Advertisement

Global Optimization by Parallel Constrained Biased Random Search

  • I. Garcia
  • G. T. Herman
Part of the Nonconvex Optimization and Its Applications book series (NOIA, volume 7)

Abstract

The main purpose of this paper is to demonstrate that even a very minimal cooperation between multiple processors (each executing the same general purpose probabilistic global optimization algorithm) can significantly improve the computational efficiency as compared to executing the algorithm without cooperation. We describe one such cooperative general purpose algorithm for global optimization and its implementation on a parallel computer. The algorithm, called Parallel Constrained Biased Random Search (PCBRS), can be classified as a probabilistic random search method. It needs just one user supplied parameter which is related to the accuracy of the solution. Comparisons to several algorithms using the Dixon-Szegö test functions are presented. PCBRS has been implemented on a multiprocessor system and on a distributed system of workstations following a Multiple Instruction Multiple Data model. Its parallel performance is evaluated using an eight-dimensional pattern classification problem. Our results make apparent that the PCBRS algorithm is computationally efficient for large problems and especially for functions with many local minima. It is shown that the cooperative work of several processors ensures an efficient solution to the global optimization problem.

Keywords

Global Optimization Parallel Algorithms 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    G.L. Bilbro. Fast stochastic global optimization. IEEE Trans, on Systems, Man, and Cybernetics, 24 (4): 684–689, 1994.MathSciNetCrossRefGoogle Scholar
  2. 2.
    P. Chaudhuri. Parallel Algorithms: Design and Analysis. Prentice Hall, 1992.zbMATHGoogle Scholar
  3. 3.
    G.T. Herman and D. Odhner. Visualizing optimization by multiple processors. IEEE Computer Graphics and Applications, 11 (6): 13–15, 1991.CrossRefGoogle Scholar
  4. 4.
    J.G. Trew, A.S. Mills and L.J. Clarke. CHIMP Concepts. Technical Report EPCC-KTP- CHIMP-CONC 1.2, Edinburgh Parallel Computing Centre. The University of Edinburgh, 1992.Google Scholar
  5. 5.
    R.A.A Bruce and J.G. Mills. CHIMP Version 1.0, Interface. Technical Report EPCC-KTP- CHIMP-IFACE 1.5, Edinburgh Parallel Computing Centre. The University of Edinburgh, 1992.Google Scholar
  6. 6.
    L.C.W. Dixon and G.P. Szegö (eds.). Towards Global Optimization 2. North Holland, Amsterdam, 1978.Google Scholar
  7. 7.
    A. Torn and A. Zilinskas. Global Optimization. Lecture Notes in Computer Science 350. Springer-Verlag, Berlin, 1989.Google Scholar
  8. 8.
    H.A. Bremermann. A method of unconstrained globed optimization. Mathematical Biosciences, 9: 1–15, 1970.MathSciNetzbMATHCrossRefGoogle Scholar
  9. 9.
    E. Fagiuoli, P. Pianca and M. Zeccchin. A mixed stochastic-deterministic technique for global optimization. In L.C.W Dixon and G.P. Szego, editor, Towards Global Optimization 2, pages 103–116. North Holland, Amsterdam, 1978.Google Scholar
  10. 10.
    W.L. Price. A controlled random search procedure for global optimization. In L.C.W Dixon and G.P. Szegö, editor, Towards Global Optimization 2, pages 71–84. North Holland, Amsterdam, 1978.Google Scholar
  11. 11.
    F.J. Solis and B. Wets. Minimization by random search techniques. Mathematics of Operations Research, 6 (1): 19–50, 1981.MathSciNetzbMATHCrossRefGoogle Scholar
  12. 12.
    A.A. Torn. A search clustering approach to global optimization. In L.C.W Dixon find G.P. Szego, editor, Towards Global Optimization 2. North Holland, Amsterdam, 1978.Google Scholar
  13. 13.
    S. Parviainen. On methods for global optimization. Licentiate Thesis, Dept. of Math., University of Jyvaskyla, Finland, 1987.Google Scholar
  14. 14.
    J. Gomulka. A user’s experience with Torn’s clustering algorithm. In L.C.W Dixon and G.P. Szego, editor, Towards Global Optimization 2, pages 63–70. North Holland, Amsterdam, 1978.Google Scholar
  15. 15.
    J. Mockus, V Tiesis and A. Zilinskas. The application of Bayesian methods for seeking the extremum. In L.C.W Dixon and G.P. Szego, editor, Towards Global Optimization 2. North Holland, Amsterdam, 1978.Google Scholar
  16. 16.
    A. Zilinskas. On the use of statistical models of multimodal functions for the construction of the optimization algorithms. In A.V. Balikrishnon and M. Thoma, editor, Lecture Notes in Control and Information Sciences, volume 23, pages 138–139. Springer-Verlag, Berlin, 1980.Google Scholar
  17. 17.
    De Biase and F. Frontini. A stochastic method for global optimization: its structure and numerical performance. In L.C.W Dixon and G.P. Szegö, editor, Towards Global Optimization 2, pages 85–102. North Holland, Amsterdam, 1978.Google Scholar
  18. 18.
    J. A. Snyman and L.P. Fatti. A multistart global minimization algorithm with dynamic search trajectories. JOTA, 54: 121–141, 1989.MathSciNetCrossRefGoogle Scholar
  19. 19.
    G. Boender, A. Rinnooy Kan, L. Stougie and G. Tommer. A stochastic method for global optimization. Mathematical Programming, 22: 125–140, 1982.MathSciNetzbMATHCrossRefGoogle Scholar
  20. 20.
    J.R. McDonnell and D. Waagen. Evolving recurrent perceptrons for time-series modeling. IEEE Trans. on Neural Network, 5 (1): 24–38, January 1994.CrossRefGoogle Scholar
  21. 21.
    N. Baba. A new approach for finding the global minimum of error fuction of neural networks. Neural Network, 2: 19–30, 1989.CrossRefGoogle Scholar
  22. 22.
    S.S. Rao. Optimization Theory and Applications. John Wiley and Sons, New York, 1978.Google Scholar
  23. 23.
    D.C. Karnopp. Random search techniques for optimization problems. Automatica, 1: 111–121, 1963.CrossRefGoogle Scholar
  24. 24.
    J. Matyas. Random optimization. Automation and Remote Control, 26: 244–251, 1965.MathSciNetzbMATHGoogle Scholar
  25. 25.
    G.T. Herman and D. Yeung. On piecewise-linear classification. IEEE Trans, on Pattern Analysis and Machine Intelligence, 14: 782–786, 1992.CrossRefGoogle Scholar
  26. 26.
    G.T. Herman, D. Odhner and D. Yeung. Optimization for pattern classification using biased random search techniques. Annals of Operations Research, 43: 419–426, 1993.zbMATHCrossRefGoogle Scholar
  27. 27.
    O.L. Mangasarian, R. Setiono, and W.H. Wolberg. Pattern recognition via linear programming: Theory and applications to medical diagnosis. In T.F. Coleman and Y. Li, editor, Large-Scale Numerical Optimization, pages 22–31. Philadelphia: SIAM, 1990.Google Scholar

Copyright information

© Kluwer Academic Publishers 1996

Authors and Affiliations

  • I. Garcia
    • 1
    • 2
  • G. T. Herman
    • 1
  1. 1.Medical Image Processing Group, Department of RadiologyUniversity of PennsylvaniaPhiladelphiaUSA
  2. 2.University of AlmeriaSpain

Personalised recommendations