Use of a new approach to automated break transposition cipher system

Abstract

Abstract Particle swarm optimization (PSO) based cryptanalysis has acquired much attention because it has a fast convergence rate. This paper investigates the use of a new approach which is PSO depending on the clustering algorithm (CLPSO) to break the transposition cipher system in appropriate time. The dynamic of CLPSO is different from existing PSO in population size, topology and the ways to find the best solution. CLPSO start’s with using high population size (particles) and then applying preprocessing operation in order to reduce this population and then partition this population into several clusters based on using semi CLARANS algorithm and allow particles to share information in different clusters and give the particles ability to move from one cluster to another depending on the values of fitness function. Experimental results of the proposed CLPSO were very promising and the results proved that the CLPSO algorithm reduces the number of tries that needed to attack long key transposition cipher in almost real time using ciphertext-only attack. This new method allows recovering key length up to 35 with no more than 58.3 second as maximum consuming time. In this case study, different parameters such as: population sizes (100-5000), key size (10 - 35) and length of ciphertext (250 - 5000) were used.