A Search-Based Approach on Metaheuristic Algorithm for Software Modularization to optimize Software Modularity

Abstract

Software refactoring is a process to maintain software quality that will further improvise the software's internal structure without amending its external behaviour. The software refactoring is implemented in the software maintenance phase by modularizing the source code of the original Software clustering is a modularization approach that is used to modularize source code objects with the purpose of boosting the code's reusability and readability. Owing to the clustering issue being NP-hard, evolutionary methods like the genetic algorithm were utilized to address it. There is no search-based technique for modularization that uses a hierarchical method in the structural refactoring literature. It is an unsubstantiated learning technique utilized to cluster software objects (for example files, classes, or modules) with similar structures. The attained clusters can be utilized for a research study, analysis, and comprehending the software objects’ structure and behaviour. While one observation is that executing software module clustering with optimum consequences is thought-provoking. In the research paper, we have utilized the local and global methods wherein a metaheuristic hierarchal search-based clustering approach has been introduced that can help in modularizing the software system. In the algorithm, the outcome is a tree that has nodes including artifacts wherein sub trees collate these artifacts and it is a cluster that is a candidate solution. This algorithm will help in extracting a newer model of the source code which will further help in deciding the correct artifacts which will redirect to obtain documents, packages, and components. The source code can be fetched from GitHub.