[10,22] presented various ways for introducing quasi-random numbers or derandomization in evolution strategies, with in some cases some spectacular claims on the fact that the proposed technique was always and for all criteria better than standard mutations. We here focus on the quasi-random trick and see to which extent this technique is efficient, by an in-depth analysis including convergence rates, local minima, plateaus, non-asymptotic behavior and noise. We conclude to the very stable, efficient and straightforward applicability of quasi-random numbers in continuous evolutionary algorithms.


Evolution Strategies Derandomization 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Atanassov, E.I.: On the discrepancy of the halton sequences. Math. Balkanica 18(12), 15–32 (2004)zbMATHMathSciNetGoogle Scholar
  2. 2.
    Braaten, E., Weller, G.: An improved low-discrepancy sequence for multidimensional quasi-monte carlo integration. J. Comput. Phys. 33, 249–258 (1979)CrossRefzbMATHGoogle Scholar
  3. 3.
    Bratley, P., Fox, B.: Algorithm 659: Implementing sobol’s quasirandom sequence generator. ACM Transactions on Mathematical Software 14(1), 88–100 (1988)CrossRefzbMATHGoogle Scholar
  4. 4.
    Cervellera, C., Muselli, M.: A deterministic learning approach based on discrepancy. In: Apolloni, B., Marinaro, M., Tagliaferri, R. (eds.) WIRN 2003. LNCS, vol. 2859, pp. 53–60. Springer, Heidelberg (2003)CrossRefGoogle Scholar
  5. 5.
    Cranley, R., Patterson, T.N.L.: Randomization of number theoretic methods for multiple integration. SIAM J. Numer. Anal. 13(6), 904–914 (1976)CrossRefzbMATHMathSciNetGoogle Scholar
  6. 6.
    Faure, H.: Good permutations for extreme discrepancy. J. Number Theory 42, 47–56 (1992)CrossRefzbMATHMathSciNetGoogle Scholar
  7. 7.
    Hansen, N., Ostermeier, A.: Completely derandomized self-adaptation in evolution strategies. Evolutionary Computation 11(1) (2003)Google Scholar
  8. 8.
    Hardy, G.H.: On double fourier series, and especially those which represent the double zeta-function with real and incommensurable parameters. Quart. J. Mathematics 37, 53–79 (1905)zbMATHGoogle Scholar
  9. 9.
    Hickernell, F.J.: A generalized discrepancy and quadrature error bound. Math. Comp. 67, 299–322 (1998)CrossRefzbMATHMathSciNetGoogle Scholar
  10. 10.
    Kimura, S., Matsumura, K.: Genetic algorithms using low-discrepancy sequences. In: GECCO, pp. 1341–1346 (2005)Google Scholar
  11. 11.
    L’Ecuyer, P., Lemieux, C.: Recent Advances in Randomized Quasi-Monte Carlo Methods, pp. 419–474. Kluwer Academic Publishers, Dordrecht (2002)Google Scholar
  12. 12.
    Mascagni, M., Chi, H.: On the scrambled halton sequence. Monte Carlo Methods Appl. 10(3), 435–442 (2004)zbMATHMathSciNetGoogle Scholar
  13. 13.
    Morokoff, W.J., Caflish, R.E.: Quasi-random sequences and their discrepancies. SIAM J. Sci. Comput. 15(6), 1251–1279 (1994)CrossRefzbMATHMathSciNetGoogle Scholar
  14. 14.
    Niederreiter, H.: Random Number Generation and Quasi-Monte Carlo Methods (1992)Google Scholar
  15. 15.
    Okten, G., Srinivasan, A.: Parallel quasi-monte carlo methods on a heterogeneous cluster. In: Niederreiter, H., Fang, K.-T., Hickernell, F.J. (eds.) Monte Carlo and Quasi-Monte Carlo Methods 2000, pp. 406–421. Springer, Heidelberg (2002)CrossRefGoogle Scholar
  16. 16.
    Owen, A.B.: Quasi-Monte Carlo Sampling, A Chapter on QMC for a SIGGRAPH 2003 course (2003)Google Scholar
  17. 17.
    Sarkar, P.K., Prasad, M.A.: A comparative study of pseudo and quasi random sequences for the solution of integral equations. J. Computational Physics 68, 66–88 (1978)CrossRefMathSciNetGoogle Scholar
  18. 18.
    Sloan, I.H., Woźniakowski, H.: When are quasi-Monte Carlo algorithms efficient for high dimensional integrals? Journal of Complexity 14(1), 1–33 (1998)CrossRefzbMATHMathSciNetGoogle Scholar
  19. 19.
    Sobol, I.M.: On the systematic search in a hypercube. Siam journal on Numerical Analysis 16(5), 790–793 (1979)CrossRefzbMATHMathSciNetGoogle Scholar
  20. 20.
    Srinivasan, A.: Parallel and distributed computing issues in pricing financial derivatives through quasi-monte carlo. In: Proceedings of the 16th International Parallel and Distributed Processing Symposium (2002)Google Scholar
  21. 21.
    Teytaud, O., Gelly, S., Mary, J.: Active learning in regression, with application to stochastic dynamic programming. In: Proceedings of ICINCO 2007, pp. 47–54 (2007)Google Scholar
  22. 22.
    Teytaud, O., Gelly, S.: Dcma: yet another derandomization in covariance-matrix-adaptation. In: GECCO 2007: Proceedings of the 9th annual conference on Genetic and evolutionary computation, pp. 955–963. ACM Press, New York (2007)Google Scholar
  23. 23.
    Tuffin, B.: On the use of low discrepancy sequences in monte carlo methods (1996)Google Scholar
  24. 24.
    Tuffin, B.: A new permutation choice in halton sequences. Monte Carlo and Quasi-Monte Carlo 127, 427–435 (1997)MathSciNetGoogle Scholar
  25. 25.
    van der Corput, J.G.: Verteilungsfunktionen. Proc. Ned. Akad. v. Wet. 38, 813–821 (1935)Google Scholar
  26. 26.
    Vandewoestyne, B., Cools, R.: Good permutations for deterministic scrambled halton sequences in terms of l2-discrepancy. Computational and Applied Mathematics 189(1,2), 341–361 (2006)CrossRefzbMATHMathSciNetGoogle Scholar
  27. 27.
    Wang, X., Hickernell, F.: Randomized halton sequences. Math. Comput. Modelling 32, 887–899 (2000)CrossRefzbMATHMathSciNetGoogle Scholar
  28. 28.
    Warnock, T.T.: Computational investigations of low-discrepancy point sets. In: Zaremba, S.K. (ed.) Applications of Number Theory to Numerical Analysis (Proceedings of the Symposium). University of Montreal, pp. 319–343 (1972)Google Scholar
  29. 29.
    Warnock, T.T.: Computational investigations of low-discrepancy point sets ii. In: Niederreiter, H., Shiue, P.J.-S. (eds.) Monte Carlo and Quasi-Monte Carlo Methods in Scientific Computing. Springer, Berlin (1995)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Olivier Teytaud
    • 1
  1. 1.TAO (Inria), LRI, UMR 8623(CNRS - Univ. Paris-Sud), Bât 490, Univ. Paris-SudOrsayFrance

Personalised recommendations