Constructin Of S-Boxes Based On Intelligent Algorithms

S-boxes are the only nonlinear components in many block cipher. Their cryptographic properties have determined the security of the whole cipher algorithms. It is quite complicated and difficult to construct S-boxes with good characteristics by traditional mathematic methods. Intelligent algorithms have shown strong vitality in the field of NP-C problems in recent years. Three algorithms based on intelligent algorithms is proposed to construct S-boxes, and three principal achievements have been obtained in this thesis.Firstly, S-boxes are optimized by genetic algorithm. The guarantee of orthogonality is become effective by coding type. Crossover and mutation probabilities are adaptive in iterative process. Based on this construction method, the flow chart of constructing S-boxes is given and the group of S-boxes with high nonlinearity and low difference uniformity are generated.Secondly, S-boxes are optimized by immune algorithm. To analyze WHT transformation and autocorrelation of Boolean functions, two rules to optimize Boolean functions are deduced, and then extraction vaccine based on these two rules. Based on this construction method, the flow chart of constructing S-boxes is given and the group of S-boxes with high nonlinearity and low autocorrelation uniformity are generated.Finally, S-boxes are optimized by, which combines the mechanism of feedback organically with the thinking of evolution, accordingly develops enough advantages of the two algorithms. Based on this construction method, the flow chart of constructing S-boxes is given and the group of S-boxes with high nonlinearity and low difference uniformity are generated. The simulation results show that this kind of construction method is better at reducing a large redundancy calculation and quickening the speed of convergence, compared with the one based on genetic algorithm.