• Sorted by Date • Classified by Publication Type • Classified by Research Category •
Tom Croonenborghs, Karl Tuyls, Jan Ramon, and Maurice Bruynooghe. Multi-Agent Relational Reinforcement Learning. Explorations in Multi-State Coordination Tasks. In K. Tuyls, P.J. ’t Hoen, K. Verbeeck, and S. Sen, editors, Learning and Adaptation in Multi-Agent Systems, Lecture Notes in Artificial Intelligence, pp. 198–212, Springer Verlag, Berlin, 2006.
In this paper we report on using a relational state spacein multi-agent reinforcement learning. There is growing evidence in theReinforcement Learning research community that a relational representationof the state space has many benefits over a propositional one.Complex tasks as planning or information retrieval on the web can berepresented more naturally in relational form. Yet, this relational structurehas not been exploited for multi-agent reinforcement learning tasksand has only been studied in a single agent context so far. In this paperwe explore the powerful possibilities of using Relational ReinforcementLearning (RRL) in complex multi-agent coordination tasks. More precisely,we consider an abstract multi-state coordination problem, whichcan be considered as a variation and extension of repeated stateless DispersionGames. Our approach shows that RRL allows to represent acomplex state space in a multi-agent environment more compactly andallows for fast convergence of learning agents. Moreover, with this technique,agents are able to make complex interactive models (in the senseof learning from an expert), to predict what other agents will do andgeneralize over this model. This enables to solve complex multi-agentplanning tasks, in which agents need to be adaptive and learn, withmore powerful tools.
@incollection(Croonbook06,
author = "Tom Croonenborghs and Karl Tuyls and Jan Ramon and Maurice Bruynooghe",
title = "Multi-Agent Relational Reinforcement Learning. Explorations in Multi-State
Coordination Tasks",
booktitle = "Learning and Adaptation in Multi-Agent Systems",
editor = "K. Tuyls and P.J. {’t Hoen} and K. Verbeeck and S. Sen",
series = "Lecture Notes in Artificial Intelligence",
volume = "3898",
pages = "198-212",
publisher = "Springer Verlag",
address = "Berlin",
year = "2006",
abstract = {In this paper we report on using a relational state space
in multi-agent reinforcement learning. There is growing evidence in the
Reinforcement Learning research community that a relational representation
of the state space has many benefits over a propositional one.
Complex tasks as planning or information retrieval on the web can be
represented more naturally in relational form. Yet, this relational structure
has not been exploited for multi-agent reinforcement learning tasks
and has only been studied in a single agent context so far. In this paper
we explore the powerful possibilities of using Relational Reinforcement
Learning (RRL) in complex multi-agent coordination tasks. More precisely,
we consider an abstract multi-state coordination problem, which
can be considered as a variation and extension of repeated stateless Dispersion
Games. Our approach shows that RRL allows to represent a
complex state space in a multi-agent environment more compactly and
allows for fast convergence of learning agents. Moreover, with this technique,
agents are able to make complex interactive models (in the sense
of learning from an expert), to predict what other agents will do and
generalize over this model. This enables to solve complex multi-agent
planning tasks, in which agents need to be adaptive and learn, with
more powerful tools.},
owner = "K.Tuyls",
bib2html_pubtype = {Book Chapter},
bib2html_rescat = {Multi-Agent Learning},
timestamp = {2006.11.23},
)
Generated by bib2html.pl (written by Patrick Riley ) on Mon Jan 18, 2010 09:51:52