Adjusting Organization to Improve Coordination for Large Heterogeneous Multi-agent Teams

Large scale heterogeneous multi-agent and multirobot teams have been popular in many domains. Similar to human society infrastructure, agents only closely coordinate with some of the others. Their organization has been proven as a key factor to influence the team performance. In this paper, we made an initial effort on finding how to adjust its organization to expedite team performance. According to the intrinsic nature of the team if we can observe the interactions between agents in the team, we are able to setup an adaptive algorithm to adjust the network. In this paper, based on our coordination simulation, we have proposed a novel organization adaptive approach to improve the coordination efficiency. Moreover, we extend this idea and design an integrated algorithm by combining social effects. Our experiment results support our ideas, and we also show that the proposed coordination method is more efficient to coordinate heterogeneous agents.