State-of-the-art review of pavement deterioration models for enhancing transportation infrastructure management

Pavement deterioration modeling plays a critical role in modern pavement management systems by enabling accurate prediction of performance and supporting maintenance and rehabilitation planning. Pavement deterioration results from the combined effects of traffic loading, environmental conditions, material characteristics, and maintenance strategies. This paper presents a comprehensive review of pavement deterioration models, categorizing them into deterministic, probabilistic, Machine learning and AI techniques and hybrid approaches. Deterministic models, including empirical, mechanistic, and ME formulations, use predefined mathematical relationships to predict pavement condition. Probabilistic models, particularly those based on Markov processes and hazard analysis, account for uncertainty and are suitable for long-term forecasting. Machine learning and AI techniques such as ANN, SVM, genetic programming, and ensemble learning have recently demonstrated superior performance by capturing nonlinear and multivariate deterioration behavior. Hybrid models integrating deterministic, probabilistic, and data-driven approaches further enhance prediction reliability by combining the strengths of multiple modeling paradigms. Key performance indicators such as the IRI, PCI, and PSI are reviewed in terms of their evolution, use, and limitations. ME sensitivity studies are discussed to identify influential parameters, including layer thickness, resilient modulus, asphalt stiffness, and cumulative traffic loading. The review also highlights persistent challenges such as limited data availability, lack of globally adaptable models, and inconsistent integration of maintenance effects. Future directions include the development of transferable models calibrated for diverse conditions, real-time monitoring using sensor-based data, and explainable AI frameworks for transparent decision-making. This representative synthesis provides a structured understanding of current pavement deterioration modeling practices and offers guidance for advancing predictive accuracy and infrastructure management efficiency.