A Generic Genetic Algorithm to Automate an Attack on Classical Ciphers

Abstract

The work presented in this paper describes a generic genetic algorithm called DUREHA’s (Dominance, Universal stochastic sampling and Rank-based Emulation of a Heuristic Algorithm) Algorithm for cryptanalysis of classical ciphers. The underlying objective of this paper is to automate the process of cryptanalysis in order to render salvage of time, and resources available, preserve population diversity, minimize the convergence rate and control mutation rates. While numerous algorithms have been proposed to automate this process for variegated ciphers, these approaches are yet isolated from each other. The existence of a generic algorithm to cryptanalyze any type of cipher is yet not true. The algorithm proposed in this paper aspires to address such issues. The implementation and experimentation of the proposed algorithm is accomplished using three types of classical ciphers namely monosubstitution, poly-substitution and columnar transposition. The theoretical validation and experimental results indicate that the proposed algorithm is able to decrypt the ciphers by reclaiming80.71% ,87.31%and 77.66% of letters in correct position in Mono-substitution, Columnar Transposition and Vignere cipher respectively. It is also able to distinguish between the three types of ciphers correctly and is able to correctly control the mutation andconvergence rates and preserve population diversity.