基于非对称模糊优化模型的多学科协同满意优化  

Multidisciplinary satisficing collaborative optimization using asymmetrical fuzzy model

作　　者：陶冶[1] 黄洪钟[2] 刘志杰[1] 吕显强[3] 

机构地区：[1]大连理工大学精密与特种加工教育部重点实验室,辽宁大连116024 [2]电子科技大学机械电子工程学院,四川成都610054 [3]大连水产学院信息工程学院,辽宁大连116023

出　　处：《大连理工大学学报》2007年第5期671-677,共7页Journal of Dalian University of Technology

基　　金：国家自然科学基金资助项目(59685003);高等学校全国优秀博士学位论文作者专项基金资助项目(200232);教育部优秀青年教师资助计划资助项目(1766)

摘　　要：协同优化(CO)作为一种多学科设计优化(MDO)方法,包括两级优化:系统级和若干个学科级.然而工程设计实践中往往存在很多不确定因素,而且互相冲突的多个优化目标的存在,进一步增加了CO建模的求解复杂性.采用模糊截集水平法为各个学科设计了非对称CO模型,扩展了学科分析模块,使学科级能独立获取本学科最优的截集水平,以构造相应决策空间.利用满意度原理设计了学科级优化的目标函数,并将其作为一致性方程的组成部分传递到系统级.系统级在满足一致性约束的同时,最大化综合满意度方程,使所有学科同时获得最满意的优化解.电子封装问题的多学科协同满意优化实例表明该方法是行之有效的.Collaborative optimization （CO） in multidisciplinary design optimization （MDO） consists of system-level and several discipline-levels. However, in practical engineering design, lots of fuzzy factors, especially, the mutual conflicting objectives further increase the complexity of CO. The cuts level method is used to design the asymmetrical fuzzy model for CO, which guarantees each discipline to acquire its optimum cuts level independently and constructs the relevant decision space. At the same time, the satisfaction degree is used to design the discipline objective function, which is used as the component of the discrepancy equation and is delivered to system-level. As system-level maximizes the overall satisfaction degree function, each discipline can acquire the optimum satisfaction solutions, simultaneously. The example of electronic packaging problem demonstrates that the proposed method is available and efficient.

关 键 词：协同优化 满意度 多学科设计优化 截集水平法 

分 类 号：TH122[机械工程—机械设计及理论]