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.

Conventional PMSs often rely on static age-based assumptions, which can fail to capture nonlinear, state-dependent deterioration and improvement-like responses observed in long-term monitoring data. This study addresses these limitations by proposing a reaction-oriented analytical framework using eight years of Korea Highway PMS data (2015–2022). We construct a Δ–State Vector by combining the previous-year condition grade with noise-filtered annual changes in the International Roughness Index (IRI) and Rut Depth (RD). Measurement noise is separated from structural signals via MAD-based noise bands (ΔIRI: ±0.089 m/km; ΔRD: ±0.993 mm), with a global MAD floor (minimum-threshold constraint) to avoid degenerate zero-band cases under sparse or near-constant transitions. The resulting vectors are embedded into a low-dimensional Reaction Space using UMAP and clustered with HDBSCAN. To validate interpretability, a rule-based Trend × Mode Reaction Signature taxonomy is used to assess the semantic consistency of unsupervised clusters. Five dominant reaction regimes are identified, showing strong agreement with signature-based labels (weighted purity = 0.927; coverage for purity ≥ 0.60 = 0.911). Overall, the results indicate that deterioration dynamics are governed by lane–segment heterogeneity and prior-state dependence rather than chronological age, providing a reproducible foundation for future event-sensitive, dynamic age reset frameworks.

Modern practices of planning road maintenance activities are increasingly based on forecasting changes in pavement condition over time. Under conditions of limited funding and growing traffic loads, scientifically substantiated determination of the timing and scope of maintenance and repair actions based on pavement deterioration models becomes particularly important. The key parameter that integrally reflects the influence of structural, climatic, and traffic-related factors is the structural capacity of the pavement structure. This paper examines the patterns of changes in the structural capacity of road pavement structures during operation. The stages of deterioration of flexible pavements are analyzed, and the relationship between the reduction in structural capacity, the development of pavement distress, deterioration of ride quality, and traffic safety conditions is established. The obtained dependencies can be used to substantiate maintenance strategies and to plan repair works. Existing approaches to pavement condition forecasting are often based on simplified empirical models or do not adequately account for the stage-wise nature of pavement structural capacity degradation. In addition, universal models developed for other countries do not always adequately reflect the real operating conditions of Ukrainian highways, which are characterized by excessive axle loads, significant irregularity of funding, and additional destructive impacts caused by military actions and changes in traffic flow structure. The purpose of this study is to generalize and further develop scientific approaches to modeling pavement deterioration based on the stage-wise change in pavement structural capacity, as well as to substantiate the applicability of such models for predicting the serviceability and operational condition of automobile roads under modern operating conditions. This study is of a review nature. A systems approach was applied, which represents a set of general scientific methodological principles based on considering the objects of study as complex systems. It was established that pavement deterioration has a clearly defined stage-wise character, in which the reduction of pavement structural capacity is the determining factor governing the development of pavement distress, the increase in the area of potholes, and the deterioration of surface evenness. The interrelation between structural capacity indicators, ride quality, and traffic safety was demonstrated. The feasibility of using combined deterioration models that integrate mechanistic–empirical and probabilistic approaches was substantiated. It was shown that the application of locally calibrated models improves the reliability of residual service life prediction for pavements. The proposed approach to modeling pavement deterioration based on the stage-wise change in pavement structural capacity can serve as a scientific basis for decision-making in road asset management systems. Adapting deterioration models to the actual operating conditions of Ukrainian highways is a necessary prerequisite for improving the efficiency of road maintenance and the rational use of limited financial resources.

