S box construction and result analysis based on optimal tabu-genetic algorithm

This paper proposes an s-box optimization method based on tabu-genetic algorithm. In the algorithm niche technique is introduced to maintain population diversity and can avoid premature convergence. Other performances such as s-box avalanche criteria and diffusion properties are treated as evolution targets and s-box optimization is studied in more depth. The experimental result shows that the improved tabu-genetic algorithm based on this paper is feasible and efficient to construct s-box. And the improved algorithm can not only acquire a series of s-boxes of high degree of nonlinearity and low differential evenness degree, but also effectively reduce redundant computations and fasten convergence speed.