On Two Types of GA-Learning
Nicolaas J. Vriend, Queen Mary, University of London
in: S.-H. Chen (Ed.), Evolutionary Computation in Economics and Finance. (Studies in Fuzziness and Soft Computing 100), Physica-Verlag, Heidelberg, 2002, p. 233-243

pdf Full paper (PDF format)

See also: An Illustration of the Essential Difference between Individual and Social Learning, and its Consequences for Computational Analyses, Journal of Economic Dynamics and Control, 2000.

Abstract. We distinguish two types of learning with a Genetic Algorithm. A population learning Genetic Algorithm (or pure GA), and an individual learning Genetic Algorithm (basically a GA combined with a Classifier System). The difference between these two types of GA is often neglected, but we show that for a general class of problems this difference is essential as it may lead to widely differing performances. The underlying cause for this is a so called spite effect.

J.E.L. classification codes. C72, D43, D83

Keywords. Cournot oligopoly, Individual learning, Social learning, Spite effect, Genetic Algorithm

Nick Vriend, n.vriend@qmul.ac.uk
Last modified 2012-12-07