Upper bounds in spectral test for multiple recursive random number generators with missing terms

Abstract One method for generating random numbers (RNs) of long period recommended by many scholars is the multiple recursive generator (MRG), in that the current RN is essentially a linear combination of the k preceding ones. In this paper, the upper bounds for a figure of merit adopted in the spectral test are derived for the k th order MRG with p ≤ k terms. As p gets smaller, the bounds become smaller as well. The simplest form of the k th order MRG with two terms frequently discussed in literature is found to have the worst bound.

[1]  L. R. Moore,et al.  An Exhaustive Analysis of Multiplicative Congruential Random Number Generators with Modulus $2^{31} - 1$ , 1986 .

[2]  Pierre L'Ecuyer,et al.  A search for good multiple recursive random number generators , 1993, TOMC.

[3]  U. Dieter,et al.  How to calculate shortest vectors in a lattice , 1975 .

[4]  Chiang Kao,et al.  Several extensively tested random number generators , 1994, Comput. Oper. Res..

[5]  Chiang Kao,et al.  An exhaustive analysis of prime modulus multiplicative congruential random number generators with modulus smaller than 215 , 1996 .

[6]  C. A. Rogers,et al.  An Introduction to the Geometry of Numbers , 1959 .

[7]  Donald E. Knuth,et al.  The art of computer programming. Vol.2: Seminumerical algorithms , 1981 .

[8]  Chiang Kao,et al.  Symmetry property of multiplicative congruential random number generator in chi-square test , 1995, Int. J. Comput. Math..

[9]  G. Marsaglia.  Random numbers fall mainly in the planes. , 1968, Proceedings of the National Academy of Sciences of the United States of America.

[10]  Harald Niederreiter,et al.  Random number generation and Quasi-Monte Carlo methods , 1992, CBMS-NSF regional conference series in applied mathematics.

[11]  S. K. Park,et al.  Random number generators: good ones are hard to find , 1988, CACM.

[12]  G. S. Fishman.  Multiplicative congruential random number generators with modulus 2^{}: an exhaustive analysis for =32 and a partial analysis for =48 , 1990 .

[13]  R. R. Coveyou,et al.  Fourier Analysis of Uniform Random Number Generators , 1967, JACM.