Clonal selection algorithm for the cryptanalysis of a simple substitution cipher

Abstract

The development of Clonal Selection Algorithm (CSA) is inspired by elements from clonal selection theories. The main aim of the algorithm is to create a group of memories for antibodies which is used to solve engineering problems. CSA has been used in several applications such as optimization of dynamic functions and scheduling problem. In this paper CSA is used to cryptanalyse a simple substitution cipher, whereby the key used to encrypt the plaintext is searched. The algorithm is based on the Clonalg algorithm developed by de Castro and Von Zuben. The performance of CSA and Genetic Algorithm (GA) is studied by comparing the number of correct characters and the progress of fitness values using different population sizes. In the first test, it was found that CSA is capable of detecting more correct characters in short ciphertexts as compared to GA. However, GA demonstrated a better performance in small population size in the second test while CSA showed higher fitness values than GA when the population size is higher. Based on the results, CSA exhibits a good performance and can be further improved in the cryptanalysis of a simple substitution cipher.