Simple Markov Models of the Genetic Algorithm in Classifier Systems: Multi-step Tasks

  • Authors:
  • Larry Bull

  • Affiliations:
  • -

  • Venue:
  • IWLCS '00 Revised Papers from the Third International Workshop on Advances in Learning Classifier Systems
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

Michigan-style Classifier Systems use Genetic Algorithms to facilitate rule-discovery. This paper presents a simple Markov model of the algorithm in such systems, with the aim of examining the effects of different types of interdependence between niches in multi-step tasks. Using the model it is shown that the existence of, what is here termed, partner rule variance can have significant and detrimental effects on the Genetic Algorithm's expected behaviour. Suggestions are made as to how to reduce these effects, making connections with other recent work in the area.