Enhanced AES cryptosystem by using genetic algorithm and neural network in S-box

Cryptography based on block ciphers use Key-dependent ciphers for encryption and decryption. The efficiency of these systems depends on the security and the speed of the algorithm. The encryption process needs to be adaptive and dynamic in order to face any cryptanalytic attacks. Increasing the complexity of the algorithm is one way to prevent the attacks. The introduced complexity increases the execution time of the algorithm which leads to timing attacks. This paper attempts to propose two enhanced AES cryptosystem by employing Genetic algorithm (GA) in SPboxes and modification of AES by implementing nonlinear neural network (NN) in SP network to increase the security against timing attack and reduce the computational time of the proposed system. Both GA and NN are used in key expansion and key distribution of the AES algorithm.