Assessing pavement conditions in Nepal is costly and time-consuming, with rising traffic and aging infrastructure making maintenance increasingly challenging. This study developed and compared pavement deterioration models to predict the Surface Distress Index (SDI) without manual assessment, using historical road data. SDI was modeled as a function of five key factors: International Roughness Index (IRI), pavement age, total annual rainfall, annual temperature range, and commercial vehicle traffic. Data were collected from relevant government sources, covering 157 road sections with a combined length of 15,783 km, for the period from 2012 to 2022. Multiple Linear Regression (MLR) and Artificial Neural Network (ANN) models were developed for SDI prediction. MLR analysis, conducted in Microsoft Excel, assessed statistical significance through ANOVA, R² values, and regression coefficients. In contrast, ANN modeling utilized a Multi-Layer Perceptron (MLP) architecture implemented in TensorFlow and Keras. The ANN model was optimized through iterative experimentation with varied architectures, employing ReLU activation and the Adam optimizer for adaptive learning. The study evaluated a range of architectures, beginning with simple single-layer networks and extending to Deep Neural Networks (DNNs) with up to four hidden layers. Results showed that, during model development, MLR achieved an R² of 0.735, whereas the ANN model, with a 5-232-1 structure and 104 epochs, outperformed MLR with an R² of 0.809. Validation of both models indicated strong alignment between observed and predicted values, with ANN demonstrating superior predictive accuracy (R² = 0.816) compared to MLR (R² = 0.74). The error histogram further confirmed ANN’s better performance, which confirms its improved reliability. The study highlights the effectiveness of both models while emphasizing ANN’s advantage in capturing complex nonlinear relationships. These findings suggest that integrating ANN into Nepal’s pavement management framework can enhance predictive accuracy, reduce assessment costs, and support more efficient maintenance planning.

Accurate modeling of pavement performance is vital to maintaining safe, reliable, and sustainable transportation infrastructure. This review synthesizes current approaches to pavement deterioration modeling, with emphasis on key influencing factors, performance indicators, and methodologies employed within Pavement Management Systems (PMS). Primary deterioration drivers, including traffic loading and environmental stressors, are analyzed for their impact on degradation patterns. Performance indicators such as the Pavement Surface Evaluation and Rating (PASER), Pavement Condition Index (PCI), and International Roughness Index (IRI) are evaluated for their effectiveness in capturing pavement condition and guiding maintenance decisions. Modeling techniques are broadly categorized into deterministic, probabilistic, and intelligent (machine learning–based) frameworks to illustrate the evolution of predictive approaches. Across these approaches, the notion of symmetry can be interpreted as the balance and consistency achieved between model assumptions, input variables, and predicted pavement behavior, while asymmetry represents deviations caused by uncertainty, variability, and nonlinearity inherent in real-world conditions. Recognizing these symmetrical and asymmetrical relationships helps unify different modeling paradigms and provides insight into how each framework handles equilibrium between accuracy, complexity, and interpretability. The review also highlights persistent challenges in data availability, quality, and standardization. Notably, the increasing adoption of machine learning reflects its capacity to handle high-dimensional and spatiotemporal datasets. Recommendations are proposed to improve the robustness, scalability, and transparency of future deterioration models, thereby enhancing their role in data-driven, resilient, and cost-effective pavement management strategies.

This study compiled a dataset of published works relating to fatigue testing in asphalt mixes, covering 2020–2025. The dataset was subjected to bibliometric and textual analyses, including a systematic review, to explore emerging trends and patterns in experimental protocols. Bibliometrix, VOSviewer, and IRaMuTeQ were employed to map the scientific landscape of 368 articles. Following PRISMA guidelines, the 100 most-cited articles were reviewed to identify prevailing test setups and parameters. The results showed a growing scientific production (9.1% per year), concentrated in a few high-impact journals and dominated by China, with emphasis on sustainability. A comparison between scientific output and a road quality index revealed a disconnect between academic research and field implementation. Five thematic clusters emerged: sustainable pavement management, mechanical characterization, binder modification, performance modeling, and evaluation of innovative materials. Indirect tensile and four-point bending tests were the most common loading modes. Considerable variability in protocols, frequent omissions of methodological details, and limited statistical treatment were also observed. The study highlighted the importance of standardized reporting and robust analysis, offering a reproducible framework to understand fatigue behavior and support future research.

Accurate and reliable modeling of pavement deterioration is critical for effective infrastructure management. This study proposes a probabilistic machine learning framework using Bayesian-optimized Natural Gradient Boosting (BO-NGBoost) to predict the International Roughness Index (IRI) of asphalt pavements in cold climates. A dataset only for cold regions was constructed from the Long-Term Pavement Performance (LTPP) database, integrating multiple variables related to climate, structure, materials, traffic, and constructions. The BO-NGBoost model was evaluated against conventional deterministic models, including artificial neural networks, random forest, and XGBoost. Results show that BO-NGBoost achieved the highest predictive accuracy (R2 = 0.897, RMSE = 0.184, MAE = 0.107) while also providing uncertainty quantification for risk-based maintenance planning. BO-NGBoost effectively captures long-term deterioration trends and reflects increasing uncertainty with pavement age. SHAP analysis reveals that initial IRI, pavement age, layer thicknesses, and precipitation are key factors, with freeze–thaw cycles and moisture infiltration driving faster degradation in cold climates. This research contributes a scalable and interpretable framework that advances pavement deterioration modeling from deterministic to probabilistic paradigms and provides practical value for more uncertainty-aware infrastructure decision-making.

