Advertisement

Are long path problems hard for genetic algorithms?

  • Christian Höhn
  • Colin Reeves
Theoretical Foundations of Evolutionary Computation
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1141)

Abstract

Long path problems have been introduced into GA literature as problems that are hard for hill-climbers but empirical results suggest they can easily be solved by GAs. However, no formal analysis has been carried out which would explain this apparent success. The paper shows that the Root2path problem resembles a Royal Road function which allows the GA to exploit short cuts in the crossover landscape.

To fully understand the GA behaviour the original problem as introduced by Horn et. al. has been decomposed into its components: the Root2path and the slope function. Although both functions may be classified as GA-easy, they impose different requirements on the constitution of the population. These requirements are shown to be incompatible, so that the combination of Root2path and slope function results in a problem that is hard for a GA.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    J. Horn, D.E. Goldberg, and K. Deb. Long path problems. In Yuval Davidor, Hans-Paul Schwefel, and Reinhard Männer, editors, Parallel Problem Solving from Nature, volume 3. Springer-Verlag, Berlin, 1994.Google Scholar
  2. 2.
    M. Mitchell, S. Forrest, and J.H. Holland. The royal road for genetic algorithms: Fitness landscapes and genetic algorithm performance. In F.J. Varela and P. Bourgine, editors, Toward A Practice of Autonomous Systems: Proceeding of the First European Conference on Artificial Life. MIT Press, Cambridge, MA, 1991.Google Scholar
  3. 3.
    Melanie Mitchell, John H. Holland, and Stephanie Forrest. When will a genetic algorithm outperform hill-climbing? In J.D. Cowan, G. Tesauro, and J. Alspector, editors, Advances in Neural Information Processing Systems 6. Morgan Kaufmann, San Mateo, CA, 1994.Google Scholar
  4. 4.
    J.C. Culberson. Mutation-crossover isomorphisms and the construction of discriminating functions. Evolutionary Computation, 2(3), 1995.Google Scholar
  5. 5.
    C.R. Reeves and C. Höhn. Integrating local search into genetic algorithms. In Applied Decision Technologies., volume 2, London, 1995.Google Scholar
  6. 6.
    T. Jones. Crossover, macromutation, and population-based search. In Larry J. Eshelman, editor, Proceedings of the Sixth International Conference on Genetic Algorithms. Morgan Kaufmann, San Mateo, 1995.Google Scholar
  7. 7.
    D.H. Ackley. A Connectionist Machine for Genetic Hillclimbing. Kluwer Academic Publishers, Boston, 1987.Google Scholar
  8. 8.
    C. Höhn and C.R. Reeves. The genetic algorithm landscape for the onemax problem. Submitted to Second Nordic Workshop on Genetic Algorithms, 1996.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1996

Authors and Affiliations

  • Christian Höhn
    • 1
  • Colin Reeves
    • 2
  1. 1.Institut für Grundlagen Elektrotechnik/ElektronikTU DresdenGermany
  2. 2.School of Mathematical and Information SciencesCoventry UniversityUK

Personalised recommendations