论文标题:面向科学试验的工程数据仓库模型研究
Research on Model of Engineering Data Warehouse for Scientific Experimentation
论文作者 王传栋
论文导师 黄志球,论文学位 硕士,论文专业 计算机应用技术
论文单位 南京航空航天大学,点击次数 889,论文页数 83页File Size1056k
2005-01-01论文网 http://www.lw23.com/lunwen_43652/ 工程数据仓库;信息供应链;元数据;元模型;互操作;联机分析处理
engineering data warehouse; information supply chain; meta data; metamodel; interoperable; on-line analytical processing
在面向科学试验的工程领域中,异构分布环境中的工程数据流动形成一个信息供应链(ISC),有效集成 ISC 上的工程数据成为数据处理领域的一个研究热点。元数据集成是确保 ISC 上工程数据有效流动的先决条件,由于工程数据具有组织分散、模式复杂多变、缺乏统一的结构描述等特点,使得 ISC 上的元数据集成呈现更为复杂的一面。本文以面向科学试验的工程领域和数据仓库技术为研究背景,对模型驱动的元数据集成进行了研究。 首先,本文分析了工程数据的基本特点,总结了工程应用的概念本体,提出了一个全面的元数据管理策略和一套完整的基于模型的元数据集成方案,定义了一个基于本体语义共识的四层元建模框架,用于工程数据仓库(EDW)的领域建模。其次,给出了基于模型驱动的元数据集成体系结构,参照 OMG 组织定义的公共仓库元模型 CWM,定义了一种共享的、与平台无关的元模型,作为描述元数据的形式化语言,使得不同问题域的模型以互操作的方式表现工程数据仓库,并针对仓库的 OLAP 问题领域,详细设计了 OLAP 元建模所涉及的领域元模型。最后,分析了工程数据仓库 OLAP 场景所独有的领域特征,从仓库终端用户的角度出发,详细探讨了使用 OLAP 元模型和映射元模型进行逻辑建模和物理部署建模的实现过程。
It is a hotspot to study the integration of engineering data in Information Supply Chain (ISC), which is an engineering data flow in distributed and heterogeneous engineering applications of Scientific Experimentation. Meta data integration is a key prerequisite to ensure the effectiveness of engineering data flow in ISC and it appears much complicated due to the features of engineering data, such as separate organization, complex and variable schemas, lack of uniform structure, etc. Based on the development of engineering applications of Scientific Experimentation and Data Warehouse technology, this thesis primarily focuses on the model-driven method about meta data integration. This thesis firstly analyzes the characteristics of engineering data, and summarizes the concept ontology of engineering applications. Then a comprehensive strategy of meta data management and a complete model-based meta data integration solution are proposed, and a four-layer meta modeling framework based on the semantics agreement of each ontology concept which used to model Engineering Data Warehouse (EDW) is defined. Secondly, a model-driven architecture of meta data integration is introduced. A shared and platform-independent metamodel is defined according to Common Warehouse Metamodel (CWM) which is adopted standard by the Object Management Group (OMG). As a formal language capable of representing meta data, the metamodel enables different domain models to represent EDW in an interoperable manner. And then a domain-specific metamodel is designed aims at On-Line Analytical Processing (OLAP) in EDW. Finally, the specific features of OLAP scenario related with EDW are analyzed. The implementation of logical modeling and physical deployment modeling with OLAP metamodel and Mapping metamodel is discussed in detail from the perspective of an end user of the system.
|
