UneVEn: Universal Value Exploration for Multi-Agent Reinforcement Learning release_rhss3n6jrrejbauu5jxyokhaii

by Tarun Gupta, Anuj Mahajan, Bei Peng, Wendelin Böhmer, Shimon Whiteson

Released as a article .

2021  

Abstract

VDN and QMIX are two popular value-based algorithms for cooperative MARL that learn a centralized action value function as a monotonic mixing of per-agent utilities. While this enables easy decentralization of the learned policy, the restricted joint action value function can prevent them from solving tasks that require significant coordination between agents at a given timestep. We show that this problem can be overcome by improving the joint exploration of all agents during training. Specifically, we propose a novel MARL approach called Universal Value Exploration (UneVEn) that learns a set of related tasks simultaneously with a linear decomposition of universal successor features. With the policies of already solved related tasks, the joint exploration process of all agents can be improved to help them achieve better coordination. Empirical results on a set of exploration games, challenging cooperative predator-prey tasks requiring significant coordination among agents, and StarCraft II micromanagement benchmarks show that UneVEn can solve tasks where other state-of-the-art MARL methods fail.
In text/plain format

Archived Files and Locations

application/pdf  3.8 MB
file_2khcppbhnva5nhyr7l2bqmz344
arxiv.org (repository)
web.archive.org (webarchive)
Read Archived PDF
Archived
Type  article
Stage   submitted
Date   2021-02-12
Version   v2
Language   en ?
arXiv  2010.02974v2
Work Entity
access all versions, variants, and formats of this works (eg, pre-prints)
Catalog Record
Revision: f01e4a43-cbf7-4b11-ad6e-b7f0127199c6
API URL: JSON