A comparison of the correlational behavior of random number generators for the IBM 360
- 1 September 1968
- journal article
- Published by Association for Computing Machinery (ACM) in Communications of the ACM
- Vol. 11 (9) , 641-644
- https://doi.org/10.1145/364063.364095
Abstract
Hutchinson states that the “new” (prime modulo) multiplicative congruential pseudorandom generator, attributed to D.H. Lehmer, has passed the usual statistical tests for random number generators. It is here empirically shown that generators of this type can produce sequences whose autocorrelation functions up to lag 50 exhibit evidence of nonrandomness for many multiplicative constants. An alternative generator proposed by Tausworthe, which uses irreducible polynomials over the field of characteristic two, is shown to be free from this defect. The applicability of these two generators to the IBM 360 is then discussed. Since computer word size can affect a generator's statistical behavior, the older mixed and simple congruential generators, although extensively tested on computers having 36 or more bits per word, may not be optimum generators for the IBM 360.Keywords
This publication has 13 references indexed in Scilit:
- Testing a random number generatorCommunications of the ACM, 1967
- Random-number generation on digital computersIEEE Spectrum, 1967
- Fourier Analysis of Uniform Random Number GeneratorsJournal of the ACM, 1967
- A new uniform pseudorandom number generatorCommunications of the ACM, 1966
- Random numbers generated by linear recurrence modulo twoMathematics of Computation, 1965
- Uniform Random Number GeneratorsJournal of the ACM, 1965
- Mixed Congruential Random Number Generators for Binary MachinesJournal of the ACM, 1964
- The behavior of pseudo-random sequences generated on computers by the multiplicative congruential methodMathematics of Computation, 1962
- An a priori determination of serial correlation in computer generated random numbersMathematics of Computation, 1961
- Distribution of the Serial Correlation CoefficientThe Annals of Mathematical Statistics, 1942