Multi-View Model Refactoring using a Multi-Objective Evolutionary Algorithm

To improve the quality of software systems, one of the widely used techniques is refactoring defined as the process of improving the design of existing system by changing its internal structure without altering the external behavior. The majority of existing refactoring works focus mainly on the source code level. The suggestion of refactorings at the model level is more challenging due to the difficulty to evaluate: a) the impact of the suggested refactorings applied to a diagram on other related diagrams to improve the overall system quality, b) their feasibility, and c) interdiagram consistency. We propose, in this paper, a novel framework that enables software designers to apply refactoring at the model level. To this end, we used a multiobjective evolutionary algorithm to find a trade-off between improving the quality of different diagrams at the same time such as class diagrams and activity diagrams. The proposed multi-objective approach provides a multi-view for software designers to evaluate the impact of suggested refactorings applied to class diagrams on related activity diagrams in order to evaluate the overall quality, and check their feasibility and behavior preservation. The statistical evaluation performed on models extracted from four open source systems confirms the efficiency of our approach.