Predictive modeling of rutting depth in modified asphalt mixes using gene-expression programming (GEP): A sustainable use of RAP, fly ash, and plastic waste

Annually, billions of tons of fly ash, waste plastic, and reclaimed asphalt pavement (RAP) are produced and pollute our environment. The best way to use these waste products in the construction industry is to use them on those projects that are in excess and have shorter life spans, i.e., highways and roads. To enhance the usage of these waste products in hot mix asphalt (HMA), a sophisticated machine learning approach named gene-expression programming (GEP) has been opted in this study to predict its rutting depth. The developed prediction model relates the rutting depth to ten effective parameters. The modified asphalt samples were prepared by varying the percentages of plastic-ash composite (PAC) (0 %, 50 %, 0.75 %, and 1 %) and RAP (0 %, 20 %, 30 %, and 40 %) by weight of the mix. The experimental results indicate that as the percentage of PAC and RAP increases the rutting depth decreases, and upon addition of 40 % RAP and 1 % PAC the rutting depth was decreased to 80 % compared to the non-modified mix. The accuracy of the predictive capacity of the developed GEP model is accessed using various statistical indices such as R, MAE, RSE, NSE, ρ, and OBF. A second level of validation was carried out by performing sensitivity and parametric analysis, which produced a comparable variation and also indicated the contribution of input parameters to rutting.