An empirical study on effect of variations in the population size and generations of genetic algorithms in cryptography

Abstract

Genetic algorithm (GA) belongs to a large class of evolutionary algorithms. It is a metaheuristic, multi-dimensional bio-inspired optimization method which follows the process of natural selection. Genetic algorithms provide optimization solution to combinatorial optimization problems. The solution depends on the probability of GA operators, number of generations and the size of the population. Genetic algorithm can be applied for cryptography, where the conventional symmetric algorithm is enhanced with GA. This work projects, the changes in generations and size of population affects the efficiency of encryption time, decryption time, throughput time and the memory utilized by the genetic algorithm. The study reveals that the efficiency of the cryptographic algorithm treated with GA is dependent on the variations in the number of generations and initial population size. The result shows that an optimum population size has less encryption and decryption time. Among the sample population size taken for the experiment, almost the average population size has minimum encryption and decryption time. Results from iteration variations shows that the average number of iterations has less encryption and decryption time.