Semantic Analysis For CASE Design Documation And Adaptive Evaluation For Components

With the rapid development of software industry, high quality software system should be developed to meet the customer’s requirement, which will appy to the actual backgroud and business. Component-based Software Development (CBSD) method supports software reuse and speeds up development, which is an important step in program design automation. There are four main steps in software reuse process:system analysis, component description, component query and compoment assembly. However, the following problems exist in the CBSD process by far:(1) What components does the system require;(2) What’s the exsiting form of components (e.g. what kind of functions and interfaces do they provide);(3) How to estabilsh the mapping relations to query and match components. Because of the existence of these problems, it’s not smooth in software reuse. Current technology focuses on the research of component retrieval and assembly usually, which is the problem (3), however, for the question (1)(2) which helps make clear the query target of the components by extractting comcept components in the target system (the system which will be developed), we lack unified and normative mehod of extractting components, but the question (1)(2) are the premise to sovle the problem (3). Besides, adaptability level of components for system also affects the assembly quality directly.To solve the above problems, this paper focused on system analysis and evaluation in component query.In system analysis, this paper used CASE(Computer Aided Software Engineering) tools to design the target system, forming a detailed documation such as use cases, classes, components and their assembly deployment, so the concept components and their calling relations can be extracted through the semantic understanding of the documentations.In component evaluation, firstly, in order to facilitate the querying and matching, ontology description was given for component entities in the current component library, in this way, it built a component ontology library. Secondly, with the help of a domain dictionary, to some extent, it gived mapping relations between concept components and component ontologies to complete the component query. At the same time, it had established an intelligent Agent system, in this system candidate components were given adaptability evaluation, which established single component evaluation model and multi-components evaluation model respectively. The single component evaluation model used multiple evaluation degree.Combining with weight matrix, the multi-components evaluation model used dynamic propraming method to get appropriate component combination for assembly.At last, this paper had implemented and tested the model, the result showed that the model was consistent with the expected outcomes, which not only selected the components that meet the assembly requirement successfully, but also avoided confilict in assembly process, in this way, it supported software reuse better.