Artificial intelligence-based gene expression programming (GEP) model for assessing sprayed seal performance

This research predicts residual solvent (alpha), which is a key component of the performance assessment for a sprayed/chip seal. In this study, conventional equations for alpha were assessed that showed prediction inefficiency (R-2 value as low as 0.82) under different experimental conditions. Accordingly, gene expression programming (GEP), an emerging branch in artificial intelligence, was utilised to resolve these difficulties by developing empirical models for alpha. The data required for model development was obtained from extensive laboratory tests conducted on bitumen-solvent binder films in this research. Model evaluation results showed an excellent degree of correspondence between predictions and experimental results (R-2 = 0.94). This is the first study to model a key component of sprayed seal performance using GEP. The model is recommended for pre-design purposes or as a tool to determine residual solvent in a sprayed seal when laboratory testing is not feasible, thereby saving time and expenditure.

Automated summary

This research developed an empirical model using gene expression programming (GEP) to predict residual solvent (alpha) in sprayed seal performance assessment. The model showed a good correspondence with experimental results and can save time and expenditure in laboratory testing.