An Effective Genetic Algorithm for Self-Inverse S-Boxes

Hua Chen, Dengguo Feng

Abstract: As unique nonlinear components of most block ciphers, S-boxes provide the most important confusion effect. How to design secure S-boxes is a key problem in the design of block ciphers. There have existed many methods to design S-boxes, among which the method with genetic algorithms has attracted much interest in recent years. In this paper, an effective genetic algorithm is provided for the current pop- ular self-inverse S-boxes, which consists of fitness function, breeding function, selection function and mutation function. Under the given algorithm, a large number of good self- inverse S-boxes can be obtained.