Advertisement

Improving XCS Performance by Distribution

  • Urban Richter
  • Holger Prothmann
  • Hartmut Schmeck
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5361)

Abstract

Learning Classifier Systems (LCSs) are rule-based evolutionary reinforcement learning (RL) systems. Today, especially variants of Wilson’s eXtended Classifier System (XCS) are widely applied for machine learning. Despite their widespread application, LCSs have drawbacks: The number of reinforcement cycles an LCS requires for learning largely depends on the complexity of the learning task. A straightforward way to reduce this complexity is to split the task into smaller sub-problems. Whenever this can be done, the performance should be improved significantly. In this paper, a nature-inspired multi-agent scenario is used to evaluate and compare different distributed LCS variants. Results show that improvements in learning speed can be achieved by cleverly dividing a problem into smaller learning sub-problems.

Keywords

Noise Signal Real Robot Emergent Behaviour Learn Classifier System Chicken Population 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Richter, U., Mnif, M., Branke, J., Müller-Schloer, C., Schmeck, H.: Towards a generic observer/controller architecture for Organic Computing. In: INFORMATIK 2006 – Informatik für Menschen!, pp. 112–119. Köllen Verlag (2006)Google Scholar
  2. 2.
    Wilson, S.W.: Classifier fitness based on accuracy. Evolutionary Computation 3(2), 149–175 (1995)CrossRefGoogle Scholar
  3. 3.
    Kovacs, T.: Learning classifier systems resources. Soft Computing 6(3–4), 240–243 (2002)CrossRefzbMATHGoogle Scholar
  4. 4.
    Dam, H.H., Abbass, H.A., Lokan, C.: Be real! XCS with continuous-valued inputs. In: Proceedings of the 2005 Workshops on Genetic and Evolutionary Computation (GECCO 2005), pp. 85–87. ACM, New York (2005)CrossRefGoogle Scholar
  5. 5.
    Barry, A.: Hierarchy formation within classifier systems – A review. In: Proceedings of the 1st International Conference on Evolutionary Algorithms an their Applications (EVCA 1996), Moscow, pp. 195–211 (1996)Google Scholar
  6. 6.
    Baneamoon, S.M., Salam, R.A., Talib, A.Z.H.: Learning process enhancement for robot behaviors. Int. Journal of Intelligent Technology 2(3), 172–177 (2007)Google Scholar
  7. 7.
    Dorigo, M.: Alecsys and the autonomouse: Learning to control a real robot by distributed classifier systems. Machine Learning 19(3), 209–240 (1995)Google Scholar
  8. 8.
    Dam, H.H., Abbass, H.A., Lokan, C.: DXCS: An XCS system for distributed data mining. In: Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2005), pp. 1883–1890. ACM, New York (2005)CrossRefGoogle Scholar
  9. 9.
    Gershoff, M., Schulenburg, S.: Collective behavior based hierarchical XCS. In: Proceedings of the 2007 Genetic And Evolutionary Computation Conference (GECCO 2007), pp. 2695–2700. ACM, New York (2007)CrossRefGoogle Scholar
  10. 10.
    Mnif, M., Richter, U., Branke, J., Schmeck, H., Müller-Schloer, C.: Measurement and control of self-organised behaviour in robot swarms. In: Lukowicz, P., Thiele, L., Tröster, G. (eds.) ARCS 2007. LNCS, vol. 4415, pp. 209–223. Springer, Heidelberg (2007)CrossRefGoogle Scholar
  11. 11.
    Mnif, M., Müller-Schloer, C.: Quantitative emergence. In: Proceedings of the 2006 IEEE Mountain Workshop on Adaptive and Learning Systems (IEEE SMCals 2006), pp. 78–84 (July 2006)Google Scholar
  12. 12.
    Richter, U., Mnif, M.: Learning to control the emergent behaviour of a multi-agent system. In: Proceedings of the 2008 Workshop on Adaptive Learning Agents and Multi-Agent Systems at AAMAS 2008 (ALAMAS+ALAg 2008), pp. 33–40 (May 2008)Google Scholar
  13. 13.
    Butz, M.V.: XCSJava 1.0: An implementation of the XCS classifier system in Java. Technical Report 2000027, Illinois Genetic Algorithms Laboratory (2000)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Urban Richter
    • 1
  • Holger Prothmann
    • 1
  • Hartmut Schmeck
    • 1
  1. 1.Karlsruhe Institute of Technology – Institute AIFBKarlsruheGermany

Personalised recommendations