Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (24): 112-114.DOI: 10.3778/j.issn.1002-8331.2008.24.033

• 网络、通信、安全 • Previous Articles     Next Articles

Optimum algorithm of S-boxes based on master-slave parallel genetic algorithm

YIN Xin-chun1,2,QIU Liang1   

  1. 1.Institute of Information Engineering,Yangzhou University,Yangzhou,Jiangsu 225009,China
    2.State Key Laboratory for Novel Software Technology,Nanjing University,Nanjing 210093,China
  • Received:2007-10-22 Revised:2008-01-21 Online:2008-08-21 Published:2008-08-21
  • Contact: YIN Xin-chun

基于主从式并行遗传算法的S盒优化算法

殷新春1,2,仇 亮1   

  1. 1.扬州大学 信息工程学院,江苏 扬州 225009
    2.南京大学 计算机软件新技术国家重点实验室,南京 210093
  • 通讯作者: 殷新春

Abstract: The calculation time of fitness function evaluation in the optimization of S-boxes based on genetic algorithm was much great.We utilized master-slave parallel genetic algorithm with MPI environment binding for C to improve the sequential genetic algorithm and it was implemented on the Massive Parallel Processors(MPP) DeepComp 1800(Lenovo).The testing results show that the optimization of 8×8 bijective S-boxes based on the parallel algorithm has good speedup.Meanwhile,it is validated that the master-slave parallel algorithm is suitable to solve the problem which the fitness function evaluation calculation time is much great.

Key words: S-boxes, master-slave, parallel genetic algorithms, Message Passing Interface(MPI)

摘要: 针对传统遗传算法优化高阶S盒中适应度评价值计算量相对大的问题,利用MPI+C环境结合主从式并行遗传算法框架来改进传统串行遗传算法,在MPP处理机——“深腾-1800”(联想)上进行测试,实验结果表明,运用此并行算法优化6阶、8阶的双射S盒,只在优化8阶的情况下能得到更好的加速比,从而验证了主从模式在解决适应度评价值计算量大的问题时,可达到接近线性的加速比。

关键词: S盒, 主从模式, 并行遗传算法, 消息传递接口