Integrating Extreme Vertices Design and Genetic Algorithm to Enhance the Moisture Resistance of Bituminous Concrete Modified with Recycled High Density Polyethylene

This study integrates Extreme Vertices Design (EVD) and Genetic Algorithm (GA) to enhance the moisture resistance of bituminous concrete modified with recycled high-density polyethylene (HDPE), using tensile strength ratio (TSR) as the performance indicator. EVD is applied to design the mixtures, perform sensitivity analysis, and develop regression models, which are validated using F-statistics, coefficient of determination (R²), and mean absolute percentage deviation (MAP.D). GA is then employed to optimize the mixture composition within the constrained design space to further improve TSR. The TSR of HDPE-modified mixtures ranges from 82.73% to 95.82%, all exceeding ASTM and AASHTO minimum requirements and outperforming the unmodified mixture (82.43%), with several mixtures achieving TSR values above 85%. Analysis reveals that granite content positively influences TSR, while excessive sand and bitumen have detrimental effects, and HDPE improves TSR only up to an optimum content beyond which mixture cohesion declines. ANOVA highlights significant interactions among granite, bitumen, and HDPE, and the developed regression model exhibits strong predictive capability (R² = 93.27%, MAP.D = 0.74%). GA optimization yields a maximum TSR of 97.88% with an optimal blend of granite, sand, bitumen, and HDPE, demonstrating that recycled HDPE can substantially improve both performance and sustainability of asphalt mixtures while satisfying international standards.