DocumentCode :
2896985
Title :
On improving pseudo-random number generators
Author :
Deng, Lih-Yuan ; George, E. Olusegun ; Chu, Yu-Chm
Author_Institution :
Dept. of Math. Sci., Memphis State Univ., TN, USA
fYear :
1991
fDate :
8-11 Dec 1991
Firstpage :
1035
Lastpage :
1042
Abstract :
The authors provide some theoretical justifications for the asymptotic uniformity and asymptotic independence of the combination generators without the usual assumption of independence of the generators. The theorems presented give a general method of transforming a possibly bad generator into a much better generator which will yield an asymptotically independent and uniformly distributed random sequence. An empirical study demonstrates that only a small number of terms is required in the author´s asymptotic theory to achieve a much more uniform random sequence. As pointed out by S.K. Park and K.W. Miller (1988), some generators provided by certain computer systems may not be very random. Combining several generators into a single generator will provide some protection against the possibility of bad system-provided generators
Keywords :
random number generation; statistical analysis; asymptotic independence; asymptotic uniformity; combination generators; pseudo-random number generators; statistical analysis; Character generation; Computer displays; Convergence; Multidimensional systems; Random number generation; Random variables;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Simulation Conference, 1991. Proceedings., Winter
Conference_Location :
Phoenix, AZ
Print_ISBN :
0-7803-0181-1
Type :
conf
DOI :
10.1109/WSC.1991.185721
Filename :
185721
Link To Document :
بازگشت