Transportation Research Record Journal of the Transportation Research Board · 1987 · 197 citations · 0 references
EngineeringDecision SciencesCondition AssessmentMaintenance EngineeringPavement EngineeringMaintenance OptimizationPavement ConstructionPavement DesignPavement EvaluationPavement MaintenancePavement Performance AnalysisNonhomogeneous Markov ChainsPavement Condition IndexPavement ManagementPavement Performance ModelingPavement Performance PredictionPerformance Prediction ModelsPrediction Model DevelopmentRepair Strategies
Accurate and efficient pavement performance prediction models are essential for effective pavement management systems. The study aims to develop a model that informs optimal maintenance and repair decisions for pavements. The model uses a ten‑state Pavement Condition Index and divides pavement life into six‑year zones, applying homogeneous and nonhomogeneous Markov chains with zone‑specific transition matrices derived via nonlinear programming. The model accurately predicts future pavement conditions and is benchmarked against a constrained least‑squares model.
A good pavement-management system requires an accurate and efficient pavement performance and prediction model. A pavement performance and prediction model based on the Pavement Condition Index and the age of the pavement has been developed. The Pavement Condition Index ranging from 0 to 100 has been divided into ten equal condition states. A combination of homogeneous and nonhomogeneous Markov chains has been used in the development of the model. The life span of the pavement is divided into zones, with each zone representing a period of 6 years. The transition matrix of each zone is determined using nonlinear programming. If the state of any given pavement section is known, its future condition can be predicted efficiently from the corresponding transition matrices. The model presented in this paper will play an integral part in the decision-making procedure for determining optimal maintenance and repair strategies. A comparison between the Markov model and the constrained least-squares model is presented.