The Evolutionary Component-based Software Reuse Approach

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 (), which is 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.