Probabilistic Adaptive Mapping Developmental Genetic Programming (PAM DGP): A New Developmental Approach

  • Garnett Wilson
  • Malcolm Heywood
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4193)

Abstract

Developmental Genetic Programming (DGP) algorithms have been introduced where the search space for a problem is divided into genotypes and corresponding phenotypes that are connected by a mapping (or “genetic code”). Current implementations of this concept involve evolution of the mappings in addition to the traditional evolution of genotypes. We introduce the latest version of Probabilistic Adaptive Mapping DGP (PAM DGP), a robust and highly customizable algorithm that overcomes performance problems identified for the latest competing adaptive mapping algorithm. PAM DGP is then shown to outperform the competing algorithm on two non-trivial regression benchmarks.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Banzhaf, W.: Genotype-Phenotype Mapping and Neutral Variation. In: Davidor, Y., Schwefel, H.-P., Manner, R. (eds.) Proceedings of Parallel Problem Solving from Nature III, pp. 322–332. Springer, Heidelberg (1994)Google Scholar
  2. 2.
    Keller, R., Banzhaf, W.: Genetic Programming using Genotype-Phenotype Mapping from Linear Genomes in Linear Phenotypes. In: Koza, J., et al. (eds.) Genetic Programming 1996. Proceedings of the First Annual Conference, pp. 116–122. MIT Press, Cambridge (1996)Google Scholar
  3. 3.
    Keller, R., Banzhaf, W.: The Evolution of Genetic Code in Genetic Programming. In: Banzhaf, W., et al. (eds.) Proceedings of the Genetic and Evolutionary Computation Conference—GECCO 1999, pp. 1077–1082. Morgan Kaufmann, San Francisco (1999)Google Scholar
  4. 4.
    Keller, R., Banzhaf, W.: Evolution of Genetic Code on a Hard Problem. In: Spector, L., et al. (eds.) Proceedings of the Genetic and Evolutionary Computation Conference—GECCO 2001, pp. 50–56. Morgan Kaufmann, San Francisco (2001)Google Scholar
  5. 5.
    Margetts, S.: Adaptive Genotype to Phenotype Mappings for Evolutionary Algorithms. Ph. D. thesis, Cardiff University, Wales, Great Britain (2001)Google Scholar
  6. 6.
    Margetts, S., Jones, A.: Phlegmatic Mappings for Functional Optimisation with Genetic Programming. In: Whitley, L., et al. (eds.) Proceedings of the Genetic and Evolutionary Computation Conference—GECCO 2000, pp. 82–89. Morgan Kaufmann, San Francisco (2000)Google Scholar
  7. 7.
    Margetts, S., Jones, A.: An Adaptive Mapping for Developmental Genetic Programming. In: Miller, J., Tomassini, M., Lanzi, P.L., Ryan, C., Tetamanzi, A.G.B., Langdon, W.B. (eds.) EuroGP 2001. LNCS, vol. 2038, pp. 97–107. Springer, Heidelberg (2001)CrossRefGoogle Scholar
  8. 8.
    Wilson, G., Heywood, M.: Probabilistic (Genotype) Adaptive Mapping Combinations for Developmental Genetic Programming. In: Proceedings of the IEEE Congress on Evolutionary Computation—CEC 2006, IEEE Press, Los Alamitos (2006) (upcoming)Google Scholar
  9. 9.
    Koza, J.: Genetic Programming II: Automatic Discovery of Reusable Programs. MIT Press, Cambridge (1994)MATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Garnett Wilson
    • 1
  • Malcolm Heywood
    • 1
  1. 1.Faculty of Computer ScienceDalhousie UniversityHalifaxCanada

Personalised recommendations