Saved in:
Bibliographic Details
Main Authors: Bolliger, Diego, Zauter, Lorenz, Ziegler, Robert
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.06747
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916647946682368
author Bolliger, Diego
Zauter, Lorenz
Ziegler, Robert
author_facet Bolliger, Diego
Zauter, Lorenz
Ziegler, Robert
contents In this paper, we devise three actor-critic algorithms with decentralized training for multi-agent reinforcement learning in cooperative, adversarial, and mixed settings with continuous action spaces. To this goal, we adapt the MADDPG algorithm by applying a networked communication approach between agents. We introduce surrogate policies in order to decentralize the training while allowing for local communication during training. The decentralized algorithms achieve comparable results to the original MADDPG in empirical tests, while reducing computational cost. This is more pronounced with larger numbers of agents.
format Preprint
id arxiv_https___arxiv_org_abs_2503_06747
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Fully-Decentralized MADDPG with Networked Agents
Bolliger, Diego
Zauter, Lorenz
Ziegler, Robert
Machine Learning
Artificial Intelligence
Multiagent Systems
In this paper, we devise three actor-critic algorithms with decentralized training for multi-agent reinforcement learning in cooperative, adversarial, and mixed settings with continuous action spaces. To this goal, we adapt the MADDPG algorithm by applying a networked communication approach between agents. We introduce surrogate policies in order to decentralize the training while allowing for local communication during training. The decentralized algorithms achieve comparable results to the original MADDPG in empirical tests, while reducing computational cost. This is more pronounced with larger numbers of agents.
title Fully-Decentralized MADDPG with Networked Agents
topic Machine Learning
Artificial Intelligence
Multiagent Systems
url https://arxiv.org/abs/2503.06747