| 基于演化构件的软件复用方法 |
| 其他题名 | The Evolutionary Component-Based Software Reuse Approach
|
| 贾育
|
| 专业 | 计算机应用技术
|
| 2002
|
| 学位授予单位 | 中国科学院软件研究所
|
| 学位 | 博士
|
| 学位授予地点 | 中国科学院软件研究所
|
| 关键词 | 基于构件开发
特征建模
描述逻辑
演化计算
软件复用
|
| 摘要 | 该文将演化计算（EvolutionaryComputation,简称EC）的思想和方法应用到基于构件的软件开发之中,提出一种"基于演化构件的软件复用方法"（简称演化构件方法）.在对软件复用深刻认识的基础上,首先,提出一种以语义为中心的形式化构件模型,认为反映构件功能和性能的语义、面向构件联接的语法和负责计算的实现三者之间应该相互分离,其中语义的结构包括领域、定义和语境三个部分,文中用Z语言定义的构件模型是构件演化方法的基础;其次,介绍了特征建模（FeatureModeling）方法,提出用基于知识的特征空间描述构件语义的内容和结构,在深入分析了特征本质的基础上,完善了特征空间的可视化描述工具和特征建模过程,使得演化构件方法在软件工程上具备一定的可操作性;再次,借助于知识描述语言-描述逻辑,设计了一种适合表达构件语义的形式描述语言-特征描述逻辑（FeatureDescriptionLogics,简称FDL）,用FDL表达语义内涵及实现自动推理,为构件的演化计算方法的实现提供了可能.最后,提出构件演化计算方法,它是前三项研究的目的所在,即从自然演化的角度认识构件和CBD方法及过程,设计一组基于演化计算原理的构件运算算法,充分发挥演化计算方法的优势,在一定程度内（半）自动地建立最优或接近最优的基于构件软件系统. |
| 其他摘要 | Nowadays the Component-Based Development (CBD) is a hot topic in field of Software Reuse. It is regarded as a practical approach to improve software developing production and software quality, as well as the inevitable passing road for software industrializing. However, to date CBD is not comprehensively adopted in software engineering practice. One of the crucial reasons is that it fails to effectively solve the problems of representing and processing the component semantics, as well as the shortage of analysis tools to support the its methods and life cycle. In this paper the rationale and methods of Evolutionary Computing (EC) are introduced into CBD what is called the Evolutionary Component-Based Software Reuse Approach (Abbr. CBD/EC). Based on the profound understanding of the Software Reuse, firstly, a semantics-intensive, formal component model is proposed, which separates the component into three parts: the semantics regarding the function and extra-function of component, the syntax oriented to connection among components and the component implementation. Furthermore, the semantics is composed of domain, definition and context. This paper defines the component model in language Z which is the foundation of CBD/EC. Secondly, the Feature Modeling method is introduced, suggesting that the knowledge-based Feature Space can be used to describe the content and structure of component semantics. After profoundly analyzing the essence of the feature concept, the visual describing tool of Feature Space and feature modeling route are greatly extended. In the end the CBD/EC turns out to be a practical method in the perspective of Software Engineering. Thirdly, originally coming from the knowledge representation language - Description Logics (DLs), a formal describing language is designed to represent the component semantics called Feature Description Logics (FDL), which s used to express the intension of semantics and implement the automatic reasoning. Thus there exists the possibility for CBD/EC to turn into reality. Lastly, the component evolutionary computing method is provided as the general aim of all mentioned before. That is, recognizing the component and CBD process from the perspective of natural evolutionary, a set of EC theory based component operational algorithms is created. Taking advantage of the power of EC methodology, the component-based systems in a way are (semi-) automatically built to be the best or nearly best ones. To sum up, the Evolutionary Component-Based Software Reuse Approach provides a possible way for the automation and engineering of Software Reuse. |
| 页数 | 120
|
| 语种 | 中文
|
| 内容类型 | 学位论文
|
| URI标识 | http://ir.iscas.ac.cn/handle/311060/5722
|
| 专题 | 中科院软件所_中科院软件所
|
推荐引用方式
GB/T 7714 |
贾育. 基于演化构件的软件复用方法[D]. 中国科学院软件研究所. 中国科学院软件研究所,2002.
|
修改评论