An adaptive Particle Swarm Optimization to Solve Modern Security System

Nonlinear stream ciphers hold a vital position in ensuring confidentiality within contemporary cryptographic systems. In this context, we interduce a new approach known as the Adaptive Three Sub-Swarm PSO based on Regrouping Strategy (ATSPSO-RG) to enhance the effectiveness of deciphering nonlinear stream ciphers, particularly those based on the Bruer generator system. Our method significantly reduces the complexity of cipher-only attacks, making the extraction of the correct secret key more efficient. The proposed system dividing the population size into three distinct subswarm, each incorporating its evolving strategies, namely the single score, four score, and nine score strategies. In the single score strategy, we employed, the traditional PSO evolving process. In the four scores strategy, we used three additional scores in addition to the single score of the traditional PSO to update the velocity and position of particles. lastly, the nine score strategy employed five scores to update the velocity and position, as well as the four scores of the previous score. To counteract the issue of premature convergence, we employ a regrouping strategy that periodically reinitializes the population. Experimental results vividly demonstrate the success of the ATSPSO-RG method in deciphering the Bruer cipher, even when dealing with a varying number of Linear Feedback Shift Registers (LFSRs) and polynomial degrees, spanning up to 15, and secret initial values extending up to 39. The method proves its efficiency in cryptanalysis, offering a potent solution for nonlinear stream ciphers with substantial key sizes while ensuring a reasonable timeframe for analysis. The research outcomes underscore the efficacy of ATSPSO-RG, highlighting both its accuracy and time efficiency in deciphering nonlinear stream ciphers.