Multi-objective Ranking based Non-Dominant Module Clustering

Although there has been a demarcation between development and evolution (maintenance) of software, this is increasingly irrelevant as fewer and very fewer systems are completely new. Additionally, once a software engineer understands a system's structure, it is difficult to preserve this understanding, because the structure tends to change during maintenance. So Software engineers greatly emphasize on good modular structures as well modularized software is easier to develop and maintain. In this paper, I propose two algorithms, one as a search optimization technique and another as a multiobjective fitness evaluation function of the search technique. The former one, I have labeled as Modified Pareto Optimal Genetic Algorithm (Modified par-op GA) and the later one as Non-Dominance Module Ranking Algorithm (Non-DMR). As the Fitness function is a component of GA, I have embedded the Non-DMR within the Modified par-Op GA. Keywords—Non-hierarchical clustering optimization, cohesion, coupling, reverse engineering, Agglomerative hierarchical clustering, Pareto optimality.