A new S-box three-layer optimization method and its application

As a common nonlinear component in block cipher, S-box plays a role of confusion and determines the security strength of the whole block cipher algorithm. How to generate a high-performance S-box has always been a hot research issue. Based on four new proposed theorems, a new proposed chaotic map and artificial jellyfish optimization algorithm, a three-layer optimization method for generating high-performance S-box is proposed in this paper. In the first layer, a S-box with high nonlinearity satisfying bijection is generated by combining the new chaotic map. In the second layer, the differential uniformity of S-box is optimized without reducing the nonlinearity of S-box. In the third layer, based on the artificial jellyfish optimization algorithm, a multi-objective optimization algorithm with nonlinearity and differential uniformity as constraints is proposed to optimize other criteria of the S-box. Finally, a block cipher algorithm is designed by using high-performance S-box, chaotic map, improved generalized Feistel structure, etc. Many tests are carried out for the algorithm including key sensitivity analysis, key space size analysis, confusion and diffusion test and ciphertext statistical analysis, and good results are obtained.