Algorithmically supported team composition for software engineering project courses

Abstract

Composing project teams of students for software engineering courses is a challenging problem for instructors: they need to take into account objectives and constraints such as project motivation, balance of experience in the teams, team size and cultural criteria. In this paper, we present TEASE, a system that algorithmically supports instructors in creating project teams. TEASE proposes possible assignments and enables the instructor to manually adapt them based on their experience. The system also visualizes the effects of these changes on the previously defined objectives and constraints. Our evaluation in a multi-project capstone course shows that TEASE helps instructors create teams with better mean project priority while enabling them to satisfy constraints that were broken during manual assignment. Moreover, TEASE reduced the time needed for team composition by over 60%, making es­pecially the beginning of large project courses more manageable for instructors.