基于可拓学的设计方案进化推理方法  

A New Evolution-reasoning Method in Conceptual Design Based on Extension Theory

作　　者：赵燕伟[1] 刘海生[2] 张国贤[1] 

机构地区：[1]上海大学机电工程与自动化学院,上海200072 [2]浙江工业大学机电工程学院,杭州310014

出　　处：《中国工程科学》2003年第5期63-69,共7页Strategic Study of CAE

基　　金：国家自然科学基金项目 ( 5 0 175 10 3);国家"八六三"高技术计划资助项目 ( 2 0 0 2AA411110 )

摘　　要：在分析现有概念设计求解策略的基础上 ,提出了基于可拓学理论与遗传算法相结合的概念设计求解模型。利用遗传算法模拟物元变换过程 ,建立了产品方案物元描述的内部模型和外部模型 ,探讨了内、外模型的转换关系并通过关联函数将内、外部模型联系起来。根据内部模型给出了遗传算法的编码形式以及与之相适应的交叉、变异策略 ,并通过可拓评价方法建立了遗传算法的适应值函数 ,基本解决了产品概念设计中的知识组合爆炸和矛盾冲突问题。最后通过求解减速器方案验证了该方法的可行性。After existing study methods are analyzed, a new researching model for conceptual design based on extension and genetic algorithms was presented. The inner-model and outer-model of product project were set up with matter elements. In genetic algorithms the coding for individual, the means for cross-over and mutation were all founded based on the inner-model. The fitness function was set up combined with relationship function. The technique solves the innovative and incompatible problems in conceptual design. The feasibility of the new researching model is testified by three conceptual design examples for retarder.

关 键 词：概念设计 物元变换 遗传算法 进化推理 可拓评价 

分 类 号：TH122[机械工程—机械设计及理论]