基于遗传算法的协作优化方法研究  

Study of Collaborative Optimization Based on Genetic Algorithm

作　　者：蒙文巩[1] 黄俊[1] 

机构地区：[1]北京航空航天大学航空科学与工程学院,北京100083

出　　处：《航空计算技术》2006年第6期24-28,共5页Aeronautical Computing Technique

摘　　要：多学科设计优化技术是国内外迅速发展的一门学科,而协作优化方法是多学科设计优化技术中最常用的一种有效的分布式并行优化方法。然而,标准的协作优化方法很难收敛到系统的最优解,而且对初始点的选择非常敏感,不同的初始点可能收敛到不同的解。遗传算法是一种成熟的智能寻优算法,具有很强的全局搜索能力。本文正是为提高协作优化的可靠性和鲁棒性,将遗传算法引入协作优化框架,并且对传统遗传算法进行相应的改进以适用于协作优化框架。算例证明本文提出的基于遗传算法的协作优化方法行之有效,消除了标准协作优化可能收敛不到系统最优解及对初始点敏感的缺陷。Multidisciplinary Design Optimization （MDO） is a methodology developed rapidly at home and abroad. And Collaborative Optimization （CO） is an efficacious distributed concurrent optimization method widely applied in MDO. Yet, standard Collaborative Optimization can hardly get systemic optimal value, and is sensitive to the selection of primary points. Different primary points would gain different optimal values. Genetic Algorithm （GA） is a mature intelligent optimization algorithm with powerful global searching ability. This paper just aims to improve the reliability and robustness of CO, by applying GA to CO framework and improving the traditional genetic algorithm in order to fit the CO framework. The example demonstrates that CO based on GA disposes of the defect that the standard CO probably would not gain the systemic optimal value and be sensitive to primary point.

关 键 词：多学科设计优化 协作优化 遗传算法 

分 类 号：TH122[机械工程—机械设计及理论]