Robust team formation with role-specific abilities using evolutionary algorithms

Team-formation problems involving multiskilled individuals and heterogeneous roles pose a critical challenge for maintaining robustness under structural and budgetary constraints. This paper addresses robust team optimization with role-dependent abilities. Specifically, we propose a role-specific ability modelling (RSAM) framework to quantify role-dependent abilities, formulate a robust team utility (RTU) objective that balances core role performance and backup coverage, and develop a constraint-aware evolutionary algorithm (CA-EA) for efficient constrained optimization. Experiments on two real world, structurally diverse datasets demonstrate that RTU improves substitute role balance, enhances robustness under core role perturbations and preserves core team performance, while CA-EA efficiently explores constrained solution spaces and outperforms baseline methods. HighlightsA framework transforms individual competencies into quantitative role-specific abilities for optimization.A utility objective captures both core role contributions and nonlinear backup coverage to limit team performance degradation when key roles are unavailable.An evolutionary algorithm directly incorporates role assignments and feasibility requirements into evolutionary operators.Experiments on two real-world datasets demonstrate improved team robustness while maintaining core role strength.