A New Architecture of XCS to Approximate Real-Valued Functions Based on High Order Polynomials Using Variable-Length GA

Authors:
Ali Hamzeh;Adel Rahmani
Affiliations:
Iran University of Science and Technology, Iran;Iran University of Science and Technology, Iran
Venue:
ICNC '07 Proceedings of the Third International Conference on Natural Computation - Volume 03
Year:
2007

Citing 0
Cited 2

Learning classifier systems: a complete introduction, review, and roadmap

Journal of Artificial Evolution and Applications
Learning classifier systems: a complete introduction, review, and roadmap

Journal of Artificial Evolution and Applications

Quantified Score

Hi-index	0.00

Visualization

Abstract

XCSFG is a new version of XCSF with the ability of computing the environmental payoff using Genetic Algorithm. In the first version of XCSFG, this computation was done by evolving coefficients of the associated linear payoff functions. In this paper, we extend XCSFG to approximate the payoff functions in the form of higher order polynomials. Our newly proposed method uses GA with variable-length chromosome with real-valued representation. This new version of XCSFG is called XCSFG, the Continuous version or XCSFGC. It is applied to some benchmark problems and is compared with the original XCSF and its newly introduced extensions.