Steel slag has been commonly used in road engineering as cementitious material; however, its application in base course is not widely reported. Four contents of steel slag (0%, 30%, 50%, 75% by volume) were blended into different cement (3%, 4%, 5%, 6% by weight)-treated aggregates. Mechanical properties, volume stability, economic benefits and environmental influences of steel slag mixtures were investigated for the feasibility of applying steel slag in semi-rigid base course. Abrasion, crushing and elongated particle content were compared against limestone aggregate, showing that steel slag has the potential of replacing natural aggregate in concrete. Steel slag is beneficial for reinforcement of the strength and stiffness. The mixture has the highest strength and stiffness when bended with 50% steel slag at 4% cement content. By treating steel slag with CH₃COOH or adding silica fume, volume expansion of steel slag can effectively be controlled. Larger size steel slag (>4.75 mm) and higher cement content are recommended due to heavy metal leaching risk, especially in salty humid areas. Steel slag has sound economic benefits due to the relatively low price. Environmental benefits can also be achieved given that the transport CO_2eq emission of steel slag is accounted for. With proper control in production process, steel slag is a very promising alternative material to be utilized in cement-stabilized base course in road engineering.

The design of asphalt mixture has, for a long time, been an empirical and proof process, causing the mismatch between material design and pavement structure design. To enhance the rationality of asphalt pavement design, this study seeks a path to bridge the gap between asphalt mixture modulus and structural behavior. Firstly, pavement models with different base rigidities, including cement concrete base, cement-treated granular base, and granular base, were constructed to calculate the pavement responses under different dynamic modulus master curve parameters. The influence of master curve parameters on critical pavement responses was identified by the response surface method (RSM). Furthermore, a Whale Optimization Algorithm–Back Propagation (WOA-BP) artificial-neural-network-based pavement response prediction model was established. Then, a database mapping over 100 thousand pavement responses and dynamic modulus master curve parameters was built for determining the dynamic modulus master curve parameters by optimizing the pavement responses. The results show that the impacts of dynamic modulus master curve parameters on critical pavement responses depend on pavement structures. In general, parameter δ has the greatest impact, followed by α, while the effects of β and γ are relatively small. The Artificial Neural Network (ANN) performance prediction model, optimized by the WOA algorithm, has a high accuracy. The methodology for determining the dynamic modulus master curve parameter based on the critical response of pavement was successfully implemented. The findings can bridge the gap between material design and structure design of asphalt pavement and provide a basis for more accurate and reasonable asphalt pavement design.