Hierarchical Multi-Agent Reinforcement Learning
Full Text: agents01.pdfHierarchical reinforcement learning methods have previously been shown to speed up learning primarily in single-agent domains. In this paper we explore the use of this spatio-temporal abstraction mechanism to speed up a complex multi-agent reinforcement learning task. The focus here is on domains requiring multiple agents to co-operatively learn tasks which require coordination among the agents. We adopt the MAXQ framework of hierarchical learning as it naturally extends to the multi agent.
Citation
M. Ghavamzadeh, S. Mahadevan. "Hierarchical Multi-Agent Reinforcement Learning". pp 246-253, June 2001.Keywords: | |
Category: | In Conference |
BibTeX
@incollection{Ghavamzadeh+Mahadevan:01, author = {Mohammad Ghavamzadeh and Sridhar Mahadevan}, title = {Hierarchical Multi-Agent Reinforcement Learning}, Pages = {246-253}, year = 2001, }Last Updated: June 11, 2007
Submitted by Staurt H. Johnson