Optimization design of uav wing structure based on high-dimensional multi-objective integer optimization

Abstract

The structural performance of UAV wings is critical to flight safety and payload capacity. Optimizing UAV wing structures requires simultaneously minimizing total mass and maximum deflection under strength constraints, which constitutes a computationally expensive high-dimensional multi-objective problem. For high-aspect-ratio fixed-wing UAVs, this paper proposes a surrogate-assisted optimization method based on the NSGA-III framework. A Kriging surrogate model is employed to approximate the objective functions. The Expected Hypervolume Improvement (EHVI) criterion guides sequential sampling, and importance sampling is introduced to accelerate EHVI evaluation. Additionally, a diversity-preserving strategy based on Pareto front distance is incorporated to maintain solution diversity. Taking a UAV wing as a case study, the structure is parameterized into 25 thickness variables, and a dual-objective optimization is performed using ANSYS static analysis. Numerical results show that the proposed method generates a well-distributed Pareto front. The selected trade-off solution (Solution B) reduces total mass by 27.35% and maximum deflection by 11.91%, demonstrating the effectiveness and engineering applicability of the method.