Abstract
Wilson’s XCS has rapidly become the most popular classifier system of all time, and is a major focus of current research. XCS’s primary distinguishing feature is that it bases rule fitness on the accuracy with which rules predict reward, rather than the magnitude of the reward predicted (as traditional, strength-based systems do). XCS is a complex system and differs from other systems in a number of ways. In order to isolate the source of XCS’s adaptive power, and, in particular to study the difference between strength and accuracy-based fitness, we introduce a system called Strength-Based XCS (SB–XCS), which is as similar to the accuracy-based XCS as we could make it, apart from being strength-based. This work provides a specification of SB–XCS and initial results for it and XCS on the 6 multiplexer and woods2 tasks. It then analyses the solutions found by the two systems and finds that each prefers a particular type of solution. A sequel paper provides further analysis.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Anderson, C.W.: Learning and Problem solving with multilayer connectionist systems. PhD thesis, University of Massachusetts, Amherst, MA, USA (1986)
Bäck, T., Fogel, D.B., Michalewicz, Z. (eds.): Handbook of Evolutionary Computation. Institute of Physics Publishing/Oxford University Press (1997)
Barry, A.: XCS Performance and Population Structure within Multiple-Step Environments. PhD thesis, Queens University Belfast (2000)
Barto, A.G., Anandan, P., Anderson, C.W.: Cooperativity in networks of pattern recognizing stochastic learning automata. In: Proceedings of the Fourth Yale Workshop on Applications of Adaptive Systems Theory, pp. 85–90 (1985)
Bonelli, P., Parodi, A., Sen, S., Wilson, S.: NEWBOOLE: A Fast GBML System. In: International Conference on Machine Learning, San Mateo, California, pp. 153–159. Morgan Kaufmann, San Francisco (1990)
Booker, L.B.: Triggered rule discovery in classifier systems. In: David Schaffer, J. (ed.) Proceedings of the 3rd International Conference on Genetic Algorithms (ICGA 1989), George Mason University, pp. 265–274. Morgan Kaufmann, San Francisco (1989)
Butz, M.V., Goldberg, D.E., Stolzmann, W.: Investigating Generalization in the Anticipatory Classifier System. In: Deb, K., Rudolph, G., Lutton, E., Merelo, J.J., Schoenauer, M., Schwefel, H.-P., Yao, X. (eds.) PPSN 2000. LNCS, vol. 1917, Springer, Heidelberg (2000); Also technical report 2000014 of the Illinois Genetic Algorithms Laboratory
Butz, M.V., Kovacs, T., Lanzi, P.L., Wilson, S.W.: How XCS Evolves Accurate Classifiers. In: Spector, L., Goodman, E.D., Wu, A., Langdon, W.B., Voigt, H.-M., Gen, M., Sen, S., Dorigo, M., Pezeshk, S., Garzon, M.H., Burke, E. (eds.) GECCO-2001: Proceedings of the Genetic and Evolutionary Computation Conference, pp. 927–934. Morgan Kaufmann, San Francisco (2001)
Butz, M.V., Wilson, S.W.: An Algorithmic Description of XCS. In: Lanzi, P.L., Stolzmann, W., Wilson, S.W. (eds.) IWLCS 2000. LNCS (LNAI), vol. 1996, pp. 253–272. Springer, Heidelberg (2001)
Cribbs III, H.B., Smith, R.E.: What Can I do with a Learning Classifier System? In: Karr, C., Freeman, L.M. (eds.) Industrial Applications of Genetic Algorithms, pp. 299–320. CRC Press, Boca Raton (1998)
de Boer, B.: Classifier Systems: a useful approach to machine learning? Master’s thesis, Leiden University (1994), ftp://ftp.wi.leidenuniv.nl/pub/CS/MScTheses/deboer.94.ps.gz
De Jong, K.A., Spears, W.M.: Learning Concept Classification Rules Using Genetic Algorithms. In: Proceedings of the International Joint Conference on Artificial Intelligence, Sidney, Australia, pp. 651–656 (1991)
Goldberg, D.E.: Genetic Algorithms in Search, Optimization, and Machine Learning. Addison-Wesley, Reading (1989)
Greene, D.P., Smith, S.F.: COGIN: Symbolic induction using genetic algorithms. In: Proceedings 10th National Conference on Artificial Intelligence, pp. 111–116. Morgan Kaufmann, San Francisco (1992)
Greene, D.P., Smith, S.F.: Using Coverage as a Model Building Constraint in Learning Classifier Systems. Evolutionary Computation 2(1), 67–91 (1994)
