Read the papers.
Find the gap.
Write the paper.

1. 1Machine learning prediction of asphalt pavement rutting across climate zones
2. 2Do data-driven rutting models transfer? Evidence from two climates
3. 3Training data, not algorithms: limits of ML rutting prediction

• RQ1. How accurately do ensemble models predict rut depth from traffic, climate and binder data?
• RQ2. How much accuracy is lost when a model is applied to a different climate zone?

• Optimizing asphalt mix design through predicting the rut depth of asphalt pavement using machine learning Construction and Building Materials · Elsevier · 2022 Indexed78 cites
• Creating rutting prediction models through machine learning techniques utilizing the Long-Term Pavement Performance database Sustainability · MDPI · 2023 Indexed68 cites
• Modeling the rutting performance of asphalt pavements: a review J. Infrastructure Preservation and Resilience · Springer Nature · 2023 Indexed45 cites
• Predicting rutting development using machine learning methods based on RIOCHTrack data Applied Sciences · MDPI · 2024 Indexed19 cites

30 papers selected · + 26 more

Design

Comparative modelling study with a cross-climate test

Data

Pavement sections from two climate zones (Alnaqbi et al., 2023)

Models

Random forest, gradient boosting and a neural network (Cheng et al., 2024)

Evaluation

RMSE and R² within a zone and on the held-out zone

Validity

Same features and tuning budget for every model

1. 1Introduction 450 w
2. 2Literature Review 900 w
3. 3Methodology 800 w
4. 4Results 700 w
5. 5Discussion 450 w
6. 6Conclusion 200 w

3,500 words · 30 papers assigned to sections

Machine-learning models of rut depth have been trained on long-term field data and on full-scale test tracks (Alnaqbi et al., 2023; Cheng et al., 2024). Most are evaluated on the dataset they were trained on, which leaves open whether their accuracy holds in a different climate or road network.

Scores for guidance only. Read and revise before you submit.

• MajorJustify the choice of the two climate zones and report their temperature ranges.
• MinorGive the tuned hyper-parameters of each model in a table.

1. Alnaqbi, A. J., Zeiada, W., Al-Khateeb, G. G., Hamad, K., & Barakat, S. A. (2023). Creating rutting prediction models through machine learning techniques utilizing the Long-Term Pavement Performance database. Sustainability, 15(18), 13653.
2. Cheng, C., Wang, L., Zhou, X., & Wang, X. (2024). Predicting rutting development using machine learning methods based on RIOCHTrack data. Applied Sciences, 14(8), 3177.
3. Zhao, J., Wang, H., & Lu, P. (2023). Machine learning analysis of overweight traffic impact on survival life of asphalt pavement. Structure and Infrastructure Engineering, 19(5), 606–616.