Advertisement

Numeric-stability by all-integer simplexiterations

  • Walter Schneider
II Mathematical Programming II.3 Linear Programming And Complementarity
  • 92 Downloads
Part of the Lecture Notes in Control and Information Sciences book series (LNCIS, volume 180)

Abstract

It can be shown, that each inverse of an LP-problem owns one common integer denominator, which results without any additional computation. Representing the inverse by its integer numerators and the common denominator impedes numeric failures.

Nevertheless this procedure seems to be at fault, as denominator and numerator sometimes exceed and overflow occurs. A simple numeric trick hindering this will be presented. Last not least numeric aspects, as well as comparisons with software using LU-triangulisation, round up the article.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Literature

  1. (1).
    Aitkinson, Kendall: Elementary Numerical Analysis; John Wiley & Sons, New York (1985).Google Scholar
  2. (2).
    Borse: Fortran 77 and Numerical Methods for Engineers; Prindle, Weber & Schmidt, Boston (1985)zbMATHGoogle Scholar
  3. (3).
    Dew P.M., James K.R.: Introduction to Numerical Computation in Pascal; Springer-Verlag, Berlin, Heidelberg, N.Y. (1983)Google Scholar
  4. (4).
    Edwards: Calculus and the Personal Computer; Prentice-Hall Inc., Englewood Cliffs, N.Y. (1986)Google Scholar
  5. (5).
    Fröberg, C.E.: Numerical Mathematics — Theory and Computer Applications; The Benjamin-Cummings Publishing Comp., Menlo Park California, Amsterdam, Sydney, Tokyo (1985)zbMATHGoogle Scholar
  6. (6).
    Hager W.W.: Applied Numerical Linear Algebra; Prentice Hall International Inc., London, Sydney, Tokyo (1988)zbMATHGoogle Scholar
  7. (7).
    James, Smith, Wolford: Applied Numerical Methods f.Digital Computation,3rd ed.;Harper & Row Publishers Inc., N.Y., (1985)Google Scholar
  8. (8).
    King, Thomas: Introduction to Numerical Computation; Mc Graw-Hill Book Company, N.Y. (1984)Google Scholar
  9. (9).
    Kistner K.P.: Optimierungsmethoden; Physika-Verlag, Heidelberg (1988)zbMATHGoogle Scholar
  10. (10).
    Knuth D.E.: The Art of Computer Programming; Addison-Wesley Publ.Company Inc., Reading Mass. (1981)zbMATHGoogle Scholar
  11. (11).
    Mathews: Numerical Methods for Computer Science, Engineering & Mathematics;Prentice-Hall Inc.,Englewood Cliffs,N.Y. (1987)Google Scholar
  12. (12).
    Miller, Webb: The engineering of Numerical Software; Prentice-Hall Inc., Englewood Cliffs, N.Y. (1984)Google Scholar
  13. (13).
    Rice: Numerical Methods, Software and Analysis, IMSL Reference Edition; Mc Graw-Hill Book Company, N.Y. (1983)Google Scholar
  14. (14).
    Schneider: Ein Verfahren zur Rechen-und Speicherplatzeinsparung beim Simplexalgorithmus, OR-Proceedings 1984, pp440–446; Springer Verlag, Berlin-Heidelberg-N.Y. (1984)Google Scholar
  15. (15).
    Shoup, Terry: Numerical Methods for the Personal Computer; Prentice-Hall, Inc., Englewood Cliffs, N.Y. (1983)Google Scholar

Copyright information

© International Federation for Information Processing 1992

Authors and Affiliations

  • Walter Schneider
    • 1
  1. 1.Institut f.InformatikUniversität LinzAustria

Personalised recommendations