Optimization of asphalt mix design considering mixture performance, environmental impact, and life cycle cost

High mechanical performance of asphalt mixture used in the surface layers of pavement infrastructure often comes at the expense of increased environmental impact and cost, while eco-friendly alternatives may compromise its durability. Optimizing the mix design and achieving the trade-off between these aspects are critical for cleaner mixture production and longer-performing pavement. This study proposes an integrated framework that combines Balanced Mix Design (BMD), Life Cycle Assessment (LCA), and Life Cycle Cost Analysis (LCCA) to systematically evaluate and balance the competing objectives of asphalt mixture performance, environmental impact, and cost. A deep neural network (DNN) was developed to capture the nonlinear interactions among mix design variables and key indicators, and was coupled with Particle Swarm Optimization (PSO) for multi-objective optimization of mix combinations. To enhance transparency and guide decision-making, Shapley Additive Explanations (SHAP) analysis was applied to identify critical sustainable variables—such as high-temperature performance grade (PGHT), recycled asphalt pavement (RAP) content, and gradation parameters—that significantly affect pavement performance and life-cycle outcomes. The optimized designs demonstrated improved rutting and cracking resistance, while significantly reducing global warming potential, respiratory effect, ecotoxicity, and life-cycle costs. This framework effectively bridges laboratory research with practical implementation, offering a comprehensive and scalable approach to designing durable, cost-effective, and environmentally sustainable asphalt pavements, and contributing to the advancement of AI- and optimization-aided multidimensional and sustainable mix design.