6轮数据加密标准的演化分析  

Cryptanalysis of Six-Round DES Using Evolutionary Algorithm

作　　者：宋军[1,2] 张焕国[1] 王丽娜[1] 

机构地区：[1]武汉大学计算机学院,湖北武汉430072 [2]中国地质大学计算机学院,湖北武汉430074

出　　处：《武汉大学学报（理学版）》2009年第1期71-74,共4页Journal of Wuhan University:Natural Science Edition

基　　金：国家高技术研究发展计划(863)项目(2006AA012442,2007AA012411);国家自然科学基金资助项目(60473023,60673071)

摘　　要：采用线性逼近的思想,提出了一种基于Hamming距离的遗传算法适应度函数模型.在此基础上,利用遗传算法对1~6轮数据加密算法的不同模式长度的部分密钥进行了定量的对比分析实验,对6轮数据加密算法进行演化分析可获得21 bit的未知密钥.结果显示：适应度函数设计与密码分析对象的安全强度是影响演化密码分析效果的关键因素;由于演化技术是有指导的随机性或概率性技术,在一定数量的实验中,运行时间和成功率存在较明显差异;在一定程度上增大种群规模,明密文数量或演化代数能够取得更好的分析效果.Based on the idea of linear approximation,this paper designs a genetic algorithm fitness function model of Hamming distance,in which we have the quantitative experiments of the genetic algorithm analysis on part of the DES keys in the 1st to the 6th round under different fixed bits lengths.The search of experimental result can acquire 21 bits unknown key of 6-round DES.Furthermore,the experimental results show that the designing of fitness function and the security of cryptanalysis＇ object are important factor affecting the effect of the evolutionary cryptanalysis.Because of the intrinsic randomicity or probability for the evolutionary technology,there exist different for the running time and different success rates of acquiring results in plenty of experiments.Further test indicate that the better results can be acquired by increasing population size,the running generation or the amount of the plaintext-ciphertext pairs to some extent.

关 键 词：密码分析 演化计算 演化分析 线性逼近 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]