Abstract In real world, the software systems often need to be continuously modified to satisfy the ever changing requirements and environment. Mostly, it is carried out without following the original design principles of the system. Over a period of time, such a continuous modification deteriorates the structural quality, hence increases the system complexity. To improve the structural quality of whole system, the software clustering seems more feasible technique. Recently, the search – based approach gain more attention to solve the software clustering problem. In this paper, we propose a search – based multi – objective optimization to re-structure the object – oriented software system using different coupling strength scheme such as binary coupling, absolute coupling and relative coupling scheme. The approach is evaluated over four real – world and three random software applications. The experimentation results show that how the use of absolute and relative coupling strength scheme leads to generate more effective solutions compared binary coupling strength.
Read full abstract