<![CDATA[Digital data security demands robust encryption systems in which key randomness quality serves as the primary determining factor. Metaheuristic algorithms such as the Genetic Algorithm (GA) and Simulated Annealing (SA) exhibit significant potential for key generation optimization. However, each is individually susceptible to premature convergence and slow computational time, respectively, motivating their sequential hybridization. This study proposes a GA-SA hybrid cryptographic architecture with dynamic population sizing to optimize pseudo-random keystream generation in XOR encryption, evaluated using 15 PDF document datasets across three key configurations: 16 characters (128-bit), 32 characters (256-bit), and 64 characters (512-bit). The hybrid system consistently reduced local optima entrapment across all configurations, with the 64-character key achieving the highest randomness quality at a Shannon Entropy of 7.9288 bits/byte and a mean NIST SP 800-22 Monobit Frequency Test P-Value of 0.2999, though this does not constitute a full NIST SP 800-22 suite evaluation. Runtime analysis showed near-linear empirical growth within the tested range, from 0.0361 seconds to 0.1305 seconds, without exponential bottleneck effects, suggesting the proposed architecture is a promising candidate for pseudo-random keystream generation under tested conditions, with further validation recommended before production deployment.]]>
Copyrights © 2026