Purpose: The purpose of this study is to assess the performance of double bituminous surface treatment on the Malekhu–Dhading Beshi (MDB) Road in Nepal-Asia. Design/Methodology/Approach: The descriptive study was conducted to evaluate the compatibility of the road condition assessment method for analysing results and assessing road conditions between 2012 and 2021. The assessment method consists of the International Roughness Index (IRI), the Surface Distress Index (SDI) and the Pavement Serviceability Rating (PSR). The correlation between SDI and IRI, SDI and Average Annual Daily Traffic (AADT), IRI and AADT, SDI and Age of pavement, and IRI and Age of pavement were obtained from the correlation analysis. Research Limitation: The study lacked adequate data on the quality and availability of the performance of the road projects and the delays in the study area. Findings: The relation between IRI-Traffic and SDI-IRI is positive, with R2 values of 0.0713 and 0.6831, respectively. The relation between IRI and Traffic is poor, and the relation between SDI and IRI is good. The relation between SDI-Traffic and SDI-Age of pavement is logarithmic, with R2 values of 0.4786 and 0.4319, respectively, which is a moderate relationship. The relation between the Ages of pavement and IRI is polynomial with an R2 value of 0.2676, indicating a poor relationship. Pavements in this category (value of PSR between 1.00 and 2.00) have deteriorated to such an extent that they affect the speed of free-flow traffic. Practical Implication: Understanding performance characteristics enables the strategic timing of applications, the selection of appropriate treatment types, and the prediction of maintenance cycles. This leads to a more efficient allocation of public resources and extended pavement life cycles. Social Implication: Enhanced road surfaces facilitate emergency vehicle access, school bus transportation, and agricultural product movement, directly impacting quality of life and social equity. Originality/Value: This pavement deterioration model can be used for the forecast of future values of IRI. This model is the basis for the assessment of Double Bituminous Surface Treatment pavement.

Stochastic-based pavement performance and deterioration models: A review of techniques and applications

A Multiple Regression Pavement Deterioration Model for National Highways of Nepal

Pavement deterioration is influenced by various factors with degradation rates varying widely depending on the type of pavement, its use, and the environment in which it is located. In Nepal, where the climate varies from alpine to subtropical monsoon, understanding pavement degradation is essential for effective road asset management. This study employs a Markov deterioration hazard model to predict pavement deterioration for the national highways managed by Nepal’s Department of Roads. The model uses Surface Distress Index data from 2021 to 2022, with traffic and cumulative monsoon rainfall as explanatory variables. Monsoon rainfall data from meteorological stations were interpolated using Inverse Distance Weighted and Empirical Bayesian Kriging 3D methods for comparative analysis. To compare the accuracy of interpolated values from the IDW and EBK3D methods, error metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Bias Error (MBE) were employed. Lower values for MAE, RMSE, and MBE indicate that EBK3D, which accounts for spatial correlation in three dimensions, outperforms IDW in terms of interpolation accuracy. The monsoon rainfall interpolated values using the EBK3D method were then used as an explanatory variable in the Markov deterioration hazard model. The Bayesian estimation method was applied to estimate the unknown parameters. The study demonstrates the potential of integrating the Markov deterioration hazard model with monsoon rainfall as an environmental factor to enhance pavement deterioration modeling. This model can be adapted for regions with a similar monsoon climate and pavement types making it a practical framework for supporting decision-makers in strategic road maintenance planning.

