Abstract: To actualize the flexible and rapid design for complex products of customization,a linguistic multiple attribute group decision model of maximum deviation based on Quality Function Deployment (QFD) and the design model based on Artificial Immune System (AIS) were proposed.To translate the customer fuzzy requirements into engineering characteristics of customized products design,the semantics of customer requirements were analyzed and evaluated,and the output engineering characteristics of House of Quality (HOQ) were regard as antigen and described with gene coding through the improved information entropy immune algorithm.The self-learning and self-adaption were realized through the crossover and mutation of AIS.By updating memory cell bank,immune response was achieved,and the best customized product design was acquired.The customized CNC machine design was taken as an example to translate the personalized customized requirements into engineering characteristics through HOQ,and these characteristics were described with antigen encoding.The immune response for antibodies in database was occurred,which could design CNC machine with maximum customer satisfaction accurately,efficiently and intelligently.The feasibility and effectiveness of the design based on QFD and AIS was validated.

Key words: complex products customization, customer requirements, engineering characteristics, quality function deployment, artificial immune system, product design, CNC machine

摘要： 为实现复杂产品定制设计的柔性化和智能化,提出了基于质量功能配置的最大离差语言多属性群决策模型和基于人工免疫系统的复杂产品定制设计模型。对客户语义进行分析和评估,将客户模糊需求转换为定制产品设计的工程特性;利用改进的信息熵免疫算法对质量屋输出的工程特性进行抗原基因编码,通过交叉和变异实现人工免疫系统的自学习、自适应,更新记忆细胞库,对抗原进行免疫应答,获取最佳定制产品设计方案。以定制数控机床设计为例,客户个性化需求经质量屋转换为工程特性,并对其进行抗原编码,检索数据库中的抗体以实现免疫应答,从而准确、高效、智能地设计出客户综合满意度最大化的数控机床,验证了所提方法的可行性和有效性。

关键词: 复杂产品定制, 客户需求, 工程特性, 质量功能配置, 人工免疫系统, 产品设计, 数控机床