【作者】 邓武；

【导师】 杨鑫华；

【作者基本信息】 大连交通大学 ， 计算机应用技术， 2006， 硕士

【摘要】 冲模设计是一个复杂的过程，严重依赖于设计者的经验，如何提高冲模设计智能化程度一直是冲模CAD领域的研究重点。本文将基于实例推理技术和粗糙集理论应用到冲模CAD领域，以期充分利用已有设计方案，实现冲模设计的智能推理，提高设计质量和效率，缩短开发周期，降低冲模生产成本。 本文工作主要包括以下几个方面： 第一、对基于实例推理技术和粗糙集理论进行了较为详尽的理论阐述。对当前应用于实例检索的主要方法进行了讨论，总结了它们的适用领域和优缺点。 第二、建立了一个支持实例推理的冲模产品逻辑结构。在此基础上，讨论了实例描述、表达策略和组织方法，采用基于框架面向对象的实例描述方法对冲模实例进行了描述，并利用关系数据库构建了冲模实例库。 第三、在分析冲模的实际设计过程和智能CAD系统基础上，建立了一个基于IDEFO的冲模CAD系统功能模型，设计实现了一个基于实例推理和粗糙集理论的智能冲模CAD原型系统。 第四、利用粗糙集理论对冲模实例库进行分析和处理，给出了一种基于分级聚类法和决策属性支持度的离散化方法处理属性表中的定量问题，确定各特征属性的重要程度，并以此为依据，抽取实例的关键特征属性来建立检索，用最近邻策略，实现了目标实例和源实例的相似匹配。更多还原

【Abstract】 Die design is a complex processing, and it seriously depends on the experience of the designer. How to improve the extent of the Die design intelligence that is always the keystone to research in the field of the Die CAD. This paper will apply Case-based Reasoning (CBR) technique and Rough Sets Theory (RST) in the field of the Die design, so it can implement intelligentized reasoning of Die design using existed project.It should improve the design quality and efficiency, shorten the period of development and reduce the cost of Die.The following points are finished in the dissertation:Firstly, this paper expatiates CBR technique and RST in details. Then it discusses the mainly technique of case retrieval in current, and summarizes their advantages and disadvantages in applied areas.Secondly, the paper builds a Die-product logical framework which supports CBR, based on which it discusses case description, expression strategy and organization method. It adopted case description method based on object oriented framework to describe the die cases, and realized using Related-database System to design Die cases database.Thirdly, in the foundation of the analysis to the practical die design processing and the intelligent CAD system, it builds a functional model based on IDEFO’s Die CAD system. Therefore, this paper designs and realizes a intelligentized Die CAD that based on CBR and RST.Fourthly, the thesis analyzes die case database using RST, It puts forward a method using grade classification and decision attributes support degree to deal with the quantitative features. And it confirms the important degree of all types of characteristic attributes. It aims to build up a retrieval based on case’s key attributes. Then it makes use of the nearest neighbor strategy to implement the similar matching between the target case and source case.更多还原