How to Use Metaheuristics for Design of Symmetric-Key Primitives

Abstract

The ultimate goal of designing a symmetric-key cryptographic primitive often can be formulated as an optimization problem. So far, these problems mainly have been solved with trivial algorithms such as brute force or random search. We show that a more advanced and equally versatile class of search algorithms, called metaheuristics, can help to tackle optimization problems related to design of symmetric-key primitives. We use two nature-inspired metaheuristics, simulated annealing and genetic algorithm, to optimize in terms of security the components of two recent cryptographic designs, SKINNY and AES-round based constructions. The positive outputs of the optimization suggest that metaheuristics are non-trivial tools, well suited for automatic design of primitives.