Abstract:Generally a normal distribution random sequence is generated by a single seed, but in fact each random number in this sequence can also be generated by different seeds. The method to generate different seeds is put forward and the measure to avoid generating same sequences when random number generator is transferred continuously is proposed. Four statistical parameters being used, the qualities of a great lot of normal random sequences, which are generated by single seed and multi-seed methods respectively at different time, are contrastively studied in the aspects of normality, mean, deviation and randomicity by virtue of statistical simulation and hypothesis test theory. The results show that random sequences generated by multi-seed method have the evident advantage in normality and deviation quality, while in mean and randomicity quality, the effect of these two methods are basically same to each other, so as a whole, the multi-seed method is better than the single seed one.
赵学智;叶邦彦. 单种子和多种子对正态随机序列质量的影响[J]. , 2011, 30(1): 49-55.
ZHAO Xue-zhi;YE Bang-yan. Influence of Single Seed and Multi-seed on Quality of Normal Distribution Random Sequence. , 2011, 30(1): 49-55.