Prediction and Optimization of Bituminous Mixture Performance Using Response Surface Methodology

Bituminous mixtures are fundamental to flexible pavement systems, engineered to provide essential durability and resilience against mechanical stresses, including traffic loads, permanent deformation (rutting), cracking, and environmental impacts. Traditional test methods that predict these performances involve extensive experimental testing which is time-consuming, resource-intensive, and requires costly repeats. This study uses the response surface methodology (RSM) to develop predictive models for estimating and optimizing the key performance parameters specifically rutting and cracking in Marshall-designed bituminous mixtures. This study analyses 2,940 bituminous mixtures with design factors including aggregate gradation, aggregate type, air voids, binder grade, and mixture type. Results show that air void content significantly influences the performance, with higher air voids increasing permanent deformation but enhancing cracking resistance. Finer aggregate gradations require more binder content improving cracking tolerance while reducing rutting resistance. Higher binder grades like viscosity graded (VG-40) and polymer-modified binder (PMB-40) boost the durability of the mixture, while aggregate and mixture types have moderate to minimal effects. Further, statistical analyses including analysis of variance (ANOVA), confirmed that air voids, aggregate gradation and binder grade are critical in optimizing performance. Predictive models developed in this study achieved high accuracy ( R 2 ≥ 0.78 ), providing reliable tools for balancing permanent deformation and cracking resistance in bituminous mixtures. Also, the study finds that a balanced performance (optimal trade-off between rutting resistance and cracking resistance) can be achieved by managing air voids within a moderate range of 4.0% to 5.0% using granite as the aggregate type and selecting PMB-40 as the binder type and G1 (midgradation) as aggregate gradation depending on the specific requirements of the mixture type required for mix design for optimal performance. This study provides a reliable and efficient framework for designing durable, high-performance bituminous pavements.