Stage-oe-small.jpg

Inproceedings3047: Unterschied zwischen den Versionen

Aus Aifbportal
Wechseln zu:Navigation, Suche
(Die Seite wurde neu angelegt: „{{Publikation Erster Autor |ErsterAutorNachname=Lode |ErsterAutorVorname=Clemens }} {{Publikation Author |Rank=2 |Author=Urban Richter }} {{Publikation Author |Ra…“)
 
 
Zeile 16: Zeile 16:
 
|Year=2010
 
|Year=2010
 
|Month=Juli
 
|Month=Juli
|Booktitle=Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2010)
+
|Booktitle=Proceedings of the 12th Annual Conference on Genetic and Evolutionary Computation (GECCO 2010)
|Publisher=ACM SIGEVO
+
|Pages=1015-1022
 +
|Publisher=ACM
 +
|Address=New York, NY, USA
 +
|Editor=Martin Pelikan, Jürgen Branke
 
}}
 
}}
 
{{Publikation Details
 
{{Publikation Details
Zeile 23: Zeile 26:
  
 
In this paper, LCSs are investigated in an instance of the generic homogeneous and non-communicating predator/prey scenario. A group of predators collaboratively observe a (randomly) moving prey as long as possible, where each predator is equipped with a single, independent XCS. Results show that improvements in learning are achieved by cleverly adapting a multi-step approach to the characteristics of the investigated scenario. Firstly, the environmental reward function is expanded to include sensory information. Secondly, the learners are equipped with a memory to store and analyze the history of local actions and given payoffs.
 
In this paper, LCSs are investigated in an instance of the generic homogeneous and non-communicating predator/prey scenario. A group of predators collaboratively observe a (randomly) moving prey as long as possible, where each predator is equipped with a single, independent XCS. Results show that improvements in learning are achieved by cleverly adapting a multi-step approach to the characteristics of the investigated scenario. Firstly, the environmental reward function is expanded to include sensory information. Secondly, the learners are equipped with a memory to store and analyze the history of local actions and given payoffs.
 +
|ISBN=978-1-4503-0072-8
 +
|DOI Name=10.1145/1830483.1830669
 
|Projekt=OCCS, OCCS (Phase III)
 
|Projekt=OCCS, OCCS (Phase III)
 
|Forschungsgruppe=Effiziente Algorithmen
 
|Forschungsgruppe=Effiziente Algorithmen

Aktuelle Version vom 13. Juli 2010, 21:21 Uhr


Adaption of XCS to Multi-Learner Predator/Prey Scenarios


Adaption of XCS to Multi-Learner Predator/Prey Scenarios



Published: 2010 Juli
Herausgeber: Martin Pelikan, Jürgen Branke
Buchtitel: Proceedings of the 12th Annual Conference on Genetic and Evolutionary Computation (GECCO 2010)
Seiten: 1015-1022
Verlag: ACM
Erscheinungsort: New York, NY, USA

Referierte Veröffentlichung

BibTeX

Kurzfassung
Learning classifier systems (LCSs) are rule-based evolutionary reinforcement learning systems. Today, especially variants of Wilson’s extended classifier system (XCS) are widely applied for machine learning. Despite their widespread application, LCSs have drawbacks, e. g., in multi-learner scenarios, since the Markov property is not fulfilled.

In this paper, LCSs are investigated in an instance of the generic homogeneous and non-communicating predator/prey scenario. A group of predators collaboratively observe a (randomly) moving prey as long as possible, where each predator is equipped with a single, independent XCS. Results show that improvements in learning are achieved by cleverly adapting a multi-step approach to the characteristics of the investigated scenario. Firstly, the environmental reward function is expanded to include sensory information. Secondly, the learners are equipped with a memory to store and analyze the history of local actions and given payoffs.

ISBN: 978-1-4503-0072-8
DOI Link: 10.1145/1830483.1830669

Projekt

OCCSOCCS (Phase III)



Forschungsgruppe

Effiziente Algorithmen


Forschungsgebiet

Organic Computing, Maschinelles Lernen, Agentensysteme