Group-Agent Reinforcement Learning with Heterogeneous Agents

TL;DR

This paper introduces a novel heterogeneous group-agent reinforcement learning framework that enables agents with different algorithms to share knowledge and improve learning efficiency, demonstrated through extensive Atari game experiments.

Contribution

It proposes new group-learning mechanisms for heterogeneous agents, allowing them to adopt better policies and accelerate learning in diverse environments.

Findings

96% of agents achieved learning speed-up

72% learned over 100 times faster

41% achieved higher rewards in less than 5% of the time

Abstract

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.