产品概念设计阶段的案例相似性检索技术研究  

Case similarity retrieval technique in product conceptual design

作　　者：谷建光[1] 张为华 王中伟[1] 

机构地区：[1]国防科技大学航天与材料工程学院

出　　处：《计算机集成制造系统》2008年第4期625-629,721,共6页Computer Integrated Manufacturing Systems

基　　金：国防预研资助项目(41328010504)~~

摘　　要：以提高产品设计结果的论域覆盖面和增加可行设计结果的多样性为目的,研究了产品概念设计阶段的基于案例推理的产品设计案例相似性检索技术。首先,对产品设计案例相似性检索进行了问题描述;然后指出了传统相似性检索方法的不足;进而提出一种新的基于划分聚类和一般模糊极小极大神经网络的相似案例检索方法,并建立检索模型;最后,通过基于案例推理的固体火箭发动机总体设计实例表明,与传统方法相比,新方法可有效提高相似结果案例集的论域覆盖面,增加设计结果的多样性,提高检索效率,进而提高了固体发动机总体设计质量。To improve the universe of discourse coverage and the diversity of feasible product design results, similarity retrieval technique in the conceptual product design case was discussed. Firstly, similarity retrieval technique on product design case was described. SecondLy, shortcomings of traditional similarity retrieval technique were pointed out. Thirdly, a new similarity retrieval technique based on partitioned clustering and General Fuzzy Min-Max （GFMM） neural network was proposed, and its model was constructed. Finally, this new technique was compared with traditional technique in the example of Solid Rocket Motor （SRM） system design. Example analysis revealed that the new technique could increase the universe of discourse coverage, improve the diversity of feasible SRM design results, and enhance the retrieval speed so as to improve overall quality of SRM system design.

关 键 词：产品设计 案例相似性检索 划分聚类 模糊极小极大神经网络 固体火箭发动机总体设计 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程] V435[自动化与计算机技术—控制科学与工程]