Advertisement

Two Approaches for Parallelizing the UEGO Algorithm

  • P. M. Ortigosa
  • I. García
  • M. Jelásity
Part of the Applied Optimization book series (APOP, volume 59)

Abstract

In this work UEGO, a new stochastic optimization technique for accelerating and/or parallelizing existing search methods is described and parallelized. The skeleton of the algorithm is a parallel hillclimber. The separate hillclimbers work in restricted search regions (or clusters) of the search space. The volume of the clusters decreases as the search proceeds which results in a cooling effect similar to simulated annealing. Вesides this, UEGO can be effectively parallelized because the amount of information exchanged among clusters (communication) is minimal. The purpose of this communication is to ensure that a hill is explored only by one hillclimber. UEGO makes periodic attempts to find new hills to climb. Two parallel algorithms of UEGO have been implemented and empirical results are also presented including an analysis of speedup.

Keywords

Global Optimization Stochastic Methods Parallelism Evolutionary Algorithms 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    D. Beasley, D. R. Bull, and R. R. Martin. A sequential niche technique for multimodal function optimization. Evolutionary Computation, 1(2):101–125, 1993.CrossRefGoogle Scholar
  2. 2.
    K. Deb. Genetic algorithms in multimodal function optimization. TCGA report no. 89002, The University of Alabama, Dept. of Engineering mechanics, 1989.Google Scholar
  3. 3.
    K. Deb and D. E. Goldberg. An investigation of niche and species formation in genetic function optimization. In J. D. Schaffer, editor, The Proceedings of the Third International Conference on Genetic Algorithms. Morgan Kaufmann, 1989.Google Scholar
  4. 4.
    A. Ferreira and P. M. Pardalos. Solving Combinatorial Optimization Problems in Parallel: Methods and Techniques, volume 1054 of Lecture Notes in Computational Science. Springer-Verlag, 1996.Google Scholar
  5. 5.
    C. A. Floudas and P. M. Pardalos (eds.). A Collection of Test Problems for Constrained Global Optimization Algorithms. Lecture Notes in Computer Science. Springer-Verlag, Berlin, 1978.Google Scholar
  6. 6.
    M. Jelasity. UEGO, an abstract niching technique for global optimization. In Agoston E. Eiben, Thomas Bäck, Marc Schoenauer, and Hans-Paul Schwefel, editors, Parallel Problem Solving from Nature - PPSN V, volume 1498 of Lecture Notes in Computational Science, pages 378–387. Springer-Verlag, 1998.Google Scholar
  7. 7.
    M. Jelasity and J. Dornbi. GAS, a concept on modeling species in genetic algorithms. Artificial Intelligence, 99(1):1–19, 1998.MathSciNetzbMATHCrossRefGoogle Scholar
  8. 8.
    M. Jelasity, P. M. Ortigosa, and I. Garcia. UEGO, an abstract clustering technique for multimodal global optimization. Journal of Heuristics, To appear.Google Scholar
  9. 9.
    A. Migdalas, P. M. Pardalos, and S. Storoy. Parallel Computing in Optimization. Kluwer Academic Publishers, 1998.Google Scholar
  10. 10.
    F. J. Solis and R. J.-B. Wets. Minimization by random search techniques. Mathematics of Operations Research, 6(1):19–30, 1981.MathSciNetzbMATHCrossRefGoogle Scholar
  11. 11.
    G. W. Walster, E. R. Hansen, and S. Sengupta. Test results for a global optimization algorithm. In P. T. Boggs, Richard H. Byrd, and Robert B. Schnabel, editors, Numerical Optimization, pages 272–287. SIAM, 1984.Google Scholar
  12. 12.
    Z. B. Zabinsky, D. L. Craesser, M. E. Tuttle, and G. L. Kimm. Global optimization of composite laminates using improving hit-and-run. In C. Floudas and P. Pardalos, editors, Recent Advances in Global Optimization, pages 343–368. Princeton University Press, Princeton, 1992.Google Scholar

Copyright information

© Kluwer Academic Publishers 2001

Authors and Affiliations

  • P. M. Ortigosa
    • 1
  • I. García
    • 1
  • M. Jelásity
    • 2
  1. 1.Department of Computer Architecture and ElectronicsUniversity of AlmeríaSpain
  2. 2.Research Group on Artificial IntelligenceMTA-JATESzegedHungary

Personalised recommendations