Indian Journal of Science and Technology

Abstract

Background/Objectives: Module based software development provides one of the best ways for the development of big projects. Project can easily be divided into different modules so that the development process is faster. In our approach we are determining the reusable components of a software system and enhancing the accuracy of the methods for determining them. Method/Statistical Analysis: We are using genetic algorithm and fuzzy c mean algorithm to find the cohesion and coupling between the components of a software system on basis of which reusability of the components is determined. Already existing software projects are collected and technique was applied on them to determine the reusable components of those projects and value of f-measure is calculated as a factor of comparison with previous techniques. Finding: In our approach we have found the reusable components of a software system by not only finding the dependencies between the different elements of a single package but also the finding dependencies between the elements of different packages of an already existing project that makes our approach different from existing study. The combination of two algorithms is used for identification and classification of the functions, classes, packages and sub packages in order to find out coupling and cohesion between these elements of a software system and and yields higher accuracy as compared to existing methods
Keywords: Cohesion, Coupling, Clustering, Interlobular Density, Packages, Software Reusability