Pseudo-random number generators have always been important in experimental design, computer simulation, cryptography and statistical analysis. This paper presents a method of comparing the degree of independence exhib...Pseudo-random number generators have always been important in experimental design, computer simulation, cryptography and statistical analysis. This paper presents a method of comparing the degree of independence exhibited by various random number generators, a procedure, based on consideration of the largest (in modulus) non-unit eigenvalue of the observed Markov transition matrix, is used to assess the 'randomness' of a random number generator.展开更多
文摘Pseudo-random number generators have always been important in experimental design, computer simulation, cryptography and statistical analysis. This paper presents a method of comparing the degree of independence exhibited by various random number generators, a procedure, based on consideration of the largest (in modulus) non-unit eigenvalue of the observed Markov transition matrix, is used to assess the 'randomness' of a random number generator.