1 paper · 1 filter
Jean-Philippe Garnier
Current multi-agent AI systems operate with a fixed number of agents whose roles are specified at design time. No formal theory governs when agents should be created, destroyed, or…