基于层次分析法的MDO算法性能评价

在互补判断矩阵排序理论的基础上,结合最优化方法,将互补判断矩阵特征向量法引入层次分析中,提出了一种基于互补判断矩阵特征向量法的层次分析方法。通过对几种常见MDO算法的设计变量、约束、计算精度和计算复杂度进行分析,给出了改良的序列二次规划算法(NLPQL)、自适应模拟退火算法(ASA)、多岛遗传算法(MIGA)和序列二次规划算法(DONLP)在涡轮叶片多学科优化中选择的实例说明,并与实际优化结果进行对比,验证了本文所采用方法的有效性。

MDO Algorithm Performance Evaluation Based on Analytic Hierarchy Process

Using the priority theory of complementary judgement matrices and the optimization method,the paper aims at proposing the analytic hierarchy process(AHP) method based on the eigenvector method of the complementary judgement matrices.Through analyzing the numbers of design variables and constraints,global convergence calculation accuracy and complexity of several common multiple discipline optimization(MDO) algorithms,it presented the calculation instances of the MDO algorithm selections among the improved sequential qradratic programming(SQP) algorithm,the adaptive simulated annealing(ASA) algorithm, the multi-island genetic algorithm(MIGA) and the SQP algorithm during the MDO of turbine blades of an aeroengine.Finally it compared their optimization results using the four MDO algorithms,and the comparison results show that the present AHP method is effective in seleeting MDO algorithms(see Table 1 in the full paper).