Evolving Dyadic Strategies for a Cooperative Physical Task
This work addresses the problem of role mediation in dyadic cooperative tasks for researchers in human-robot interaction or multi-agent systems, but it is incremental as it uses simulations to generate hypotheses without direct human validation.
The paper tackled the problem of understanding how roles are delegated and reassigned in cooperative physical tasks by evolving simulated agents with a genetic algorithm to explore role-switching policies. The result showed that the best performing dyads exhibited high temporal coordination (anti-synchrony), which was correlated with symmetry between agent parameters.
Many cooperative physical tasks require that individuals play specialized roles (e.g., leader-follower). Humans are adept cooperators, negotiating these roles and transitions between roles innately. Yet how roles are delegated and reassigned is not well understood. Using a genetic algorithm, we evolve simulated agents to explore a space of feasible role-switching policies. Applying these switching policies in a cooperative manual task, agents process visual and haptic cues to decide when to switch roles. We then analyze the evolved virtual population for attributes typically associated with cooperation: load sharing and temporal coordination. We find that the best performing dyads exhibit high temporal coordination (anti-synchrony). And in turn, anti-synchrony is correlated to symmetry between the parameters of the cooperative agents. These simulations furnish hypotheses as to how human cooperators might mediate roles in dyadic tasks.