ZCS Redux
- 1 June 2002
- journal article
- research article
- Published by MIT Press in Evolutionary Computation
- Vol. 10 (2), 185-205
- https://doi.org/10.1162/106365602320169848
Abstract
Learning classifier systems traditionally use genetic algorithms to facilitate rule discovery, where rule fitness is payoff based. Current research has shifted to the use of accuracy-based fitness. This paper re-examines the use of a particular payoff-based learning classifier system—ZCS. By using simple difference equation models of ZCS, we show that this system is capable of optimal performance subject to appropriate parameter settings. This is demonstrated for both single- and multistep tasks. Optimal performance of ZCS in well-known, multistep maze tasks is then presented to support the findings from the models.Keywords
This publication has 9 references indexed in Scilit:
- Toward Optimal Classifier System Performance in Non-Markov EnvironmentsEvolutionary Computation, 2000
- Convergence Results for Single-Step On-Policy Reinforcement-Learning AlgorithmsMachine Learning, 2000
- Classifier Fitness Based on AccuracyEvolutionary Computation, 1995
- Adding Temporary Memory to ZCSAdaptive Behavior, 1994
- ZCS: A Zeroth Level Classifier SystemEvolutionary Computation, 1994
- Implicit Niching in a Learning Classifier System: Nature's WayEvolutionary Computation, 1994
- Classifier systems and the animat problemMachine Learning, 1987
- COGNITIVE SYSTEMS BASED ON ADAPTIVE ALGORITHMS11Research reported in this paper was supported in part by the National Science Foundation under grant DCR 71-01997 and by the Horace H. Rackham School of Graduate Studies under grant 387156.Published by Elsevier BV ,1978
- Some Studies in Machine Learning Using the Game of CheckersIBM Journal of Research and Development, 1959