基于模糊理论的MDO算法性能评价

Evaluation of Capability for MDO Technique Using Fuzzy Logic

多学科设计优化算法的选取其实质是一个多目标综合评价问题。文中依据模糊综合评价的基本原理,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,建立了MDO算法模糊评价选优决策模型。并采用加权欧氏空间距离进行评判,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)在涡轮叶片多学科设计优化中选择的实例说明,验证了本文所采用方法的有效性。

The essence of select multidisciplinary design technique optimization is a multiple objective synthetic evaluation problem.In this paper,considering the number of design variables and constraints,the calculation accuracy and com- plexity and the global convergence,a MDO technique fuzzy synthetic evaluation model is constructed.Using the number of Euclid distance with weight as an evaluation criterion,a practical example,the select of Improved Sequential Quad- ratic Programming(NLPQL).Adapted Simulated Annealing(ASA).Multi—island Genetic Algorithm(MIGA)and Sequential Quadratic Programming(DONLP)in the multidisciplinary design optimization of turbine blades is given. The result shows the efficiency of this method.