Affiliations: [a] College of Computer and Information Technology, China Three Gorges University, Yichang, China
| [b] School of Computer Science and Technology, Wuhan University of Technology, Wuhan, China
Correspondence:
[*]
Corresponding author. Yi Song, College of Computer and Information Technology, China Three Gorges University, Yichang 443002, China. E-mail: [email protected].
Abstract: Fault localization is the critical but most expensive step in testing manufacturing software, effectively locating faults has become an increasingly concerned study. The existing spectrum-based fault localization techniques utilize spectrum information and specific prioritization algorithm to generate the suspiciousness as well as the ranking of statements. However, the effectiveness of fault localization in manufacturing software would be dramatically reduced once the statement involving bug is assigned with the same suspiciousness as other non-faulty statements. A multi-technique fusion approach (FA) is proposed based on suspicious rankings, which merges various of randomly selected fault localization techniques to minimize the difference between the numbers of statements that need to be examined (GAP) to find the bug respectively in the worst and best assumptions, further improve the effectiveness of fault localization. In addition, a novel metric for comparing fault localization techniques is developed. Experiments on Siemens Suite shows that our approach outperforms these selected techniques in the effectiveness.
Keywords: Fault localization, manufacturing software, spectrum information, multi-technique fusion, GAP
