Double-Waylandアルゴリズムを用いた乱数の評価方法の考案

概要

論文の詳細を見る
Random numbers consist of true random numbers and pseudorandom numbers. The true random numbers are generated by using physical phenomena, and the pseudorandom numbers are generated by using computers. The pseudorandom numbers have been widely used in many fields because generating these numbers by using a computer is easy. The pseudorandom numbers should possess irregularity and equiprobability. This irregularity of the random numbers should be verified adequately because it affects the simulation accuracy and robustness of a code language. Tests of irregularity are designed by combining different hypothesis tests; such a combination allows the addition of irregularity to a null hypothesis. Therefore, technically, these tests do not perform statistical verification of the irregularity. We suggest evaluation methods for verifying the irregularity by using new amount of characteristics obtained from an attractor reconstructed in higher-dimensional space from the pseudorandom numbers. In this study, we apply the Wayland algorithm, which is a form of non-linear analyses, and the Double-Wayland algorithm devised by the authors for pseudo-uniform random numbers and normal random numbers in order to verify the irregularity. Both the pseudo-uniform random numbers and normal random numbers are evaluated to be deterministic (regular) pseudorandom numbers. The methods developed in this study can be used to evaluate the exact irregularity, unlike the previous evaluation methods.
岐阜医療科学大学の論文

Double-Waylandアルゴリズムを用いた乱数の評価方法の考案

スポンサーリンク

概要

著者

関連論文

スポンサーリンク