REAL-TIME ADAPTIVE AES KEY SCHEDULING USING ONLINE GENETIC ALGORITHM

Abstract

The proposed research seeks to offer an advanced AES key scheduling algorithm executed through the internet using the Online Genetic Algorithm (OGA), utilizing exploration strategies and artificial intelligence in dynamically pursuing approaches towards approximate key solutions. The research evaluates various cryptographic factors such as entropy, balanced bit-keys, transfer probability, and avalanche properties in assessing different cryptographic keys and hence provides adaptive intelligence in the choice of keys. The key generation process based on Monte Carlo simulation (20 trials over 100 generations) demonstrates that the keys developed from OGA are preferable in terms of high entropic values, balanced bit values, and avalanche properties compared to other methods of fixed key generation in cryptography. The key generation process and automatic evolution of key quality in cryptography are represented through statistical parameters like mean and standard deviation with use of error bars and box plots. The key generation process demonstrates that cryptosystems optimized with artificial intelligence are effective in improving resistance levels over statistical cryptanalysis and differential cryptanalysis. The research indicates that artificial intelligence should be combined with cryptosystem generation mechanisms in progressing towards advanced adaptive cryptographic systems.