Computing

, Volume 12, Issue 3, pp 223–246

Computer methods for sampling from gamma, beta, poisson and bionomial distributions

  • J. H. Ahrens
  • U. Dieter
Article

Abstract

Accurate computer methods are evaluated which transform uniformly distributed random numbers into quantities that follow gamma, beta, Poisson, binomial and negative-binomial distributions. All algorithms are designed for variable parameters. The known convenient methods are slow when the parameters are large. Therefore new procedures are introduced which can cope efficiently with parameters of all sizes. Some algorithms require sampling from the normal distribution as an intermediate step. In the reported computer experiments the normal deviates were obtained from a recent method which is also described.

Key words and phrases

Random numbers pseudorandom normal distribution gamma distribution bei distribution Poisson distribution binomial distribution simulation numerical analysis 

Computermethoden zur Erzeugung Gamma-, Beta-, Poisson- und Binomialverteilter Zufallszahlen

Zusammenfassung

Zur Erzeugung nicht-gleichverteilter Zufallszahlen braucht man Methoden, die gleichverteilte Zufallszahlen in Größen der gegebenen Verteilung transformieren. Es werden Transformationen untersucht, die Gamma-, Beta-, Poisson- oder Binomial-verteilte Zufallszahlen produzieren. Approximative Verfahren werden nicht behandelt. Die bisher bekannten Algorithmen sind langsam, wenn die Parameter der Verteilungen groß sind. Daher werden neue Methoden eingeführt, die diesen Nachteil weitgehend vermeiden. In allen Verfahren dürfen die Parameter beliebig und jedesmal neu gewählt werden. Für manche Transformationen werden normalverteilte Zufallszahlen als Zwischenschritt benötigt; die hierfür verwendete Methode ist ebenfalls angegeben.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    Ahrens, J. H., and U. Dieter: Computer methods for sampling from the exponential and normal distributions. Comm. ACM15, 873–882 (1972).CrossRefMathSciNetGoogle Scholar
  2. [2]
    Ahrens, J. H., and U. Dieter: Extensions of Forsythe's method for random sampling from the normal distribution. Math. Comput.27, 927–937 (1973).MathSciNetGoogle Scholar
  3. [3]
    Ahrens, J. H., U. Dieter, and A. Grube: Pseudo-random numbers: A new proposal for the choice of multiplicators. Computing6, 121–138 (1970).CrossRefMathSciNetGoogle Scholar
  4. [4]
    Bankövi, G.: A note on the generation of beta-distributed and gamma-distributed random variables. Publ. Math. Inst. Hung. Acad. Scie.9, 555–563 (1964).MATHGoogle Scholar
  5. [5]
    Békéssy, A.: Remarks on beta-distributed random numbers. Publ. Math. Inst. Hung. Acad. Scie.9, 565–571 (1964).MATHGoogle Scholar
  6. [6]
    Dieter, U.: Autokorrelation multiplikativ-erzeugter Pseudo-Zufallszahlen. Operations Res. Verfahren6, 69–85 (1969).MATHGoogle Scholar
  7. [7]
    Dieter, U.: Pseudo-random numbers: The exact distribution of pairs. Math. Comput.25, 855–883 (1971).MATHMathSciNetGoogle Scholar
  8. [8]
    Dieter, U.: Pseudo-random numbers: Permutations of triplets. J. Res. Nat. Bureau Standards (to appear).Google Scholar
  9. [9]
    Dieter, U., and J. H. Ahrens: An exact determination of serial correlations of pseudo-random numbers. Numer. Math.17, 101–123 (1971).CrossRefMathSciNetGoogle Scholar
  10. [10]
    Dieter, U., and J. H. Ahrens: A combinational method for the generation of normally distributed random numbers. Computing11, 137–146 (1973).CrossRefMathSciNetGoogle Scholar
  11. [11]
    Dieter, U., and J. H. Ahrens: Pseudo-random numbers. Preliminary version in preprint, 1972. (340 pages).Google Scholar
  12. [12]
    Fisz, M.: Probability Theory and Mathematical Statistics. New York: J. Wiley. 1962.Google Scholar
  13. [13]
    Forsythe, G. E.: Von Neumann's comparison method for random sampling from the normal and other distributions. Math. Comput.26, 817–826 (1972).MATHMathSciNetGoogle Scholar
  14. [14]
    Jöhnk, M. D.: Erzeugen von betaverteilten und gammaverteilten Zufallszahlen. Metrika8, 5–15 (1964).MATHMathSciNetGoogle Scholar
  15. [15]
    Jöhnk, M. D.: Erzeugen und Tasten von Zufallszahlen. (Berichte aus dem Institut für Statistik und aus dem Institut für Angewandte Statistik der FU Berlin, Heft 6.) Würzburg: Physica-Verlag. 1969.Google Scholar
  16. [16]
    Knuth, D. E.: The art of computer programming, Vol. II: Seminumerical algorithms. Reading, Mass.: Addison Wesley. 1969.Google Scholar
  17. [17]
    Léger, R.: On sampling from the negative binomial and Weibull distributions. Master's thesis, Dalhousie University, Halifax, N.S., 1973.Google Scholar
  18. [18]
    MacLaren, M. D., G. Marsaglia, and T. A. Bray: A fast procedure for generating normal random variables. Comm. ACM7, 4–10 (1964).CrossRefGoogle Scholar
  19. [19]
    Marsaglia, G.: Expressing a random variable in terms of uniform random variables. Ann. Math. Stat.32, 894–899 (1961).MATHMathSciNetGoogle Scholar
  20. [20]
    Marsaglia, G.: Random variables and computers. Trans. Third Prague Conf. Information Theory, Statistics and Decision Functions. Prague 1964, 499–512.Google Scholar
  21. [21]
    Payne, W. H., and T. G. Lewis: Conditional bit sampling: Accuracy and speed. Math. Software (Rice, J. R., ed.). Academic Press. 1971.Google Scholar
  22. [22]
    Relles, D. A.: A simple algorithm for generating binomial random variables whenN is large. J. Am. Stat. Ass.67, 612–613 (1972).MATHGoogle Scholar
  23. [23]
    von Neumann, J.: Various techniques in connection with random digits. Monte Carlo Methods. Nat. Bureau Standards, AMS12, 36–38 (1951).Google Scholar

Copyright information

© Springer-Verlag 1974

Authors and Affiliations

  • J. H. Ahrens
    • 1
  • U. Dieter
    • 2
  1. 1.Department of Applied MathematicsNova Scotia Technical CollegeHalifaxCanada
  2. 2.Lehrkanzel und Institut für Mathematische StatistikGrazAustria

Personalised recommendations