Abstract

In this thesis, we investigate the applications of two swarm-inspired artificial life optimization techniques in cryptology. In particular, we investigate the use of both Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO) for automated cryptanalysis of simple classical substitution ciphers. We also use PSO to construct Boolean functions with some desirable cryptographic properties. Both ACO and PSO based attacks proved to be effective for the cryptanalysis of simple substitution ciphers encoded with various sets of encoding keys. Purely uni-gram and bi-gram statistics are used for solving this problem. Boolean functions are vital components of symmetric-key ciphers such as block ciphers, stream ciphers and hash functions. When used in cipher systems, Boolean functions should satisfy several cryptographic properties such as balance, high nonlinearity, resiliency and high algebraic degree. Using PSO, with an unorthodox approach of spectral inversion, we are able to construct Boolean functions that achieve the maximum possible nonlinearity (Bent function) and several other important resilient functions. In fact, we were able to construct, for the first time, a 9-variable Boolean function with nonlinearity 240, algebraic degree 5, and resiliency degree 3. This construction affirmatively answers the open problem about the existence of such functions