Preserving Core Components of Object-oriented Packages while Maintaining Structural Quality

Abstract

Different software maintenance activities, carried out from time to time, lead to structural quality degradation. To improve the degraded structural quality of the software system, re-structuring of software entities is desirable and the same can be achieved by using a suitable software clustering technique. Current techniques require too many components (e.g., classes) to be moved between modules (e.g., packages) to achieve high quality software. In such a scenario, the core components of the packages may also move, resulting into loss of identity of the packages. This paper presents a multi-objective evolutionary optimization technique to improve the quality of the existing software while preserving the core components of the package. We evaluate structural quality of six real-world and one random problem instances, using MCA and ECA multi-objective approach.