A genetic strategy to design cellular automata based block ciphers

We propose an evolutionary computation approach to design a fast and secure block cipher using non-uniform second-order cellular automata. We build a flexible block ciphering model that permit the construction of a huge space of possible instances defined each one by a finite set of elementary transition rules. The constructed space is explored using a genetic algorithms strategy in order to find an optimal solution with respect to the strict avalanche criterion used as fitness measurements. The genetically designed cipher is benchmarked experimentally using conventional statistical tests, and shown to have very admissible characteristics leading to a very acceptable level of cryptographic security. Moreover, performances analysis shows that the designed cipher permit to achieve a high encryption/decryption speed, and compete many of the existing standardized ciphers.