暗号のための乱数に関する統計的検定について
Statistical Testings for Randomness in Cryptograpy

酒折文武

(Fumitake SAKAORI）
2002年度

理工学研究科 数学専攻 博士後期課程

Graduate Student, Graduate School of Science and Engineering
※現在　立教大学助手

　　　Random numbers, or pseudo-random numbers, have much to contribute to the cryptography, e.g. in stream cipher and generation of keys in symmetric-key cryptography. NIST Special Publication 800-22 (SP800-22) and DIEHARD propose many statistical testing methods of the random or pseudo-random number, which test equally probability and 'randomness' from various point of view.
　　　These statistics, however, may not be rigorous in tatistical and mathematical point of view. Some of them are hardly adequate in mathematical corroboration. They sometime use the central limit theorem without using any proofs, which are absolutely nontrivial, or use constants which are nothing more than the result of Monte Carlo simulations in the testing statistics.
　　　The purpose of this study is to give statistical and mathematical endorsement to these statistics and modify them.