Group-Agent Reinforcement Learning with Heterogeneous Agents

Group-agent reinforcement learning (GARL) is a newly arising learning scenario, where multiple reinforcement learning agents study together in a group, sharing knowledge in an asynchronous fashion. The goal is to improve the learning performance of each individual agent. Under a more general heterogeneous setting where different agents learn using different algorithms, we advance GARL by designing novel and effective group-learning mechanisms. They guide the agents on whether and how to learn from action choices from the others, and allow the agents to adopt available policy and value function models sent by another agent if they perform better. We have conducted extensive experiments on a total of 43 different Atari 2600 games to demonstrate the superior performance of the proposed method. After the group learning, among the 129 agents examined, 96% are able to achieve a learning speed-up, and 72% are able to learn over 100 times faster. Also, around 41% of those agents have achieved a higher accumulated reward score by learning in less than 5% of the time steps required by a single agent when learning on its own.

Group-Agent Reinforcement Learning with A Distributed Deep Solution

It can largely benefit the reinforcement learning process of each agent if multiple agents perform their separate reinforcement learning tasks cooperatively. These tasks can be not exactly the same but still benefit from the communication behaviour between agents due to task similarities. In fact, this learning scenario is not well understood yet and not well formulated. As the first effort, we provide a detailed discussion of this scenario, and propose group-agent reinforcement learning as a formulation of the reinforcement learning problem under this scenario and a third type of reinforcement learning problem with respect to single-agent and multi-agent reinforcement learning. We propose that it can be solved with the help of modern deep reinforcement learning techniques and provide a distributed deep reinforcement learning algorithm called DDA3C (Decentralised Distributed Asynchronous Advantage Actor-Critic) that is the first framework designed for group-agent reinforcement learning. We show through experiments in the CartPole-v0 game environment that DDA3C achieved desirable performance with very stable training and has good scalability.