Robust and Efficient Methods for Key Generation using Chaotic Maps and A2C Algorithm

In the current digital landscape, information security has become a critical necessity given the escalating frequency and sophistication of cyberattacks across global computing networks. Cryptography is the science and practice of securing information by transforming it into a format that is unreadable or inaccessible to unauthorized parties. The strength of the cryptography algorithm lies in the strength of used encryption key. Recently, the nonlinear behavior of chaotic maps has been utilized as a random source to generate robust key stream bits for cryptographic purposes. The aim of this paper is to introduce an efficient and robust system for generating strong key stream bits against different types of attacks. The proposed system implemented six proposed scenarios for generating a highly strong key stream bits based on different 5-types of chaotic maps (Tent, Ikeda, Chua's, Rössler and Double Pendulum). In the first five scenarios, use each map with Deep-Reinforcement Learning (DRL) algorithm called Advantage Actor-Critic (A2C), while in the sixth scenario, the fusion of all above five chaotic maps with Advantage Actor-Critic (A2C) are used at the same time. For each scenario of the proposed system generates three different lengths of key stream bits (128-bit, 192-bit, and 256-bit). To evaluate the robustness, randomness, and effectiveness of the proposed system, several types of tests were applied like NIST, brute-force attack, Auto Correlation (AC), Cross Correlation (CC), and Discrete Fourier Transform (DFT). All the obtained results indicated to the highly robustness and highly strength of generated key stream bits for all six proposed scenarios of the proposed system.