Execution examination of chaotic S-box dependent on improved PSO algorithm

Abstract

Achieving proper nonlinear properties and autocorrelation in the S-box structure is an open challenge in cryptography. Besides, there have been numerous articles on the optimization of S-box, using two types of fitness functions for optimization. This study investigated both types of functions and compares their performance. In addition, this study used ergodic chaotic maps. First, the performance of particle swarm optimization (PSO) was improved using these maps. Then, the new chaotic S-boxes were designed based on the ergodic maps. After that, the improved PSO was used for optimization to obtain the best S-boxes. This optimization was performed once by selecting nonlinearity as a fitness function. At the second optimization, the entropy source was selected as a fitness function for optimization by examining the P-value of the mono-test frequency. Finally, the related results for the introduced chaotic S-boxes were compared to the optimized chaotic S-boxes with two types of fitness functions. The introduced S-boxes were safe due to the use of ergodic maps with high keyspace length. Furthermore, the simulation performance was analyzed and compared with other relevant approaches.