【作者】 经有国；

【导师】 但斌；

【作者基本信息】 重庆大学 ， 管理科学与工程， 2011， 博士

【摘要】 随着消费类机电产品市场竞争的日益激烈和市场需求的不断分化,越来越多的消费类机电产品制造企业开始或已经采用大规模定制模式,通过向消费者提供定制化和个性化的产品以赢得市场竞争优势。本文主要研究了基于Web人机交互的消费类机电产品定制过程中,客户需求信息的获取及处理问题。首先,研究了考虑客户需求表达能力异质的消费类机电产品客户需求信息获取问题。对于某种消费类机电产品(如计算机、摩托车、汽车等)而言,不同客户(或消费者)由于个人学历专业、社会经历以及专业知识等的不同,他们对该产品的熟悉与了解程度有很大差别,导致不同客户的需求表达能力存在较大的异质性,因此,在基于Web人机交互的产品定制过程中,根据不同客户的需求表达能力向其提供相应的需求交互方式是很有必要的。鉴于此,本文提出了一种面向异质客户的需求信息获取方法。该方法通过获取客户的个人特征和需求情境的相关信息,然后应用模糊推理规则推理客户的需要表达能力及初步的产品需求,再根据该客户的需求表达能力情况提供相应的需求交互过程与界面,实现了便捷有效地获取需求表达能力异质的众多客户的需求信息。最后,以踏板摩托车的案例阐述了该方法的有效性和实用性。其次,研究了消费类机电产品的半结构化客户需求信息转换处理问题。对于普通消费者(非专业客户)而言,消费类机电产品的技术和结构都比较复杂,在基于Web人机交互的产品定制过程中,他们难以通过选择产品模块、功能特征及技术参数等选项的结构化方式来表达其需求,他们往往更希望用模糊词汇(或词组)的半结构化方式来描述给定的产品定制特征。为此,本文研究了一种将半结构化客户需求信息转换成精确产品功能要求的方法。提出了半结构化客户需求信息(SCN)的数学描述方法和SCN转换处理问题的形式化描述方法;建立了需求特征与功能特征之间的模糊关系矩阵;再应用模糊推理技术实现了半结构化客户需求信息到产品功能要求的转换处理。最后,以普通骑式二轮摩托车的案例阐述了该方法的有效性和实用性。最后,研究了消费类机电产品的非结构化客户需求信息结构化处理问题。当客户(普通消费者)对消费类机电产品的了解极为有限,在基于Web人机交互的产品定制过程中,通过结构化和半结构化方式表达需求信息存在困难时,若能允许客户使用自然语言短语(或语句)的非结构化方式描述产品要求,将不仅会提高客户对需求交互的满意程度,也有利于获取客户的真实需求信息。为此,本文研究了一种非结构化客户需求信息结构化处理方法。提出了建立客户需求本体以表示客户知识的方法及非结构化客户需求信息(UCN)结构化描述方法;建立了UCN结构化处理的体系结构及相应数学模型,包括句型模式匹配、结构化推理、模糊特征分析及模糊一致性检验等步骤,实现了将UCN转换成结构化的客户需求信息。最后,以普通骑式二轮摩托车的案例阐述了该方法的有效性和实用性。更多还原

【Abstract】 With the increasingly fierce market competition of consumer electromechanical products and the continuous differentiation of market demand, more and more consumer electromechanical products manufacturers start to adopt or have already adopted the model of mass customization to gain market competitiveness by providing consumers with customized and personalized products. The main purpose of this dissertation is to solve the acquiring and processing problem of customer need information for mass customization of consumer electromechanical products.First, considering the heterogeneity of customer skills of expressing needs, the dissertation studies the issue of customer need information acquisition for consumer electromechanical products. For some products (such as computer, motorcycle, automobile, etc.), different customers’(or consumers’) familiarity and understanding of the products varies severely because of their different personal qualifications of professional and social experience, which leads to the heterogeneity of customer skills of expressing needs. Therefore, it necessary to provide differrent customer appropriate interaction under Web environment in need information acquisition according to their skills of expressing needs. In view of this, the thesis studies a need acquisition approach for heterogeneous customers. The interaction process of need acquisition can be customized for each customer by carrying out following six steps: acquisition of customer personal characteristics and need scenarios, feature reasoning analysis, fuzzy weighted calculation, confidence threshold detection, feedback and need expression, and defuzzification. Precise customer needs are outputted at last. Finally, the practicability of the approach is tested and verifed by an application case of scooter.Secondly, the dissertation studies the issue of transformation of semi-structured customer need information (SCN) for consumer electromechanical products. For ordinary consumers (non-professional customers) in need interaction under Web environment, it is difficult for them to express their needs in a structured way through selecting the options of product modules, functional characteristics and technical parameters. Actually, they often hope to express their needs in a semi-structured way through fuzzy semantics to describe the product characteristics and demand characteristics which they concern. In view of this, an expressing and transforming approach for SCN is studied. The mathematical description approach of SCN and the formal description approach of SCN transforming problem are presented, and then the architecture of the SCN transformation is constructed. This architecture is composed by four steps: fuzzy semantic analyzing, intelligent fuzzy reasoning, weighted optimizing, and defuzzification. Precise product functional requirements satisfying SCN are outputted at last. Finally, the practicability of the approach is tested and verifed by an application case of two-wheeled motorcycle.Finally, the dissertation studies the issue of structuralization of unstructured customer needs information (UCN) for consumer electromechanical products. Customers (ordinary consumers) are usually unfamiliar with consumer electromechanical products, and it difficult for them to express their needs through structured and semi-structured way in need interaction under Web environment. Therefore, if we can allow customers to use unstructured natural language way (such as phrases and statements) to describe their needs, will not only increase customer satisfaction of need interaction process, but also facilitate access to the real needs of customers. In view of this, the method for UCN structuralization is studied. Firstly, Customer need ontology (CNO) and the structured description approach of UCN are presented, and the architecture of UCN structuralization is constructed. The process of UCN structuralization is composed by two key stages which are UCN structurizing and fuzzy consistency testing. The former stage includes three steps: sentence pattern matching, structurizing reasoning and fuzzy analyzing. And the latter stage consistes of constraint testing and conflict testing of fuzzy attribute values. Structured customer needs satisfying fuzzy consistency are outputted at last. Finally, the practicability of the approach is tested and verifed by an application case of two-wheeled motorcycle.更多还原