A combining encryption and decryption algorithm based on neural networks, cross-variation and FFT

Diversity and aperiodicity of populations generated by genetic algorithm and high nonlinearity of fast Fourier transformation (FFI) are absorbed to create a new method of encryption and decryption of important data based on neural networks, cross-variation and FFT algorithm. Through the cross-variation operation of key sequence, the model improves nonlinear relationship between key and initial random seeds; transformation of part of ciphertext avoids the ciphertext-only attack efficiently. The whole model enhances periodicity of key sequence and complexity of the algorithm. Because of realizing the genuine "one-time pad", it can ensure that the encryption system has a very high security standard.