Holmes, J.H.: Evolution-Assisted Discovery of Sentinel Features in Epidemiologic Surveillance. PhD thesis, Drexel University (1996)
Jacobs, R.A.: Increased rates of convergence through learning rate adaptation. Neural Networks 1, 295–307 (1988)
Kovacs, T.: Evolving Optimal Populations with XCS Classifier Systems. Master’s thesis, University of Birmingham (1996)
Kovacs, T.: XCS Classifier System Reliably Evolves Accurate, Complete, and Minimal Representations for Boolean Functions. In: Roy, Chawdhry, Pant (eds.) Soft Computing in Engineering Design and Manufacturing, pp. 59–68. Springer, Heidelberg (1997)
Kovacs, T.: Deletion schemes for classifier systems. In: Banzhaf, W., Daida, J., Eiben, A.E., Garzon, M.H., Honavar, V., Jakiela, M., Smith, R.E. (eds.) GECCO 1999: Proceedings of the Genetic and Evolutionary Computation Conference, pp. 329–336. Morgan Kaufmann, San Francisco (1999)
Kovacs, T.: Strength or accuracy? A comparison of two approaches to fitness calculation in learning classifier systems. In: Wu, A.S. (ed.) 1999 Genetic and Evolutionary Computation Conference Workshop Program, pp. 258–265 (1999)
Kovacs, T.: Towards a theory of strong overgeneral classifiers. In: Martin, W., Spears, W.M. (eds.) Foundations of Genetic Algorithms, vol. 6, pp. 165–184. Morgan Kaufmann, San Francisco (2001)
Kovacs, T.: What should a classifier system learn? In: Proceedings of the 2001 Congress on Evolutionary Computation (CEC 2001), pp. 775–782. IEEE Press, Los Alamitos (2001)
Kovacs, T.: A Comparison of Strength and Accuracy-Based Fitness in Learning Classifier Systems. PhD thesis, School of Computer Science, University of Birmingham (2002)
Liepins, G.E., Wang, L.A.: Classifier System Learning of Boolean Concepts. In: Belew, R.K., Booker, L.B. (eds.) Proceedings ofthe Fourth International Conference on Genetic Algorithms (ICGA 1991), San Mateo, CA, pp. 318–323. Morgan Kaufmann Publishers, San Francisco (1991)
Oblow, E.M.: Implementing Valiant’s Learnability Theory using Random Sets. Technical Report ORNL/TM-11512R, Oak Ridge National Laboratory (1990)
Pagallo, G., Haussler, D.: Boolean feature discovery in empirical learning. Machine Learning 5(1), 71–100 (1990)
Quinlan, J.R.: An empirical comparison of genetic and decision-tree classifiers. In: Proceedings of the Fifth International Machine Learning Conference, pp. 135–141 (1988)
Sen, S.: Classifier system learning of multiplexer function. The University of Alabama, Tuscaloosa, Alabama (1988) Class Project
Smith, R.E., Cribbs, H.B.: Is a Learning Classifier System a Type of Neural Network? Evolutionary Computation 2(1), 19–36 (1994)
Wang, L.A.: Classifier System Learning of the Boolean Multiplexer Function. Master’s thesis, Computer Science Department, University of Tennessee, Knoxville, TN (1990)
Wilson, S.W.: Classifier Systems and the Animat Problem. Machine Learning 2, 199–228 (1987)
Wilson, S.W.: Bid competition and specificity reconsidered. Complex Systems 2, 705–723 (1989)
Wilson, S.W.: Perceptron Redux: Emergence of Structure. Physica D, 249–256 (1990)
Wilson, S.W.: ZCS: A Zeroth Level Classifier System. Evolutionary Computation 2(1), 1–18 (1994)
Wilson, S.W.: Classifier Fitness Based on Accuracy. Evolutionary Computation 3(2), 149–175 (1995)
Wilson, S.W.: Personal communication, November 21 (1996)
Wilson, S.W.: Generalization in the XCS classifier system. In: Koza, J., Banzhaf, W., Chellapilla, K., Deb, K., Dorigo, M., Fogel, D.B., Gar-zon, M.H., Goldberg, D.E., Iba, H., Riolo, R. (eds.) Genetic Programming 1998: Proceedings of the Third Annual Conference, pp. 665–674. Morgan Kaufmann, San Francisco (1998)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2003 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Kovacs, T. (2003). XCS’s Strength-Based Twin: Part I. In: Lanzi, P.L., Stolzmann, W., Wilson, S.W. (eds) Learning Classifier Systems. IWLCS 2002. Lecture Notes in Computer Science(), vol 2661. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-40029-5_5
Download citation
DOI: https://doi.org/10.1007/978-3-540-40029-5_5
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-20544-9
Online ISBN: 978-3-540-40029-5
eBook Packages: Springer Book Archive