| Abstract [eng] |
Offline dictionary attacks against WPA2 remain computationally demanding, particularly in resource-constrained digital forensics environments. While conventional approaches primarily emphasize hardware acceleration, this study investigates heuristic optimization and sorting of static wordlists. Five strategies, Alphabetical, Reverse Alphabetical, Length, N-gram, and a novel Structural method, are evaluated through a three-stage experimental pipeline comprising synthetic string-matching simulations, GPU-accelerated WPA2 hash recovery using Hashcat, and cross-dataset validation. The validation includes 1,000-target stress tests for Complex and Transliterated password categories. The study also examines regional password behaviors associated with Cyrillic-to-Latin transliteration and spatial QWERTY Keyboard Layout Patterns in Cyrillic-dependent regions. Under the evaluated conditions, Alphabetical sorting was the most effective strategy for high-entropy and security-compliant passwords, producing speedups exceeding 2.0× with high statistical stability. Structural sorting performed best for low-entropy Numeric and Dictionary patterns, achieving speedups of up to 4.13×. For regionally influenced credentials, N-gram sorting accelerated Transliterated password recovery by 1.55×, although its practical value was limited by substantial preprocessing time. The results further demonstrate that dictionary size, password structure, hardware throughput, and preprocessing overhead jointly determine the operational value of each strategy. Based on these findings, an empirically grounded Heuristic Decision Framework is proposed to assist authorized digital forensics practitioners in selecting a wordlist strategy according to the expected target profile, dataset scale, available hardware, and the required balance between preparation cost and Time-to-Crack. |
