Novelty-Based Fitness: An Evaluation under the Santa Fe Trail

  • John Doucette
  • Malcolm I. Heywood
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6021)

Abstract

We present an empirical analysis of the effects of incorporating novelty-based fitness (phenotypic behavioral diversity) into Genetic Programming with respect to training, test and generalization performance. Three novelty-based approaches are considered: novelty comparison against a finite archive of behavioral archetypes, novelty comparison against all previously seen behaviors, and a simple linear combination of the first method with a standard fitness measure. Performance is evaluated on the Santa Fe Trail, a well known GP benchmark selected for its deceptiveness and established generalization test procedures. Results are compared to a standard quality-based fitness function (count of food eaten). Ultimately, the quality style objective provided better overall performance, however, solutions identified under novelty based fitness functions generally provided much better test performance than their corresponding training performance. This is interpreted as representing a requirement for layered learning/ symbiosis when assuming novelty based fitness functions in order to more quickly achieve the integration of diverse behaviors into a single cohesive strategy.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Gomez, F.J.: Sustaining diversity using behavioral information distance. In: Proceedings of the Genetic and Evolutionary Computation Conference, pp. 113–120. ACM, New York (2009)CrossRefGoogle Scholar
  2. 2.
    Koza, J.: Genetic Programming: On the Programming of Computers by Means of Natural Selection. MIT Press, Cambridge (1992)MATHGoogle Scholar
  3. 3.
    Kushchu, I.: Genetic programming and evolutionary generalization. IEEE Transactions on Evolutionary Computation 6(5), 431–442 (2002)CrossRefGoogle Scholar
  4. 4.
    Langdon, W.B., Poli, R.: Foundations of Genetic Programming. Springer, Heidelberg (2002)MATHGoogle Scholar
  5. 5.
    Lehman, J., Stanley, K.O.: Exploiting open-endedness to solve problems through the search for novelty. In: Proceedings of the International Conference on Artificial Life XI. MIT Press, Cambridge (2008)Google Scholar
  6. 6.
    Lichodzijewski, P., Heywood, M.I.: Managing team-based problem solving with symbiotic bid-based Genetic Programming. In: Proceedings of the Genetic and Evolutionary Computation Conference, pp. 363–370 (2008)Google Scholar
  7. 7.
    Liu, Y., Yao, X., Higuchi, T.: Evolutionary Ensembles with Negative Correlation Learning. IEEE Transactions on Evolutionary Computation 4(4), 380–387 (2000)CrossRefGoogle Scholar
  8. 8.
    McIntyre, A.R., Heywood, M.I.: Cooperative problem decomposition in pareto competitive classifier models of coevolution. In: O’Neill, M., Vanneschi, L., Gustafson, S., Esparcia Alcázar, A.I., De Falco, I., Della Cioppa, A., Tarantino, E. (eds.) EuroGP 2008. LNCS, vol. 4971, pp. 289–300. Springer, Heidelberg (2008)CrossRefGoogle Scholar
  9. 9.
    Mouret, J.-B., Doncieux, S.: Overcoming the bootstrap problem in evolutionary robotics using behavioral diversity. In: IEEE Congress on Evolutionary Computation, pp. 1161–1168 (2009)Google Scholar
  10. 10.
    Risi, S., Vanderbleek, S.D., Hughes, C.E., Stanley, K.O.: How novelty search escapes the deceptive trap of learning to learn. In: Proceedings of the Genetic and Evolutionary Computation Conference, pp. 153–160. ACM Press, New York (2009)CrossRefGoogle Scholar
  11. 11.
    Zongker, D., Punch, B.: lil-gp 1.0 User’s Manual. Michigan State University (1995)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • John Doucette
    • 1
  • Malcolm I. Heywood
    • 1
  1. 1.Faculty of Computer ScienceDalhousie UniversityHalifaxCanada

Personalised recommendations