Networked Agents in the Dark: Team Value Learning under Partial Observability

TL;DR

This paper introduces DNA-MARL, a novel cooperative multi-agent reinforcement learning approach enabling agents with partial observability to learn shared objectives through local communication, expanding applicability in privacy-sensitive real-world domains.

Contribution

The paper presents DNA-MARL, a distributed MARL method that uses consensus and gradient descent for networked agents under partial observability, addressing communication constraints.

Findings

DNA-MARL outperforms previous methods in benchmark scenarios.

The approach effectively handles privacy and message delivery issues.

It broadens the scope of networked agent applications.

Abstract

We propose a novel cooperative multi-agent reinforcement learning (MARL) approach for networked agents. In contrast to previous methods that rely on complete state information or joint observations, our agents must learn how to reach shared objectives under partial observability. During training, they collect individual rewards and approximate a team value function through local communication, resulting in cooperative behavior. To describe our problem, we introduce the networked dynamic partially observable Markov game framework, where agents communicate over a switching topology communication network. Our distributed method, DNA-MARL, uses a consensus mechanism for local communication and gradient descent for local computation. DNA-MARL increases the range of the possible applications of networked agents, being well-suited for real world domains that impose privacy and where the messages may not reach their recipients. We evaluate DNA-MARL across benchmark MARL scenarios. Our results highlight the superior performance of DNA-MARL over previous methods.