高维多目标设计是目前飞行器气动外形设计方法的难点，优化算法的能力决定着能否在高维目标空间内得到整体性能最优的气动设计方案。本项目拟通过研究基于性能指标高层目标集合粒子群算法进化搜索机制，为集合内部粒子提供全局最优进化方向；基于目标集合当前解集最优进化信息，耦合在线目标降维技术建立多种群协作的动态邻居粒子群搜索模式，构建基于低层气动设计目标的高维多目标粒子群算法，实现Pareto最优解集尽快收敛；针对Pareto解集指数级增长问题，引入基于改进K支配排序的占优机制及基于广义闵可夫斯基距离密度评估策略的精英保留机制，提高Pareto解选择压力，有效改善超高维Pareto曲面收敛性及分布性。本研究将为飞行器气动设计中的高维多目标问题提供新的优化设计思想及解决途径。

Many-objective design is a difficult problem in aerodynamic optimization based on CFD. The overall performance of aircraft aerodynamic design scheme in many-objective space depends on the ability of optimization algorithm. In this research, an indicator-based many-objective particle swarm optimization will be presented to provide the global optimal evolution direction for the inner set particles. Using information of the optimal reference point of each set, the multi-swarm collaboration particle swarm optimization searching modes based on dynamic neighborhood will be established. Then the many-objective particle swarm optimization based on aerodynamic objective will be built up to guide particles to converge towards Pareto optimal solutions as soon as possible. The improved -dominated sorting and the external archive based on Minkowski density estimation strategies is proposed to solve the exponential growth of the Pareto solution set. This selection mechanism will enhance Pareto selection pressure and improve the convergence and distribution of the Pareto surface effectively. This research will provide some new design ideas and solutions for the many-objective problem in aerodynamic design.