基于混合多目标粒子群算法的飞行器气动布局设计  

Aerodynamic Configuration Design of Aircraft with Hybrid Multi-objective Particle Swarm Optimization

作　　者：王允良[1] 李为吉[2] 

机构地区：[1]海军航空工程学院飞行器工程系,山东烟台264001 [2]西北工业大学航空学院,陕西西安710072

出　　处：《航空学报》2008年第5期1202-1206,共5页Acta Aeronautica et Astronautica Sinica

基　　金：国家自然科学基金(10377015)

摘　　要：为了提高多目标优化算法求解非劣解集的效率,在多目标粒子群算法的基本框架中引入了Pareto过滤算子、小生境技术和模拟退火算法,建立了全新的混合多目标粒子群算法。该算法具有运算收敛快,所得非劣解集分布均匀、广泛的特点。将其应用于求解以升阻比和效用体积最大化为目标的再入式高超声速飞行器气动布局多目标优化设计模型,将计算结果与原始多目标粒子群算法的计算结果进行对比,体现出本文提出的混合多目标粒子群算法能够更加有效地求解复杂多目标优化设计问题的非劣解集,从而为多目标决策提供有力的支持。A new hybrid multi-objective particle swarm optimization （MOPSO） is built in order to improve the searching efficiency and keep the diversity of noninferior set. Pareto operator, niche technique and simulated annealing method are incorporated into the new hybrid MOPSO. The algorithm is applied to the multi-objective optimization of aerodynamic configuration design of a re-entry hypersonic vehicle with respect to lift-to-drag ratio and volumetric efficiency given the constraint of pitching stability. The optimal solutions are compared with those generated by the basic MOPSO. It is shown that the new hybrid MOPSO can produce a robust and well distributed noninferior set for complex multi-objective design projects, which can help designers understand their design project and make decisions.

关 键 词：粒子群优化算法 小生境技术 模拟退火算法 多目标优化 气动布局设计 

分 类 号：O221[理学—运筹学与控制论] V221[理学—数学]