Multi-agent Collaboration in an Adversarial Turret Reconnaissance Task
INTELLIGENT HUMAN SYSTEMS INTEGRATION 2021(2021)
摘要
Although we have previously observed the emergence of collaboration, or more specifically, coordination between cooperative agents in a Predator-Prey Pursuit task, we have yet to demonstrate this interdependence of agents in a different task domain, where the coordination between cooperative agents is expected to depend on specific environmental parameters that may not support collaboration. In this work, we focus on a multi-agent Turret Reconnaissance task (T-RECON) to explore agent interdependence identified through a state-space perturbation technique. This work aims to 1) identify collaboration, and 2) quantify the magnitude of coordination between cooperating agents.
更多查看译文
关键词
Multi-agent reinforcement learning, Simulation experiments, Adversarial games, Collaboration
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要