Evolving cryptographically sound boolean functions

Abstract

This paper explores the evolution of Boolean functions for a cryptographic usage, with genetic algorithms and genetic programming. We also experiment with a new mutation operator and a new kind of initialization process. Results obtained show that those modifications can help in obtaining better solutions. The results indicate that it is possible to obtain high quality Boolean functions with algorithms that are not tailor-made for this purpose. Additionally, among the algorithms tested, the best performance was obtained with variations of genetic programming.