摘要
针对现有模糊聚类方法在粒度划分实现过程中存在问题,基于构建面向配置设计的产品功构模型,通过对模糊C均值算法中随机初始化策略进行改进,研究功构单元粒度划分方法,并结合产品配置设计过程中需求映射及实例映射特点,根据公理化设计原理探讨了基于信息熵的聚类评价函数构造。最后,结合项目实施给出该方法在电梯制造业中的典型应用实例。系统实现及结果表明提出方法的有效性,为面向配置设计的产品族构建提供另一种数值计算方法。
Aiming at problems existed in the traditional granularity division method realized by the fuzzy cluster means, product function and structure unit granularity division method is researched by adjusting the random initialization tactic of the fuzzy cluster means, which is based on the product function and structure model oriented to the configuration design. According to the Axiom design theory, the cluster evaluation function is constructed with the characteristic of requirement mapping and instance mapping during the configuration design process. Finally, a typical concrete instance is given related with the project of elevator design, which shows the validity and promise of the method mentioned above. Therefore it provides another numeric tool of product family construction for configuration design.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2006年第11期109-114,共6页
Journal of Mechanical Engineering
基金
国家自然科学基金(50505044
50405037)
关键词
粒度划分
功构单元
模糊C均值算法
配置设计
公理化设计
Granularity Product function-and-structure unit Fuzzy cluster means Configuration design Axiomatic design