The hierarchy of correlations in random binary sequences
- 83 Downloads
The meaning of randomness is studied for the simple case of binary sequences. Ensemble theory is used, together with correlation coefficients of any order. Conservation laws for the total amount of correlation are obtained. They imply that true randomness is an ensemble property and can never be achieved in a single sequence. The relation with entropy is discussed for different ensembles. Well-tempered pseudorandom sequences turn out to be suitable sources of random numbers, and practical recipes to generate them for use in large-scale Monte Carlo simulations are found.
Key wordsRandomness correlation coefficients entropy random-number generation Monte Carlo calculations
Unable to display preview. Download preview PDF.
- 1.D. E. Knuth,The Art of Computer Programming, Vol. 2 (Addison-Wesley, 1981).Google Scholar
- 2.L. Afflerbach,J. Comp. Appl. Math. 31:3 (1990).Google Scholar
- 3.G. Marsaglia and A. Zaman, preprint (1990).Google Scholar
- 4.F. James,Comp. Phys. Commun. 60:329 (1990).Google Scholar
- 5.B. D. Ripley,J. Comp. Appl. Math. 31:153 (1990).Google Scholar
- 6.H. Niederreiter,Ann. Oper. Res., to be published.Google Scholar
- 7.M. van Lambalgen,J. Symbolic Logic 52:725 (1987); Ph. D. Thesis, University of Amsterdam (1987).Google Scholar
- 8.A. Compagner and A. Hoogland,J. Comp. Phys. 71:391 (1987).Google Scholar
- 9.A. Compagner,Am. J. Phys., to be published (1991).Google Scholar
- 10.S. W. Golomb,Shift Register Sequences (Holden-Day, 1967).Google Scholar
- 11.N. Zierler,Inform. Control 15:67 (1969).Google Scholar