Evolutionary Design Of S-boxes

S-Box is the core component of the block cipher algorithm, so the S-Box determines the strength of the strength of encryption algorithm. S-Box designing criteria are as follows: differential, nonlinearity autocorrelation properties, give some algorithm that calculating the properties very quickly. This is not just a few criteria to measure the strength of the box, but also the evolution of S-box of an important tool. In the evolution process, the variance of the Walsh spectrum of a Boolean function, called Walsh variance, is more effective than the traditional non-linearity property.The S-boxes of the AES, Camellia encryption algorithm, have the linear redundancy. For randomly generated S-boxes, a few little S-Boxes have the affine equivalence structure, so that is a weakness of the AES algorithm. The original Fuller-Millan algorithm which is used for detecting affine equivalence of Boolean functions is very slow in some case. Based on the algorithm, cutting some branch on search tree, greatly improve the efficiency of the algorithm.In the latter part describes how the genetic algorithm will be applied to the S-box design, detail the framework of the evolutionary system, how to choose a different operator, the operator for the evolution of the impact of the results. Finally, given a number of the evolution results, for the random S-boxes, evolutionary method, the non-linearity and autocorrelation can be significantly improved, to prove the effectiveness of the algorithm.