Automatic generation of hash functions for program code obfuscation

The specifics of hash function applications in code obfuscation are considered, as well as the disadvantages of currently existing hash functions for this purpose. Considering these specifics and disadvantages, an automatic hash function generation method is proposed, that is based on a genetic programming approach. Methods of measuring the hash function resilience to automated preimage attacks based on SMT solvers usage and random collision resistance are also proposed. Generated functions were evaluated and a method of weak function detection is proposed, that allows to increase the resilience of generated functions to attacks notably.