A novel cuckoo search strategy for automated cryptanalysis: a case study on the reduced complex knapsack cryptosystem

Abstract

During the past decade several new variants of knapsack cryptosystems have been reported in the literature. Hence, there is a growing demand for automated cryptanalysis of knapsack cryptosystems. Brute force approach is capable to cryptanalyze simple stages of cryptosystems while cryptanalysis of complex cryptosystems demands efficient methods and high-speed computing systems. In the literature, several search heuristics have proven to be promising and effective in automated cryptanalysis (or attacks) of classical or reduced cryptosystems. This paper presents the automated cryptanalysis of the reduced multiplicative knapsack cryptosystem using three different search heuristics, namely, cuckoo search, particle swarm optimization and genetic algorithm. It should be noted that the considered cryptosystem is reduced but is complex and practical representative. To the best of our knowledge, this is the first time when the cuckoo search is utilized for automated cryptanalysis of the complex cryptosystem. The performance of developed techniques has been measured in terms of time taken by the algorithm (i.e., how efficient the algorithm is?), number of times the original plaintext is determined (i.e., success rate), and the number of candidate plaintexts is examined before determining the original plaintext (i.e., how effective the algorithm is?). For the case considered, performance of the proposed techniques, namely, novel binary cuckoo search (NBCS), improved genotype–phenotype binary particle swarm optimization (IGPBPSO), and new genetic algorithm (NGA) is as follows: roughly the NBCS technique is 12% and 8% more efficient, 6% and 5% more successful, and 16% and 12% more effective than IGPBPSO and NGA, respectively. This results show that the proposed NBCS strategy is superior to IGPBPSO and NGA, and therefore NBCS strategy can be used as an efficient and effective choice for solving similar binary discrete problems such as 0–1 knapsack problem, set covering problem, etc.