摘要
提出了联合采用聚类分析、模糊识别、基于实例的推理进行产品结构设计的方法,基本思想如下:①任一“产品超类”都可以按照“产品-部件-零件”的层次进行划分,由此建立一棵“树”。②一个“产品类”拥有许多以“参数”表达的技术属性。采用聚类的方法,可以把每个属性映射到“树”的某个结点,构成该结点的属性集,然后基于这个属性集、利用已有的产品实例建立一个对应于该结点的模糊模型。③采用模糊识别的方法,按照“树”的拓扑结构对当前待求的设计问题进行识别,实现产品在概念设计阶段的结构设计。④采用CBR方法,实现详细设计阶段的结构设计。本文方法适用于大多数机电产品的智能化设计。
A method to design product structure was proposed by integrating cluster analysis, fuzzy recognition and case-based reasoning (CBR). Any product super-class can be divided into three layers in terms of product-component-part to setup a tree. A product class has many technical attributes represented by parameters. Each attribute was mapped with a node of the tree to form an attribute set for this node by cluster analysis. Based on this attribute set, a fuzzy model corresponding to this node was set up by taking advantage of existing product cases. By fuzzy recognition, the problem to be designed was recognized according to the topology of the tree to carry out product structure designing during conceptual design. The detail design can be implemented by CBR. This method can be used in intelligent designing for most of products.
出处
《微计算机信息》
北大核心
2007年第01S期298-300,共3页
Control & Automation
基金
河北省博士基金资助(05547003D-3)