组合认知复杂度的程序谱软件错误定位方法.

Spectrum-based fault localization（SBFL） technology collects test case results and statements coverage information to calculate the suspiciousness of each statement. Cognitive complexity is a software complexity metric, and the code with high value is prone to make mistakes. In order to improve the performance of fault localization, this paper proposed a method combining statement level cognitive complexity and SBFL to rank statements. When two or more statements were assigned the same suspiciousness, the method calculated cognitive complex values of every statement. The higher cognitive complexity value corresponding to statement, the more likely that it contained a bug. This test dataset has 925 wrong edition, includes Java projects, C projects and C ++ projects. Experimental result demonstrates that the proposed method reduces fault localization cost that exists in SBFL approaches [ABSTRACT FROM AUTHOR]

软件缺陷的存在导致软件无法满足用户的需求,如何高效高质量地定位缺陷是消除软件缺陷的关键。基于模型的缺陷定位技术是当前的研究热点,可以用于检测软件系统故障找到软件失效的原因。现有基于模型的缺陷定位技术中,未考虑非相邻节点间传递依赖和测试用例对可疑度的影响,导致缺陷定位精度和效率低。提出了基于概率模型检测的软件缺陷定位方法（probabilistic model checking method for software fault location,PMCSFL）,首先提出一种程序概率模型用于提高模型的推理能力;然后设计了基于执行路径构建程序概率模型的学习算法;最后设计了基于概率模型检测的软件缺陷定位算法,用于缺陷定位分析。通过在公共数据集Siemens上进行实验和分析,表明了PMC-SFL方法与五种现有的缺陷定位方法 RankCP、BNPDG、Tarantula、SOBER和CT相比,具有更高的软件缺陷定位精度和效率。 [ABSTRACT FROM AUTHOR]