一种基于遗传算法的S盒优化设计  

An Optimal Design of S-box Based on Genetic Algorithm

作　　者：李亚鹏[1,2] 丁文霞[1] 

机构地区：[1]国防科学技术大学电子科学与工程学院,长沙410073 [2]武警医学院临床医学系,天津300162

出　　处：《重庆理工大学学报（自然科学）》2012年第2期79-85,共7页Journal of Chongqing University of Technology：Natural Science

摘　　要：针对基本遗传算法"爬山"能力弱,有"早熟"特征,寻优效率比较低的不足,提出一种基于遗传算法的S盒的优化设计。在初始种群的生成过程中加入由先验知识产生的部分性能较优的S盒,在一定程度上提高收敛速度和收敛效果;在遗传算子操作中采用最优个体保存法选择策略,可以大大减少额外的计算量;采用Davis顺序交叉法进行交叉操作,引入进化逆转变异法进行变异操作,补偿群体中多样性易损失的不足,同时能够提高算法的搜索效率,加快收敛速度。仿真实验结果表明,本算法构造的S盒在密码学性能、收敛速度和适应度值方面都有很好的改善。In view of the weak ability of ＂Climbing＂ of the basic genetic algorithm, the ＂premature＂ feature, and low searching efficiency, this paper put forward an optimal design of S-box based on ge- netic algorithm. In the initial population of the production process, by adding a priori knowledge gen- erated S-box with partial advantageous performance, the speed and effect of convergence were im- proved to some degree. In genetic operator, using the best individual preservation method selection strategy, additional computing can be greatly reduced. Through the simulation experiments and results analysis, this algorithm is verified by the constructed S-box in cryptography properties. Convergence speed and fitness values have a very good improvement.

关 键 词：基本遗传算法 选择算子 交叉算子 变异算子 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]