Group Role Assignment with Bound Agents

Group Role Assignment (GRA) aims to optimize group performance by assigning appropriate roles to agents. Existing GRA research has explored relationships among agents and roles including conflict, cooperation, and role coupling, which help address challenges in complex environments. Binding relationships among agents are common in real-world collaboration and typically arise from factors such as information exchange and mutual dependency, exemplified by mentor-mentee relationships. However, existing GRA research overlooks binding relationships, potentially result in the inability to complete collaborative tasks. Therefore, this paper models the binding relationships among agents by defining a binding matrix and introducing a constraint in GRA. Then we formalize the GRA problem with binding relationships among agents. To address this, we propose a novel GRA collaborative model, called GRA with bound agents (GRABA). Specifically, we introduce a constraint in GRA to ensure that bound agents are assigned to the same role, and solve it using a GRA-based method. Experimental results validate the model’s effectiveness in optimizing role assignment and its robustness to parameter variations.