混和遗传算法及其在亚音速翼型优化中的应用

A Hybrid Generic Algorithm and its Application in Optimization of a Subsonic Airfoil

为了降低遗传算法中适应度值的计算量,提出了一个新的混合算法,引入了聚类分析和下山单纯形局部寻优两个算子,聚类分析评估种群在欧氏测度上的分布特点后,给出合理的局部寻优空间,有效利用了下山单纯形算法的局部收敛特性,实现了算法的融合。在以升阻比最大为目标的RAE2822亚音速翼型单点设计中,分别采用传统GA算法和新算法进行优化设计,结果表明,新算法有效加快了计算效率,增强了算法对设计空间的挖掘能力。

To decrease the heavy computational cost of Fitness function in Genetic Algorithm, this paper advanced a new hybrid method, in which cluster analysis and downhill simplex local search was incorporated as two additional operatorsl Cluster analysis operator evaluates the diversity and distribution characters of the population according to Euclidian measure in the hyper cubic space constructed by design variables, and setup reasonable sub space for local optimization, thus the local convergence ability of Downhill Simplex could be used aggressively, and the two algorithms fused together. A single point transonic airfoil design started from the RAE2822 with a desire to achieve maximum lift- drag ratio was successfully deployed with traditional GA, and Hybrid GA. A compare between the obtained optimal solutions proved that the measure in this paper bring acceleration in the computational speed and overwhelm in the dig ability of top designs.