【作者】 吴磊；

【导师】 刘晓平；

【作者基本信息】 合肥工业大学 ， 计算机应用技术， 2005， 硕士

【摘要】 多态的概念已被广泛地应用于工程、可靠性分析、面向对象程序设计、神经网络等多个领域中。本文在科学计算辅助建模领域引入多态的概念,用来表述复杂模型由于模型精度、计算时间、计算精度的变化所产生的多种状态;基于多态模型研究模型复杂度与科学计算精度之间的关系,通过归纳、总结、表述这种关系,为科学计算的辅助建模工作提供依据和原则。本文主要的研究内容包括: 1) 在分析多态概念及其应用的基础上,将其引入科学计算辅助建模领域,提出多态模型的概念,基于多态模型研究复杂模型精度、计算时间、计算精度之间的关联。 2) 以几何模型为研究对象,基于三维特征研究多态模型的表示和变迁机理。通过分析比较三维特征识别技术,并结合多态模型的自身特点,采用基于图的特征识别技术处理模型中的特征。 3) 针对多态模型所涉及的特征,建立实验模型,分析科学计算结果,总结模型精度、计算时间、计算精度之间的关系,并通过函数的方式进行表达,为多态模型工作提供依据和原则。 4) 建立多态模型研究工作的实验平台,通过特征树存储和表达模型中所包含的特征,实现多态模型的表示和变迁。更多还原

【Abstract】 The concept of multi-state has been widely applied in many fields, such as engineering analysis, reliability analysis of system, object-oriented programming and neural network. In the thesis the concept is introduced into the field of computer-aided modeling of science computing, and the concept is used to represent different states of a complex model. A complex model’s states are defined according to model precision, computing time and computing error. The thesis studies the relation between model precision and computing error based on multi-state model. Some guidance rules are given by concluding and expressing these relations, and the rules are helpful to computer-aided modeling. The main content of the thesis includes:1) Based on the analysis of the applications of multi-state concept in different fields, the concept is imported into the field of computer-aided modeling of science computing to define the multi-state model. So the relation between model precision, computing time and computing error is studied based on the multi-state model.2) Geometry model is regarded as the study object. Representation and variance of multi-state model is studied based on 3D geometry feature. By comparing and analyzing of different 3D feature recognition methods, the method of 3D feature recognition based on graph is selected to deal with geometry features of multi-state model.3) To all 3D geometry features involved in the multi-state model, many experimental models are constructed for experimental computing. The relation between model precision, computing time and computing error is concluded by analyzing the computing result. The relation is expressed in form of mathematic function to guide study on multi-state model.4) A system of multi-state model is designed and realized, which is the experimental platform for the coming research. Feature tree is used to express and store features of a model. Representation and variance of multi-state model is realized in this system.更多还原