Heterogeneous Multi-agent Zero-Shot Coordination by Coevolution

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xue, Ke, Wang, Yutong, Guan, Cong, Yuan, Lei, Fu, Haobo, Fu, Qiang, Qian, Chao, Yu, Yang
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909444910088192
author Xue, Ke
Wang, Yutong
Guan, Cong
Yuan, Lei
Fu, Haobo
Fu, Qiang
Qian, Chao
Yu, Yang
author_facet Xue, Ke
Wang, Yutong
Guan, Cong
Yuan, Lei
Fu, Haobo
Fu, Qiang
Qian, Chao
Yu, Yang
contents Generating agents that can achieve zero-shot coordination (ZSC) with unseen partners is a new challenge in cooperative multi-agent reinforcement learning (MARL). Recently, some studies have made progress in ZSC by exposing the agents to diverse partners during the training process. They usually involve self-play when training the partners, implicitly assuming that the tasks are homogeneous. However, many real-world tasks are heterogeneous, and hence previous methods may be inefficient. In this paper, we study the heterogeneous ZSC problem for the first time and propose a general method based on coevolution, which coevolves two populations of agents and partners through three sub-processes: pairing, updating and selection. Experimental results on various heterogeneous tasks highlight the necessity of considering the heterogeneous setting and demonstrate that our proposed method is a promising solution for heterogeneous ZSC tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2208_04957
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Heterogeneous Multi-agent Zero-Shot Coordination by Coevolution
Xue, Ke
Wang, Yutong
Guan, Cong
Yuan, Lei
Fu, Haobo
Fu, Qiang
Qian, Chao
Yu, Yang
Neural and Evolutionary Computing
Artificial Intelligence
Machine Learning
Multiagent Systems
Generating agents that can achieve zero-shot coordination (ZSC) with unseen partners is a new challenge in cooperative multi-agent reinforcement learning (MARL). Recently, some studies have made progress in ZSC by exposing the agents to diverse partners during the training process. They usually involve self-play when training the partners, implicitly assuming that the tasks are homogeneous. However, many real-world tasks are heterogeneous, and hence previous methods may be inefficient. In this paper, we study the heterogeneous ZSC problem for the first time and propose a general method based on coevolution, which coevolves two populations of agents and partners through three sub-processes: pairing, updating and selection. Experimental results on various heterogeneous tasks highlight the necessity of considering the heterogeneous setting and demonstrate that our proposed method is a promising solution for heterogeneous ZSC tasks.
title Heterogeneous Multi-agent Zero-Shot Coordination by Coevolution
topic Neural and Evolutionary Computing
Artificial Intelligence
Machine Learning
Multiagent Systems
url https://arxiv.org/abs/2208.04957