Genetic Programming and Evolvable Machines

, Volume 9, Issue 2, pp 109–123 | Cite as

Quotients of Markov chains and asymptotic properties of the stationary distribution of the Markov chain associated to an evolutionary algorithm

  • Boris Mitavskiy
  • Jonathan E. Rowe
  • Alden Wright
  • Lothar M. Schmitt
Original Paper

Abstract

In this work, a method is presented for analysis of Markov chains modeling evolutionary algorithms through use of a suitable quotient construction. Such a notion of quotient of a Markov chain is frequently referred to as “coarse graining” in the evolutionary computation literature. We shall discuss the construction of a quotient of an irreducible Markov chain with respect to an arbitrary equivalence relation on the state space. The stationary distribution of the quotient chain is “coherent” with the stationary distribution of the original chain. Although the transition probabilities of the quotient chain depend on the stationary distribution of the original chain, we can still exploit the quotient construction to deduce some relevant properties of the stationary distribution of the original chain. As one application, we shall establish inequalities that describe how fast the stationary distribution of Markov chains modeling evolutionary algorithms concentrates on the uniform populations as the mutation rate converges to 0. Further applications are discussed. One of the results related to the quotient construction method is a significant improvement of the corresponding result of the authors’ previous conference paper [Mitavskiy et al. (2006) In: Simulated Evolution and Learning, Proceedings of SEAL 2006, Lecture Notes in Computer Science v. 4247, Springer Verlag, pp 726–733]. This papers implications are all strengthened accordingly.

Keywords

Markov chain Stationary distribution Quotient Coarse graining Evolutionary algorithm Uniform population Asymptotics Mutation rate Selection pressure 

References

  1. 1.
    Schmitt, L.: Theory of genetic algorithms. Theor. Comput. Sci. 259, 1–61 (2001)MATHCrossRefGoogle Scholar
  2. 2.
    Schmitt, L.: Theory of genetic algorithms ii: models for genetic operators over the string-tensor representation of populations and convergence to global optima for arbitrary fitness function under scaling. Theor. Comput. Sci. 310, 181–231 (2004)MATHCrossRefGoogle Scholar
  3. 3.
    Vose, M.: The Simple Genetic Algorithm: Foundations and Theory. MIT Press (1999)Google Scholar
  4. 4.
    Coffey, S.: An Applied Probabilist’s Guide to Genetic Algorithms. A Thesis Submitted to The University of Dublin for the degree of Master in Science (1999)Google Scholar
  5. 5.
    Mitavskiy, B., Rowe, J.: An Extension of Geiringer Theorem for a wide class of evolutionary search algorithms. Evol. Comput. 14(1), 87–118 (2006)MathSciNetGoogle Scholar
  6. 6.
    Mitavskiy, B.: A schema-based version of Geiringer’s Theorem for nonlinear Genetic Programming with homologous crossover. Found. Gen. Algorithms 8, 156–175 Springer-Verlag (2005)MathSciNetCrossRefGoogle Scholar
  7. 7.
    Mitavskiy, B., Rowe, J.: Some results about the Markov chains associated to GPs and general EAs. Theor. Comput. Sci. 361(1), 72–110 (2006)MATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    Davis, T., Principe, J.: A simulated annealing like convergence theory for the simple genetic algorithm. In: Belew, R., Bookers, L. (eds.) Proceedings of the Fourth International Conference on Genetic Algorithms, pp. 174–181. Morgan Kaufmann (1991)Google Scholar
  9. 9.
    Suzuki, J.: A further result on the Markov chain model of genetic algorithm and its application to simulated annealing-like strategy. IEEE Trans. Syst. Man Cybernet. 28(1), 95–102 (1998)CrossRefGoogle Scholar
  10. 10.
    Mitavskiy, B., Rowe, J., Wright, A., Schmitt, L.: Exploiting quotients of Markov chains to derive properties of the stationary distribution of the markov chain associated to an evolutionary algorithm. In: Simulated Evolution and Learning, Proceedings of SEAL 2006, Lecture Notes in Computer Science v. 4247, pp. 726–733. Springer Verlag (2006)Google Scholar
  11. 11.
    Aldous, D., Fill, J.: Reversible Markov Chains and Random Walks on Graphs (unpublished book) (2001) http://www.stat.berkeley.edu/
  12. 12.
    Rowe, J., Vose, M., Wright, A.: Differentiable coarse graining. Theor. Comput. Sci. 361, 111–129 (2006)MATHCrossRefMathSciNetGoogle Scholar
  13. 13.
    Spears, W., De Jong, K.: Analyzing GAs using Markov chain models with semantically ordered and lumped states. In: Belew, R., Vose, M. (eds.) Foundations of Genetic Algorithms 4, pp. 85–100. Morgan Kaufmann (1997)Google Scholar
  14. 14.
    Moey, C., Rowe, J.: Population aggregation based on fitness. Nat. Comput. 3, 5–19 (2004)MATHCrossRefMathSciNetGoogle Scholar
  15. 15.
    Moey, C.: Investigating state aggregation for Markov models of genetic algorithms. Thesis, The University of Birmingham (2006)Google Scholar
  16. 16.
    Radcliffe, N.: The algebra of genetic algorithms. Ann. Math. Artif. Intel. 10, 339–384 (1994)MATHCrossRefMathSciNetGoogle Scholar
  17. 17.
    Mitavskiy, B.: Crossover invariant subsets of the search space for evolutionary algorithms. Evol. Comput. 12(1), 19–46 (2004)CrossRefGoogle Scholar
  18. 18.
    Antonisse, J.: A new interpretation of schema notation that overturns the binary encoding constraint. In: Schaffer, J.D. (ed.) Proceedings of the Third International Conference on Genetic Algorithms, pp. 86–97. Morgan Kaufmann (1989)Google Scholar

Copyright information

© Springer Science+Business Media, LLC 2007

Authors and Affiliations

  • Boris Mitavskiy
    • 1
  • Jonathan E. Rowe
    • 2
  • Alden Wright
    • 3
  • Lothar M. Schmitt
    • 4
  1. 1.School of MedicineUniversity of SheffieldSheffieldGreat Britain
  2. 2.School of Computer ScienceUniversity of BirminghamBirminghamGreat Britain
  3. 3.Department of Computer ScienceUniversity of MontanaMissoulaUSA
  4. 4.School of computer Science and EngineeringThe University of AizuAizu-Wakamatsu CityJapan

Personalised recommendations