Abstract
In this paper, we present how a reinforcement learning approach (ELF — Evolutionary Learning of Fuzzy rules) implements an extension of the popular Q-Learning algorithm to Fuzzy Classifier Systems. We discuss how chains of fuzzy rules may be identified by an evolutionary learning system that provides delayed reinforcement.
We mention other recent proposals for Fuzzy Q-Learning, and we point out how ELF is more efficient, and more suitable to learn behaviors for autonomous agents in unstructured, real environments.
We conclude the paper with an example of the application of the Fuzzy Q-Learning features of ELF to learning a fuzzy system that implements a behavior for an autonomous agent.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
H. R. Berenji, Fuzzy Q-Learning: a new approach for fuzzy dynamic programming. Proc. Third IEEE Int. Conf. on Fuzzy Systems. IEEE Computer Press, Piscataway, NJ, pp. 486–491, 1994.
A Bonarini, ELF: learning incomplete fuzzy rule sets for an autonomous robot, Proc. of EUFIT '93, ELITE Foundation, Aachen, Germany, pp. 69–75, 1993.
Bonarini A., Evolutionary learning of general fuzzy rules with biased evaluation functions: competition and cooperation. In Proc. of IEEE WCCI — Evolutionary Computation IEEE Computer Press, Piscataway NJ, pp. 51–56, 1994.
A. Bonarini, Some methodological issues about designing autonomous agents which learn their behaviors: the ELF experience. In R. Trappl (Ed.) Cybernetics and Systems Research '94, World Scientific Publishing, Singapore, pp. 1435–1442, 1994.
A. Bonarini, Learning to coordinate fuzzy behaviors for autonomous agents. Proc. EUFIT '94 ELITE Foundation, Aachen, Germany, pp. 475–479, 1994.
A. Bonarini, Evolutionary learning of fuzzy rules: competition and cooperation. To appear in W. Pedrycz (Ed.) Fuzzy modeling: paradigms and practice, Kluwer Academic Publishers, Norwell, MA, 1995.
J. H. Connell, S. Mahadevan, Automatic programming of behavior-based robots using reinforcement learning, Artificial Intelligence, Vol. 55, no.2, pp. 311–365, 1992.
P. Y. Glorennec, Fuzzy Q-learning and Evolutionary Strategy for adaptive fuzzy control, Proc. EUFIT '94 ELITE Foundation, Aachen, Germany, pp. 35–40, 1994.
J. J. Grefenstette, Lamarckian learning in multi-agent environmemts, Proc. of the Fourth Int. Conf. on Genetic Algorithms, Morgan Kaufmann, San Mateo, CA, pp. 346–353, 1991.
J. J. Grefenstette, The evolution of strategies for multiagent environments. Adaptive Behavior, Vol. 1, no.1, pp. 65–91, 1992.
M. Valenzuela-Rendón, The fuzzy classifier system: a classifier system for continuously varying variables. Proc. of the 4th Int. Conf. on Genetic Algorithms, Morgan Kaufmann, San Mateo, CA,pp. 346–353, 1991.
C. Watkins, P. Dayan, Q-Learning, Machine Learning, Vol. 8, pp. 279–292, 1992.
S. W. Wilson, ZCS, a zeroeth level Classifier System, Evolutionary Computation, Vol. 2, no.1, pp. 1–18, 1994
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1995 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Bonarini, A. (1995). Extending Q-learning to Fuzzy Classifier Systems. In: Gori, M., Soda, G. (eds) Topics in Artificial Intelligence. AI*IA 1995. Lecture Notes in Computer Science, vol 992. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-60437-5_3
Download citation
DOI: https://doi.org/10.1007/3-540-60437-5_3
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-60437-2
Online ISBN: 978-3-540-47468-5
eBook Packages: Springer Book Archive