论文部分内容阅读
提出了用于加速气动力形状优化过程的分级型Nash基因算法.分级型算法可以看作是并行基因算法的特例,后者使用了互相联系但独立进化的子群的概念.本文在并行基因算法中引入多层分级拓扑结构以提高算法的收敛性.这种拓扑结构混合使用不同精度的模型,低精度模型用于探索搜索空间,高精度模型用于对准优解进行提纯.将此方法与Nash博弈相结合,构造了多目标优化算法,并应用于气动力优化问题.针对喷管反设计问题与多段翼型高升力优化问题,在计算机集群并行环境下进行了计算,结果表明本文的算法具有较高的加速收敛特性.
A hierarchical Nash algorithm is proposed to accelerate the aerodynamic shape optimization process.Hierarchical algorithm can be regarded as a special case of parallel genetic algorithm which uses the concept of interrelated but independently evolved subgroups.In this paper, A multilevel hierarchical topology is introduced to improve the convergence of the algorithm.This topology uses a mixture of different accuracy models, a low-precision model to explore the search space, and a high-precision model to align the optimal solution to be refined.This method and Nash game, the multi-objective optimization algorithm is constructed and applied to the aerodynamic optimization problem.According to the counter-nozzle design problem and the multi-section airfoil high-lift optimization problem, the calculation is carried out under the computer cluster parallel environment. The results show that the algorithm Has a higher acceleration convergence.