Customer requirements are divided into two types. One is ordinary requirements and the other is customized requirements. They play different roles for determining product structure in custom design. According to their different roles, as well as integrating the respective advantages of RS theory and ANN,a novel approach of agile customization design based on rough set and multi-ANN (MANN) models was proposed.In this method, RS was applied to reduce and extract the effective customer

requirements;ANN was applied to establish the matching relation between reduced customer requirements and corresponding product structure.The renewable MANN models

were formed from which a appropriate network could be selected according to different customer requirements. Consequently,the perfect design scheme could be obtained

quickly. Finally, a roll plate machine’s customization design was taken as an example to validate the validity of the proposed approach. 

agile customization design, customer requirement, rough set(RS), artificial neural network(ANN)

客户需求分为普通需求和定制需求,根据这两类需求对确定产品结构所起的不同作用，结合粗糙集（RS）和人工神经网络（ANN）方法的优势，提出一种基于粗糙集和多ANN模型的产品敏捷定制设计方法。该方法采用RS提取有效的客户需求;运用ANN技术，建立客户需求与产品结构间的选配关系，形成可不断更新的多个ANN模型。根据不同的客户需求，选取与之匹配的网络模型，快速确定合适的设计方案。以卷板机的定制设计为例，验证了该方法的有效性。

 

敏捷定制设计, 客户需求, 粗糙集, 人工神经网络