Cooperative co-evolutionary optimization of software project staff assignments and job scheduling

Abstract

This paper presents an approach to Search Based Software Project Management based on Cooperative Co-evolution. Our approach aims to optimize both developers' team staffing and work package scheduling through cooperative co-evolution to achieve early overall completion time. To evaluate our approach, we conducted an empirical study, using data from four real-world software projects. Results indicate that the Co-evolutionary approach significantly outperforms a single population evolutionary algorithm. Cooperative co-evolution has not previously been applied to any problem in Search Based Software Engineering (SBSE), so this paper reports the first application of cooperative co-evolution in the SBSE literature. We believe that co-evolutionary optimization may fit many applications in other SBSE problem domains, since software systems often have complex inter-related subsystems and are typically characterized by problems that need to be co-evolved to improve results. © 2011 Springer-Verlag.