【作者】 苏楠；

【导师】 赵燕伟；

【作者基本信息】 浙江工业大学 ， 机械电子工程， 2009， 博士

【摘要】 随着全球化市场形成和技术变革加速,相对稳定的市场环境逐步向动态多变的方向发展,企业面临的巨大挑战是如何解决不断增长的客户个性化需求和低成本生产、高效率供货之间的矛盾问题。可拓配置设计方法为有效解决“大规模”与“定制”之间的矛盾问题、顾客需求的“个性化设计”与企业效益的“共同性设计”之间的矛盾问题提供了一种全新思路。本文在国家自然科学基金项目(编号:50575207)和国家863高技术研究发展计划(编号:2002AA411110)的直接资助下,从理论到应用对产品族可拓配置设计的关键问题进行了深入研究。在全面综述国内外现有方法的基础上,综合运用可拓逻辑理论,研究产品族配置设计过程中的个性化需求与设计效率之间的矛盾问题,探索一种创新与辨证思维形式化、模型化、智能化设计方法,对于解决产品族配置设计理论研究与工程实现中的瓶颈问题,具有重要的理论意义和应用价值。一、针对产品族配置设计不同阶段的矛盾问题,建立定性定量相结合的产品族可拓配置设计模型,深入分析了关系物元配置模型的拓展性,从模型组合的角度,提出用于产品族结构模块划分阶段的网状模型、产品族实例推理阶段的树状模型和配置冲突消解过程的环状模型。二、为有效确定模块划分过程中复杂联接关系,提出基于可拓聚类的配置结构划分方法,首先建立不同设计约束下的关联距离函数,采用结构关联性因子表示结构间定性和定量的联系,其次对公共结构网状模型进行检索和变换,从而满足了模块划分中结构独立性和联接关系紧密性的要求,为产品族核心模块确定和配置需求结构化提供支持。三、针对同一设计区间内相似零件难以有效检索的问题,提出基于可拓实例推理的方案生成方法。从顾客定制需求物元出发,建立需求结构化树状模型,利用相关度得到面向顾客不同定性需求的局部相似结构;提出改进距相似度全局匹配算法,该算法能够从不同实例中得到最优结构匹配,从而提供满足顾客要求的一类个性化产品的初始设计方案。四、为提高实例修改过程所获得方案知识的有效性,提出基于可拓挖掘的配置方案再生成方法。首先利用所提出的关系物元环状模型,将一般性的配置规则转化为可拓规则;其次在方案修改过程中利用可拓规则间的转化,建立变换方式与传导效应之间的关系,挖掘出配置要素之间的相互影响,从而对配置过程中的隐含知识加以分析、提取和利用,生成能够满足设计要求的可变方案。五、开发基于可拓逻辑的产品族配置设计原型系统,该系统以某油锯产品族和某分拣机系列产品为设计对象,针对产品族配置过程中需求结构间的矛盾问题以及顾客的目标要求,有效集成所提出的可拓聚类、可拓实例推理和可拓挖掘方法,建立了可拓模块划分、可拓配置推理和可拓挖掘变换等功能模块,该原型系统验证本文所提方法的可行性,初步实现了产品族配置设计的矛盾问题智能化处理。更多还原

【Abstract】 Along with the globalization market formation and the technical changing acceleration, the traditional relatively stable market environment develops gradually to the dynamic changeable direction. The huge challenge is how to satisfy the costomizated demand at the low cost, high grade and the quick product date of delivery. This dissertation is under the National Science Foundation (50575207) and the National 863 High Technology Program Foundation (2002AA411110).Design for the product family is the inevitable result because of the market perturbation, the customer demand and the diversification, diverse, the personalized product, which share the small enterprises to provide more opportunities, but simultaneously also bring the enterprise and the product development more challenges. So how to deal with the contradiction problems by the artifical intelligence has become very important.This dissertation took the extension logic method into the product configuration design process. The main contributions of this dissertation were as following.After the existing research and the dilemmas in the product configuration design were summarized. Then six main methods design for framework, design for model, design for reasoning, design for common module, design for varieties, design for assembly and design for mass customization. Then the extension design flowchart was introduced. Finally present the main content of this paper.Firstly, to build the qualitative and quantitative extension model, the conflictions in the different steps for configuration design were analyzed. A net or tree model were built to stand for the configuration process at the support of relationship matter element extension. Another confliction resolution cycle model was presented to realize the case modifying. As a result, the basic framework to deal with the extension knowledge was advanced.Secondly, to solve the complex relationships between different parts, the division in the configuration database on extension cluster was realized. Its focus was to make sure the relationship in the different parts. Through the relationship standard, a new product structure division method which satisfies the individual and correlative need was presented. After the transforming on the common model, the database is more suitable for retrieval.Thirdly, at the purpose of the similar parts retrieval, a configuration method on extension relationship-matter reasoning was presented. Its focus was the similar structures in the same design space which were hard to retrieval. After the demand and structure mapping, the similar structures for configuration are generated. Then the general retrieval method was employed to finish the configuration process to satisfy the different customers’ demands.Forthly, one method to deal with the case transforming on Extension mining was advanced. Its aim was to realize the transforming process’s auto-generation. According relationship-matter model, the general configuration rule could be changed into an extension rule. On the confliction transforming, the conflicts between different product structures could be mined. Then the relationships between transforming effect and the configuration ability could be built. Finally a transforming case could be generated.Finally, the prototype system for extension product family configuration, to the various design spaces was realized, global conflicts and changing demands, using the extension cluster method, demand and structures mapping, extension reasoning method and extension transforming method, the system solve the conflicts in the configuration process.更多还原