A Heuristic Guided Optimized Strategy for Non-Deterministic Mutation
Zhen Cai, Honglin Wang, Xiaojun Qin · Proceedings of the 3rd International Conference on Computer Science and Application Engineering · 2019
AFL (American Fuzzy Lop) is one of the most popular fuzzy test tools. Aiming at the problem of insufficient path coverage caused by complete random mutation in the non-deterministic mutation stage, this paper proposes a heuristic guided optimized strategy for non-deterministic mutation, and implements AFLCAI on the basis of AFL. AFLCAI uses the effector map mechanism to obtain the approximation of metadata, and improves the branch coverage and the number of path coverage by heuristic guided mutation. The comparison experiment proves that AFLCAI can effectively improve the code coverage without affecting the running speed. The branch coverage rate is increased by 3.79% and the new path is increased by 9.90%, which confirms the effectiveness and advantages of the proposed method.