Multiagent Cooperation and Competition with Deep Reinforcement Learning
This work provides a tool for studying decentralized learning in complex multiagent systems, but it is incremental as it builds on existing Deep Q-Network methods.
The researchers extended Deep Q-Learning Networks to multiagent environments, specifically in the game Pong, and showed that manipulating reward schemes leads to the emergence of competitive and collaborative behaviors, with agents learning to score efficiently or keep the ball in play longer.
Multiagent systems appear in most social, economical, and political situations. In the present work we extend the Deep Q-Learning Network architecture proposed by Google DeepMind to multiagent environments and investigate how two agents controlled by independent Deep Q-Networks interact in the classic videogame Pong. By manipulating the classical rewarding scheme of Pong we demonstrate how competitive and collaborative behaviors emerge. Competitive agents learn to play and score efficiently. Agents trained under collaborative rewarding schemes find an optimal strategy to keep the ball in the game as long as possible. We also describe the progression from competitive to collaborative behavior. The present work demonstrates that Deep Q-Networks can become a practical tool for studying the decentralized learning of multiagent systems living in highly complex environments.