What question did this study set out to answer?

Develop a dual-driven framework that integrates physical constraints with machine learning to predict asphalt mixture dynamic modulus with high accuracy and engineering rationality.

August 15, 2026Engineering Applications of Artificial IntelligenceOpen Access

Physics-informed machine learning for asphalt mixture dynamic modulus: Balancing high accuracy with stability and rationality

View Full Paper

Ask AI

Bookmark

Authors

JSJiahao ShengZhejiang Chinese Medical University LLLi LiuJiangnan University ZLZhaohui LiuChangsha University of Science and Technology

Discussion

Member takes

Implication

Machine learning study demonstrates high predictive accuracy and stability for dynamic modulus in asphalt mixtures, suggesting robust physics-guided optimization for sustainable pavement design.

Key Points

Develop a dual-driven framework that integrates physical constraints with machine learning to predict asphalt mixture dynamic modulus with high accuracy and engineering rationality.
Quantified twelve material composition parameters using Grey Relational Analysis and established a sensitive feature space using K-Nearest Neighbors, Gradient Boosting Decision Trees, and Random Forest algorithms.
Embedded physical correlation rules concerning softening point and asphalt type into a Backpropagation Neural Network, validating performance through multi-error cross-validation and stability metrics.
Void volume and asphalt aggregate ratio emerged as the most dominant drivers of dynamic modulus, while mineral powder apparent density and asphalt specific gravity diminished generalization.
Achieved R² > 0.96 and narrowed prediction confidence intervals, despite minor increases in root mean square error (41–46 MPa) and mean absolute error (28–29 MPa) relative to purely data-driven baselines.

Cite This Study

Sheng et al. (2026) studied this question.

synapsesocial.com/papers/6a80194675c2e31742c85527 https://doi.org/10.1016/j.engappai.2026.115997

View Full Paper

Ask AI

Bookmark

Implication

Machine learning study demonstrates high predictive accuracy and stability for dynamic modulus in asphalt mixtures, suggesting robust physics-guided optimization for sustainable pavement design.

Physics-informed machine learning for asphalt mixture dynamic modulus: Balancing high accuracy with stability and rationality

Discussion

Member takes

Implication

Key Points

Cite This Study

Physics-informed machine learning for asphalt mixture dynamic modulus: Balancing high accuracy with stability and rationality

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider