Research On Algorithms For Component Extraction Using Existing Codes

CBSD (Component-Based Software Development) is a software development method, with software architecture as buleprint to assemble and software component as module to install. It improves productive efficiency and quality of software greatly. With the development of component techniques, extracting component from existing codes becomes a high efficiency, practicable, lower cost method for component production and it becomes the important research domain of software reusability and program comprehension.The component extraction technique based can satisfy the various requirement of client through adjusting the function parameter of extraction. This process is very complex one including solving NP hard problems during structure decomposition, determining and designing the parameter of metric model, comprehending and using the client’s requirement in the process of extraction, and so on. To solve these problems effectively, we need analyze the object that this paper is focused on, build the corresponding mathematics model and use relevant algorithm to make a solution.This paper brings forward research scheme for extracting component existing codes based, with the basis of system comprehension and software metric, and provides a more efficient method to extract component from existing system. This method includes:adopting fuzzy cluster analyze to decompose structure and extract component innovatively. The code files is viewed as cluster units and the client’s requirement as cluster factor to distinguish the existing files. After that, to solve the file’s properties distinction, this paper designs the URQCM (Usability, Reusability and Quality of Component Metrics) model to measure the usability, reusability and quality of files under study to mark and validate the candidate components. Besides these, to implement extraction from large-scale system, this paper designs and develops the RBEC (Request Based Extracting Component) genetic algorithm according with the strategy of genetic algorithm and URQCM model. Especially, we design coupling metric to measure the cohesion and coupling among the files in the extracting solution. At last, this paper uses the proposed algorithms to solve the case in fact and get satisfactory result. Through the analyze to the algorithm and experimental result, the paper points out the future research direction...