Abstract
This study proposes a new model for predicting the interlayer bonding strength (IBS) of asphalt pavement using gene expression programming (GEP). A total of 164 double-layered specimens were tested to evaluate IBS, using a customized direct tensile bonding test device by incorporating a shear force application feature. The GEP model was developed based on three parameters, including temperature, shear stress, and tack coat application rate, and R-squared (R2) and root mean square error (RMSE) were used as performance indicators of the model. The results showed that the GEP model could explain over 94% of the measured IBS data (95.7% for training and 94.4% for validation) with a very low error rate (RMSE value of 0.047 MPa in training phase and 0.054 MPa in validation phase), indicating its high accuracy. Additionally, the results of sensitivity analysis indicated that test temperature (T), horizontal shear stress (Hs), and tack coat application rate (TAR) affected the model by 57.11, 29.32, and 13.57%, respectively. The findings of this study have significant implications for informed decision-making in pavement design and maintenance. Accurately evaluating IBS is essential to ensure durable pavement performance, and findings of this study provide valuable insights into achieving this goal.