Evolving prediction weights using evolution strategy

ABSTRACT

The evolution strategy is one of the strongest evolutionary algorithms for optimizing real-value vectors. In this paper, we study how to use it for the evolution of prediction weights in XCSF in order to make the computed prediction more accurate. Our version of XCSF shows to be able to evolve more accurate linear approximations of functions. It is more efficient than the original XCSF and slightly better than XCSF with recursive least squares, in spite of its simple structure and its low complexity.

REFERENCES

• 1: Butz, M. V., and Wilson, S. W., An Algorithmic Description of XCS, Soft Computing, 6(3-4), pp. 144--153, 2002.
• 2: David E. Goldberg, Genetic Algorithms in Search, Optimization and Machine Learning, Addison-Wesley Longman Publishing Co., Inc., Boston, MA, 1989
• 3: Stefan Kern , Sibylle D. Müller , Nikolaus Hansen , Dirk Büche , Jiri Ocenasek , Petros Koumoutsakos, Learning probability distributions in continuous evolutionary algorithms– a comparative review, Natural Computing: an international journal, v.3 n.1, p.77-112, 2004 [doi>10.1023/B:NACO.0000023416.59689.4e]
• 4: Pier Luca Lanzi , Daniele Loiacono , Stewart W. Wilson , David E. Goldberg, Extending XCSF beyond linear approximation, Proceedings of the 2005 conference on Genetic and evolutionary computation, June 25-29, 2005, Washington DC, USA [doi>10.1145/1068009.1068319]
• 5: Pier Luca Lanzi , Daniele Loiacono , Stewart W. Wilson , David E. Goldberg, Classifier prediction based on tile coding, Proceedings of the 8th annual conference on Genetic and evolutionary computation, July 08-12, 2006, Seattle, Washington, USA [doi>10.1145/1143997.1144242]
• 6: Pier Luca Lanzi , Daniele Loiacono , Stewart W. Wilson , David E. Goldberg, Prediction update algorithms for XCSF: RLS, Kalman filter, and gain adaptation, Proceedings of the 8th annual conference on Genetic and evolutionary computation, July 08-12, 2006, Seattle, Washington, USA [doi>10.1145/1143997.1144243]
• 7: Loiacono, D., and Lanzi, P. L., Evolving Neural Networks for Classifier Prediction with XCSF, ECAI'2006, Workshop on Evolutionary Computation, pp. 36--40, ACM Press, New York, USA, 2006.
• 8: Rechenberg, I., Evolutionsstrategie: Optimierung technischer Systeme und Prinzipien der biologischen Evolution, Frommann-Holzboog, Stuttgart, 1973.
• 9: Hans-Paul Schwefel, Numerical Optimization of Computer Models, John Wiley & Sons, Inc., New York, NY, 1981
• 10: Hau Trung Tran , Cédric Sanza , Yves Duthen , Thuc Dinh Nguyen, XCSF with computed continuous action, Proceedings of the 9th annual conference on Genetic and evolutionary computation, July 07-11, 2007, London, England [doi>10.1145/1276958.1277327]
• 11: Tran, T. H., Approches évolutionnaires pour le comportement adaptatif d'entités autonomes, PhD thesis, University of Toulouse, IRIT lab., France, 2007.
• 12: Bernard Widrow , Marcian E. Hoff, Adaptive switching circuits, Neurocomputing: foundations of research, MIT Press, Cambridge, MA, 1988
• 13: Stewart W. Wilson, Classifier fitness based on accuracy, Evolutionary Computation, v.3 n.2, p.149-175, Summer 1995 [doi>10.1162/evco.1995.3.2.149]
• 14: Wilson, S. W., Get Real! XCS with Continuous-Valued Inputs, Festschrift in Honor of John H. Holland, pp. 111--121, 1999.
• 15: Wilson, S. W., Function Approximation with a Classifier System, GECCO'2001, Genetic and Evolutionary Computation Conference, San Francisco, CA, pp. 974--981, 2001.
• 16: Stewart W. Wilson, Classifiers that approximate functions, Natural Computing: an international journal, v.1 n.2-3, p.211-234, June 2002 [doi>10.1023/A:1016535925043]

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.6 Learning
          Subjects: Concept learning

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.6 Learning
          Subjects: Parameter learning

General Terms:
Algorithms, Performance

Keywords:
XCSF, evolution strategy, function approximation