Parallel Framework for Evolutionary Black-box optimization with Application to Algebraic Cryptanalysis

In this study, a framework aimed at implementing pseudo-Boolean optimization algorithms is presented. In this framework, called EvoGuess objective functions are calculated using the Monte Carlo random sampling method. The framework is used to construct algebraic attacks on several cryptographic keystream generators. An original cryptanalysis problem is first reduced to the Boolean satisfiability problem (SAT). The obtained problem can be simplified by assigning guessed bits values from some set of variables. A problem of finding a guessed bits set, which corresponds to a SAT-based attack with relatively low runtime estimation, is studied. This problem is considered as an optimization problem of a special black-box fitness function on the Boolean hypercube. To solve this optimization problem we develop EvoGuess, in which several evolutionary strategies are implemented. Also, the framework allows to use an arbitrary CDCL-based SAT solver. In computational experiments, SAT-based guess-and-determine attacks on several well-known keystream generators are constructed.