Evolutionary computation in cryptanalysis of classical ciphers

Abstract

Evolutionary computation have represented a very popular way of
problem-solving in the recent years. This approach is also capable of effectively solving historical cipher in a fully automated way. This paper deals with empirical cryptanalysis of a monoalphabetic substitution using a genetic algorithm (GA) and a parallel genetic algorithm (PGA). The key ingredient of our contribution
is the parameter analysis of GA and PGA. We focus on how these parameters affect the success rate of solving the monoalphabetic substitution.