To address the urgent need for accurate pavement performance modeling in pavement design, this study proposes a meta-learning-based few-shot learning method for predicting the Deflection Basin Area (DBA) of asphalt pavements. The method utilizes features such as pavement temperature and load pressure, and applies cyclic DBA data from various pavement types subjected to different pressures. The objective is to predict the trend of DBA changes over cycles at a specific pressure. By leveraging pre-training on diverse pavement datasets, the proposed meta-learning model reduces the training data required for target pavement DBA prediction, enabling better generalization to the target pavement. This approach enhances DBA prediction accuracy even with a small sample size. Compared to traditional machine learning and pre-training methods using data from a single pavement type, the proposed method achieves a Mean Square Error of 13.26 and a Mean Absolute Error of 2.85, demonstrating superior performance. Furthermore, it achieves high prediction accuracy with fewer iterations. Overall, the proposed method effectively predicts DBA across various pavement structures with a few data.

There is a growing global interest in preserving transportation infrastructure. This necessitates routine evaluation and timely maintenance of road networks. The effectiveness of pavement management systems (PMSs) heavily relies on accurate pavement deterioration models. However, there are limited comparative studies on modeling approaches for rural roads in arid climatic conditions using the same datasets for training and testing. This study compares three approaches for developing a pavement condition index (PCI) model as a function of pavement age: classical regression, machine learning, and deep learning. The PCI is a pavement management index widely adopted by many road agencies. A dataset on pavement age and distress was collected over a twenty-year period to develop reliable predictive models. The results demonstrate that the regression model, machine learning model, and the deep learning model achieved a coefficient of determination (R2) of 0.973, 0.975, and 0.978, respectively. While these values are technically equal, the average bias for the deep learning model (1.14) was significantly lower than that of the other two models, signaling its superiority. Additionally, the trend predicted by the deep learning model showed more distinct phases of PCI deterioration with age than the machine learning model. The latter exhibited a wider range of PCI deterioration rates over time compared to the regression model. The deep learning model outperforms a recently developed regression model for a similar region. These findings highlight the potential of using deep learning to estimate pavement surface conditions accurately and its efficacy in capturing the PCI-age relationship.

This paper used pavement condition data collected by the Federal Highway Administration (FHWA) between 2001 and 2006 aggregated by U.S. states to identify macroscopic factors affecting pavement roughness in time and space. To account for prior pavement conditions and preservation expenditure over time, time autocorrelation parameters were introduced in a spatial modeling scheme that accounted for spatial autocorrelation and heterogeneity. The proposed framework accommodates data aggregation in network-level pavement deterioration models. Because pavement roughness across different roadway classes is anticipated to be affected by different explanatory parameters, separate time–space models are estimated for nine roadway classes (rural interstate roads, rural collectors, urban minor arterials, urban principal arterials, and other freeways). The best model specifications revealed that different time–space models were appropriate for pavement performance modeling across the different roadway classes. Factors that were found to affect state-level pavement roughness in time and space included preservation expenditure, predominant soil type, and predominant climatic conditions. The results have the potential to assist governmental agencies in planning effectively for pavement preservation programs at a macroscopic level.

A common phenomenon in developing countries is that the function of the pavement in the road network will experience structural damage before the completion of life is reached, and the uncertainty of pavement damage is difficult to predict. Planning for maintenance treatment depends on the accuracy of predicting future pavement performance and observing current conditions. This study aims to apply the Markovian probability operational research process to develop a decision support system predicting future pavement conditions. Furthermore, it determines policies and effectiveness in managing and maintaining roads. A standard approach that can be used by observing the history of pavement damage from year to year is to estimate the transition probability as a Markovian-based performance prediction model. The results show that the application of the model is quite optimal, changes in pavement conditions after repair can be easily compared with an increase in good condition, reaching 92.8%. Routinely and consistently handling road deterioration will give favorable results regarding pavement condition value. This will ease in the management of the road network and the accomplishment of the optimal maintenance and repair policies. Doi: 10.28991/CEJ-2024-010-09-012 Full Text: PDF

Floods are one of the most extreme events that affect human life in different regions worldwide. In India, North Bihar is one of the regions affected by floods every year during the monsoon season. Every year, floods cause the loss of human and cattle lives. It has also become the reason for enormous economic loss. Damage to infrastructure is a significant cause of such economic loss. Every year, transportation infrastructure gets inundated due to damage to the pavement during severe flood conditions. Many researchers have investigated the impact of floods on pavement structures. They have reported the different factors significantly influencing the pavement during flood conditions. Different forms of pavement failure take place due to this. Pavement deterioration models were developed to predict pavement damage. None of the deterioration models are suitable for all regions. The suitability of deterioration models depends upon conditions in the selected region. In this study, the impact of floods on the pavement in North Bihar is investigated. The different forms of pavement failure were identified and their extent was analyzed. Also, a mathematical model was developed to predict different forms of failure due to floods in North Bihar. It was observed that the height of flood has significant impact on the different parameters considered in this study.

