应用博奕理论的多目标分布式气动优化设计  

Distributed Optimization Using Virtual and Real Game Strategies for Aerodynamic Multi-objective Design

作　　者：胡世祥[1] 李磊[2] 佘春东[2] 唐智礼[3] 

机构地区：[1]中国人民解放军总装备部,北京100720 [2]中国科学院软件研究所,北京100080 [3]南京航空航天大学空气动力学系,江苏南京210016

出　　处：《航空学报》2004年第4期321-326,共6页Acta Aeronautica et Astronautica Sinica

摘　　要：应用博奕理论研究了最优气动外形优化设计方法。采用的基本的优化器为基于控制理论的约束优化设计方法,通过引入标量伴随变量来执行约束条件,进而将上面的方法和对策论结合起来处理多目标翼型优化设计问题,根据各种不同的气动标准(可以是互为冲突的)来优化翼型形状。在对称Nash策略中,每一个"player"都力图优化自己的目标,而Nash平衡则提供了多个目标之间的一种解。不同的翼型分裂和设计算例表明了本文虚拟和真实Nash竞争策略在多目标翼型优化设计中的能力。成功的设计结果表明了本文方法的有效性。The question of multi-objective optimization for optimum shape design in aerodynamics via constrained control theory is described. The optimizer is based on control theory. A scalar adjoint variable to implement the constraints is introduced. The above method is combined with a formula derived from game theory to treat multi-point airfoil optimization. Airfoil shapes are optimized according to various aerodynamics criteria (in conflict). In the symmetric Nash Game, each player is responsible for one criterion. Several kinds of equilibrium provide a solution to the multi-point optimization. Several kinds of airfoil splitting and design cases are shown for virtual and real game strategies in aerodynamic design, and design results confirm the validity and efficiency of the present design method.

关 键 词：博奕理论 多目标优化设计 分布式优化 最优控制理论 约束优化 

分 类 号：V211.41[航空宇航科学与技术—航空宇航推进理论与工程]