Evaluating the Probability of Successful Knapsack Ciphersystem Analysis with Genetic Algorithms

The genetic algorithms are a well-known family of high-performance probabilistic algorithms. In this paper, we explore the possibility of using the genetic algorithm for the Knapsack problem to compromise the security of a Knapsack cipher. Despite being much faster than the exact algorithms, the genetic algorithm for the Knapsack problem may fail to find a solution. We explore the connection between the success rate of the genetic algorithm and the Knapsack problem parameters: the Knapsack Density, the items count in the solution and whether the Knapsack is modular or multiplicative. As a result, we determine whether the genetic algorithm is viable as an analysis tool for the Knapsack ciphers with specific parameters.