Decentralized multi-agent reinforcement learning algorithm using a cluster-synchronized laser network

TL;DR

This paper introduces a photonic-based decentralized multi-agent reinforcement learning algorithm leveraging laser network synchronization to solve the competitive multi-armed bandit problem efficiently.

Contribution

It presents a novel approach using chaotic laser oscillations and cluster synchronization for decentralized decision-making in MARL without explicit information sharing.

Findings

Chaotic laser oscillations enable effective exploration and exploitation.

Cluster synchronization facilitates cooperative decision-making.

Decentralized control is achieved through physical laser network dynamics.

Abstract

Multi-agent reinforcement learning (MARL) studies crucial principles that are applicable to a variety of fields, including wireless networking and autonomous driving. We propose a photonic-based decision-making algorithm to address one of the most fundamental problems in MARL, called the competitive multi-armed bandit (CMAB) problem. Our numerical simulations demonstrate that chaotic oscillations and cluster synchronization of optically coupled lasers, along with our proposed decentralized coupling adjustment, efficiently balance exploration and exploitation while facilitating cooperative decision-making without explicitly sharing information among agents. Our study demonstrates how decentralized reinforcement learning can be achieved by exploiting complex physical processes controlled by simple algorithms.