A Unified MARS‐NRBO Model for Predicting Dynamic Modulus of Asphalt Concretes and Stone Matrix Asphalts Produced by Hot and Warm Technologies
Advances in Materials Science and Engineering · 2026-01 · 2026 (1)
DOI: 10.1155/amse/5557654 ↗Kurz gefasst
Fachartikel in der Zeitschrift Advances in Materials Science and Engineering, online erschienen am 22.07.2026. Verfasst von Ali Reza Ghanizadeh und 3 weiteren Autor:innen. Die Arbeit nennt 77 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The dynamic modulus (| E ∗ |) is the key parameter to characterize the viscoelastic behavior of asphalt mixtures, making it an essential component of the mechanistic‐empirical pavement design guide (MEPDG). However, direct laboratory measurement of | E ∗ | is time‐consuming and costly, motivating the development of reliable predictive models. This goal is usually achieved by developing and | E ∗ | master curve for each asphalt mixture, covering a vast spectrum of temperatures and loading frequencies. Therefore, if the asphalt mix type and the production technology of asphalt mixture change, different master curves must be developed. This study proposes a hybrid framework that integrates multivariate adaptive regression splines (MARS) with the Newton–Raphson–based optimizer (NRBO) to propose a unique formula to predict | E ∗ | of two types of asphalt mixture (i.e., asphalt concrete (AC) and stone matrix asphalt (SMA) produced by two technology, namely, hot mix asphalt (HMA) and warm mix asphalt (WMA). Four key input variables, namely, mixture type, production technology, loading frequency, and temperature, were used as predictors. The NRBO algorithm was employed to tune critical MARS hyperparameters, including maximum basis functions, penalty parameter, threshold, and maximum interaction terms, thereby enhancing predictive accuracy while mitigating overfitting. Model performance was assessed using various statistical indexes. Results demonstrate that the optimized MARS model achieved outstanding accuracy, with R 2 = 0.9889, NSE = 0.98, a20‐index = 0.9791, and RMSE = 130.198 across the full dataset. The relative importance analysis identifies temperature as the most influential input for predicting | E ∗ |. A parametric study further highlighted the dominant influence of temperature and frequency on | E ∗ |, while SMA‐WMA mixtures consistently exhibited the highest modulus.
Zitieren
APA
Ghanizadeh, A. R., Jahanshahi, F. S., Naseralavi, S. S., & Soltanirad, M. (2026). A Unified MARS‐NRBO Model for Predicting Dynamic Modulus of Asphalt Concretes and Stone Matrix Asphalts Produced by Hot and Warm Technologies. Advances in Materials Science and Engineering, 2026(1). https://doi.org/10.1155/amse/5557654
Vancouver
Ghanizadeh AR, Jahanshahi FS, Naseralavi SS, Soltanirad M. A Unified MARS‐NRBO Model for Predicting Dynamic Modulus of Asphalt Concretes and Stone Matrix Asphalts Produced by Hot and Warm Technologies. Advances in Materials Science and Engineering. 2026;2026(1). doi:10.1155/amse/5557654
BibTeX
@article{Ghanizadeh2026,
author = {Ghanizadeh, Ali Reza and Jahanshahi, Farzad Safi and Naseralavi, Seyed Saber and Soltanirad, Mohammad},
title = {A Unified MARS‐NRBO Model for Predicting Dynamic Modulus of Asphalt Concretes and Stone Matrix Asphalts Produced by Hot and Warm Technologies},
journal = {Advances in Materials Science and Engineering},
year = {2026},
volume = {2026},
number = {1},
doi = {10.1155/amse/5557654}
}
EUVIMED