Coordinating Policies Among Multiple Agents via an Intelligent Communication Channel

TL;DR

This paper introduces a novel multi-agent communication framework where agents interact via an intelligent facilitator that interprets signals to enhance collective performance without central control, tested across various environments.

Contribution

It proposes a new communication architecture with an intelligent facilitator that improves cooperation in MARL while preventing centralized control.

Findings

Outperforms existing baselines in cooperative MARL environments

Facilitator effectively interprets signals to enhance collective decision-making

Agents are incentivized to reduce reliance on the facilitator's messages

Abstract

In Multi-Agent Reinforcement Learning (MARL), specialized channels are often introduced that allow agents to communicate directly with one another. In this paper, we propose an alternative approach whereby agents communicate through an intelligent facilitator that learns to sift through and interpret signals provided by all agents to improve the agents' collective performance. To ensure that this facilitator does not become a centralized controller, agents are incentivized to reduce their dependence on the messages it conveys, and the messages can only influence the selection of a policy from a fixed set, not instantaneous actions given the policy. We demonstrate the strength of this architecture over existing baselines on several cooperative MARL environments.