies, but this issue does not match the uncertain essence of deterioration and future M R expenditures. Hence, the uncertainty associated with pavement deterioration and budget calculation should be considered in scheduling M R activities. This study scheduled M R activities deterministically and probabilistically to compare the solutions obtained from both approaches. The uncertainty of several features in the IRI model and budget calculation was not considered in the deterministic approach. Furthermore, in the probabilistic approach, historical data were employed to fit the distribution function for uncertain features in the model. Then, Monte Carlo simulation and optimizer were run to generate probability distributions for sections’ IRI and required budget and optimize M R scheduling. The IRI model was developed using 288 data. The testing data R-square of the model was 0.917. As a case study, the research results were applied to a network, including five sections during a 5-year-planning. Additionally, the costs of M R scheduling in the deterministic and probabilistic approaches were $52,149 and $40,195. Hence, the cost of the deterministic approach was 29.7% higher than the probabilistic approach. Besides, the probabilistic method applied more preventive maintenance, desirable for users, than the deterministic one.