New Perspectives about Default Hierarchies Formation in Learning Classifier Systems

@inproceedings{Dorigo1991NewPA,
  title={New Perspectives about Default Hierarchies Formation in Learning Classifier Systems},
  author={Marco Dorigo},
  booktitle={AI*IA},
  year={1991}
}
In this paper we present some results of research in default hierarchies formation. A default hierarchy is a set of rules that models a set of environmental states in which some default rules cover most of the environmental states while specific ones cover exceptions. It is well known that default hierarchies can be used to categorize environmental states very efficiently. In fact, a default hierarchy implements a quasihomomorphic model of the world, which usually requires far less rules than a… CONTINUE READING

From This Paper

Topics from this paper.

References

Publications referenced by this paper.
Showing 1-4 of 4 references

Reinforcement Learning with Classifier Systems", Proceedings of Congress on AI, Simulation and Planning in High Autonomous Systems,University of Arizona, Tucson, March-1990

  • D. E. Goldberg, R. E. Smith
  • 1990
3 Excerpts

A Critical Review of Classifier Systems

  • D. E. Goldberg, S. W. Wilson
  • Proceedings of the Third International…
  • 1989
2 Excerpts

Similar Papers

Loading similar papers…