A new approach of known plaintext attack with Genetic Algorithm

摘要

Cryptanalysis of modern cryptosystems is viewed as NP-Hard problem. Block ciphers, a modern symmetric key cipher are characterised with the nonlinearity and low autocorrelation of their structure. In literature, various attacks were accomplished based on traditional research algorithms such the brute force, but results still insufficient especially with wide instances due to resources requirement, which increase with the size of the problem. Actual research tends toward the use of bio-inspired intelligence algorithms, which are heuristic methods able to handle various combinatorial problems due to their optimisation of search space and fast convergence with reasonable resource consumption. The paper presents a new approach based on genetic algorithm for cryptanalysis of block ciphers; we focuses especially around the problem formulation, which seems a critical factor that depends the attack success. The experiments were accomplished on various set of data; the obtained results indicate that the proposed methodology seems an efficient tool in handling such attacks. Moreover, results comparisons of the considered approach with similar heuristics such Particle Swarm Optimisation and Brute Force reports its effectiveness in solving the considered problem.