Software vulnerability detection using genetic algorithm and dynamic taint analysis

In order to solve the problems of traditional Fuzzing technique for software vulnerability detection, this paper proposes a novel method based on genetic algorithm and dynamic taint analysis. First, static analysis is applied to calculate the critical path information, including danger functions, high cyclomatic number functions and loop structures. Second, dynamic taint analysis is introduced to identify the key bytes to reduce the input space. Third, the genetic algorithm fitness function is constructed based on the critical path information to guide the test case generation and the genetic operators are executed on the reduced input space. Experiments show that the method could obtain higher vulnerability detection accuracy and efficiency.