Abstract Shock buffet is one of the common flow instability problems in aviation engineering, which can affect flight handling quality and limit flight envelope. In this study, we propose a physics-guided aerodynamic optimization design framework to improve the buffet onset in the transonic region. The buffet instability mechanism correlates strongly with both the extent of flow separation and the spatial evolution of shock waves in the underlying steady flow. Therefore, this optimization framework is based on the deep deterministic policy gradient (DDPG) algorithm, which focuses on optimizing the shock wave position and separation area from the stable flow, while comprehensively considering various aerodynamic constraints for airfoil optimization. The proposed method is applied to two airfoil optimization parameter spaces: 5% and 15% variation ranges. The results indicate an increase in the shock buffet onset of the optimized NACA0012 and RAE2822 airfoils by 0.3° and 0.7°, respectively, along with improving aerodynamic performance. Furthermore, near the leading edge of the optimized airfoil, there is a decrease in the curvature of the upper surface and the leading edge nose radius, while the curvature of the lower surface increases. These changes primarily influence the shock wave position and the peak suction value at the leading edge.
Ma et al. (2026) studied this question.