Evolutionary computation techniques for constructing SAT-based attacks in algebraic cryptanalysis
Авторы: Pavlenko A., Semenov A., Ulyantsev V.
Журнал: Lecture Notes in Computer Science
Отчётный год: 2019
Аннотация: In this paper we present the results on applying evolutionary computation techniques to construction of several cryptographic attacks. In particular, SAT-based guess-and-determine attacks studied in the context of algebraic cryptanalysis. Each of these attacks is built upon some set of Boolean variables, thus it can be specified by a Boolean vector. We use two general evolutionary strategies to find an optimal vector: (1+1)-EA and GA. Based on these strategies parallel algorithms (based on modern SAT-solvers) for solving the problem of minimization of a special pseudo-Boolean function are implemented. This function is a fitness function used to evaluate the runtime of a guess-and-determine attack. We compare the efficiency of (1+1)-EA and GA with the algorithm from the Tabu search class, that was earlier used to solve related problems. Our GA-based solution showed the best results on a number of test instances, namely, cryptanalysis problems of several stream ciphers (cryptographic keystream generators).
Индексируется WOS: 1
Индексируется Scopus: 1
Индексируется РИНЦ: 1
Публикация в печати: 0
Добавил в систему: