Evaluation and Diversity in Co-evolution

  • Rients P. T. van Wijngaarden
  • Edwin D. de Jong
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5199)


This paper studies the performance of four alternative evaluation methods; two instances of the Exponential Moving average, the Elo-rating and the Glicko-rating method. These methods are tested in a co-evolutionary setup using the LINT-game, which is known to be problematic under co-evolutionary conditions. Besides the different evaluation approaches, two methods aimed at preserving diversity are tested. By using the Objective Fitness Correlation as an analytical tool for monitoring accuracy of evaluation, it is shown that actual performance of an evaluation method strongly depends on whether co-evolutionary failure occurs and that a multi-modal approach to the LINT-problem is effective in maintaining stable progress over time.


Evaluation Method Rate Deviation Solution Concept Player Rating Learner Score 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Sevan, G.: Ficici. Solution Concepts in Coevolutionary Algorithms. PhD thesis, Brandeis University (2004)Google Scholar
  2. 2.
    Watson, R.A., Pollack, J.B.: Coevolutionary dynamics in an minimal substrate. In: Spector, L., et al. (eds.) Proceedings of the Genetic and Evolutionary Computation Conference, GECCO 2001, pp. 702–709. Morgan Kaufman, San Fransisco (2001)Google Scholar
  3. 3.
    De Jong, E.D.: Objective fitness correlation. In: Thierens, D., et al. (eds.) Proceedings of the Genetic and Evolutionary Computation Conference, GECCO 2007, pp. 440–447. ACM press, New York (2007)Google Scholar
  4. 4.
    Watson, R.A.: Personal communication (2003)Google Scholar
  5. 5.
    de Jong, E.D.: Intransitivity in Coevolution. In: Yao, X., Burke, E.K., Lozano, J.A., Smith, J., Merelo-Guervós, J.J., Bullinaria, J.A., Rowe, J.E., Tiňo, P., Kabán, A., Schwefel, H.-P. (eds.) PPSN 2004. LNCS, vol. 3242, pp. 843–851. Springer, Heidelberg (2004)CrossRefGoogle Scholar
  6. 6.
    Chou, Y.-l.: Statistical Analysis. Holt International (1975) ISBN 0030894220Google Scholar
  7. 7.
    Elo, A.E.: The Rating of Chessplayers: Past and Present. Arco (1978)Google Scholar
  8. 8.
    Glickman, M.E.: Parameter estimation in large dynamic paired comparison experiments. Applied Statistics 48(3), 377–394 (1999)zbMATHGoogle Scholar
  9. 9.
    Ficici, S.G., Pollack, J.B.: Pareto optimality in coevolutionary learning. In: Kelemen, J., Sosík, P. (eds.) ECAL 2001. LNCS (LNAI), vol. 2159, pp. 316–325. Springer, Heidelberg (2001)CrossRefGoogle Scholar
  10. 10.
    Rosin, C.D., Belew, R.K.: New methods for competitive coevolution. Evolutionary Computation 5(1), 1–29 (1997)CrossRefGoogle Scholar
  11. 11.
    Eiben, A.E., Smith, J.E.: Introduction to Evolutionary Computing. Natural Computing. Springer, Berlin (2003)CrossRefzbMATHGoogle Scholar
  12. 12.
    Mahfoud, S.W.: Crowding and preselection revisited. In: Manner, R., Manderick, B. (eds.) Parallel Problem Solving from Nature 2, PPSN-II (2nd PPSN 1992), pp. 27–36. Elsevier Science Publishers, Amsterdam (1992)Google Scholar
  13. 13.
    Harik, G.R.: Finding multimodal solutions using restricted tournament selection. In: Eshelman, L.J. (ed.) Proc. of the Sixth Int. Conf. on Genetic Algorithms, pp. 24–31. Morgan Kaufmann, San Francisco (1995)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Rients P. T. van Wijngaarden
    • 1
  • Edwin D. de Jong
    • 1
  1. 1.Algorithmic Data Analysis GroupUniversiteit UtrechtThe Netherlands

Personalised recommendations