Prediction of the fundamental viscoelasticity of asphalt mixtures using ML algorithms

Pavement performance prediction is the basis for maintenance decisions. Predicting future pavement conditions accurately and efficiently helps determine the optimal maintenance time, select appropriate measures, and allocate rehabilitation funds effectively. However, limited to the instability and variability in pavement condition data collection, deterministic models are not always reliable for all pavement situations. On the other hand, probabilistic-based models are influenced by environmental factors that are challenging to quantify. Recognizing the limitations of the above two methods, this paper proposes a cumulative distribution-based technique for developing pavement performance prediction models. First, after comparing performance metrics such as pile-by-pile single-point, probability density, and cumulative distribution, it is evident that the cumulative distribution is the most reliable method for describing pavement conditions. A continuous distribution function is created from a limited set of discrete observed field pavement condition data using the sampling theorem. With cumulative distribution-based deterioration curves changing over time, it is possible to predict future pavement deterioration rates. A case study is presented at last. Analyses of the predicted curve and observed pavement performance indicate that the cumulative distribution-based technique is effective in modeling pavement performance and can provide reliable predictive results.

Pavement distresses, such as cracks and ruts, reduce pavements’ effectiveness and serviceability and can lead to failure. This underlines the importance of predicting pavements’ deterioration in pavement management systems (PMSs) for effective maintenance and rehabilitation (M&R) strategies. Consequently, it is essential to understand the concept of service life, which represents how long a pavement will remain in service based on how reliable it is. This study introduces a pavement deterioration model using data from the Long-Term Pavement Performance program for the international roughness index (IRI) and other factors. Different machine learning methods were utilized in developing the model to incorporate eight factors that significantly affect pavement roughness; these methods are: linear regression, regression tree, Gaussian Process Regression (GPR), Support Vector Machine (SVM), Ensemble Trees, and Artificial Neural Network (ANN). For comparison, the models’ performances were evaluated using Root Mean Squared Error (RMSE), Mean Squared Error (MSE), and R squared (R2). The weights and biases of the best model and the Federal Highway Administration (FHWA) recommended IRI ranges were utilized to create the limit state function. A reliability analysis using Monte Carlo Simulation (MCS) was determined to calculate the sections’ probability of failure. This study concluded that pavement sections in the US and Canada are reliable and that the mean yearly Kilo Equivalent Single Axle Load (KESAL) significantly contributes to pavement failure. Keywords: Pavement deterioration, IRI, Machine learning, Neural network, Pavement reliability, Probability of failure.

Roadway and highway agencies across the globe spend a sizable fraction of their annual budget for the upkeep and maintenance of roadways. Different road segments deteriorate at different rates owing to variable traffic flow along the segments. In previous works, various forms of mathematical formulations were provided for roadway maintenance and pavement deterioration modeling. Numerical solutions algorithms using linear programming, dynamic programming, and genetic algorithms were proposed. The solution algorithms, however, did not benefit from the prescriptive and predictive capabilities of machine learning (ML) algorithms (e.g., random forest classifier, support vector machine, and artificial neural networks). Furthermore, previous methods treated transition probabilities of condition states of a pavement in future years to be static. In this paper, a variable transition probability is introduced based on the deterioration rate of a pavement over time. A modified capacitated arc routing formulation is developed for a highway infrastructure management information system. Prescriptive and predictive analytics are performed using ML to analyze the road network in simulation studies and from Montgomery County, Maryland, USA. The pavement condition index (PCI) for the road network is predicted using ML algorithms. The results show a good promise for PCI prediction based on variable deterioration rate and for obtaining condition states in future years subject to varying transition probabilities. Received: 2 November 2023 | Revised: 12 December 2023 | Accepted: 25 December 2023 Conflicts of Interest Manoj K. Jha is an Associate Editor for Journal of Computational and Cognitive Engineering and was not involved in the editorial review or the decision to publish this article. The author declares that he has no conflicts of interest to this work. Data Availability Statement Data sharing is not applicable to this article as no new data were created or analyzed in this study. Author Contribution Statement Manoj K. Jha: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization.