PSOFuzzer: A Target-Oriented Software Vulnerability Detection Technology Based on Particle Swarm Optimization

Coverage-oriented and target-oriented fuzzing are widely used in vulnerability detection. Compared with coverage-oriented fuzzing, concentrates more computing resources on suspected vulnerable points to improve the testing efficiency. However, sample generation algorithm detection technology has some problems, such as weak guidance, penetration, difficult generation. This paper proposes a new fuzzer, PSOFuzzer, that uses particle swarm optimization generate samples. PSOFuzzer can quickly learn high-quality features historical samples implant them into be led execute point. The experimental results show test process reach target point trigger vulnerabilities 79% 423% higher probability than AFLGo Sidewinder, respectively, tested software programs.