A Binary Program Vulnerability Mining Method Based on Heuris-tic Hybrid Symbolic Execution

This research delves deeply into the complementarity of fuzz testing and symbolic execution, aiming to overcome the path explosion problem and the difficulty of reaching deep layers of the program when using them separately, thereby significantly enhancing the efficiency and accuracy of vulnerability discovery. This paper conducts a comprehensive review of the latest advancements in binary vulnerability mining, symbolic execution, fuzz testing, and taint analysis both domestically and internationally. Through a thorough analysis and comparison, it proposes an adaptive near-target region generation strategy and a heuristic hybrid symbolic execution method. This method intelligently locates potential vulnerability regions, reduces the search space, and optimizes the execution path with heuristic strategies, effectively addressing the challenges of vulnerability discovery in complex programs and opening up a new path for efficient and accurate binary vulnerability mining.