Abstract: |
This paper discusses several evolutionary computation algorithms for improving the nonlinearity of Boolean functions, discrete combinatorial objects with applications in coding theory and cryptology. Firstly, we review the existing literature focusing on relevant properties and heuristic design of single output Boolean functions. We compare several local search algorithms with different problem representations and evolutionary operators. The effects of these choices on the dynamics of EC algorithms is studied in light of some recently developed theory, Finally a new hybrid EC search algorithm is proposed to fully take advantage of these ideas. |