计算机科学 ›› 2017, Vol. 44 ›› Issue (4): 124-130.doi: 10.11896/j.issn.1002-137X.2017.04.027

• NASAC 2015 • 上一篇    下一篇

基于故障关联的多故障概率诊断方法

徐俊洁,陈荣   

  1. 大连海事大学信息科学技术学院 大连116026,大连海事大学信息科学技术学院 大连116026
  • 出版日期:2018-11-13 发布日期:2018-11-13
  • 基金资助:
    本文受国家自然科学基金(61175056)资助

Probabilistic Diagnosis Approach to Diagnosing Multiple-fault Programs with Fault Correlation

XU Jun-jie and CHEN Rong   

  • Online:2018-11-13 Published:2018-11-13

摘要: 故障软件诊断的必要性在于真实世界中的软件几乎都会包含一个以上的故障。与单故障不同,多个故障的传播及其关联导致软件诊断更复杂,不确定性更高,概率推理因而被用于适应多故障程序的特殊性。提出了一种新的基于变形概率图FCG及其推理的软件诊断方法。相比于BARINEL方法和经典的贝叶斯网,FCG的特别之处在于采用了无向图上候选故障及其关联关系的贝叶斯推理和Noisy-or推理,而候选故障及其关联可以从程序语句间的控制依赖关系和数据依赖关系中创建。从西门子套件到更大的space,grep程序的实验,无论是在处理单故障还是处理多故障的情况下,实验结果都证明了FCG的有效性,其诊断效果比LOUPE,Ochiai,Tarantula甚至BARINEL方法都准确。

关键词: 多故障,故障关联与不确定性,概率推理,控制和数据依赖关系,变形概率图

Abstract: Diagnosing multiple-fault software is necessary because almost all real-world software contains more than one fault.Unlike single-fault,the propagation and correlation of multiple faults in software lead to more complexity and great uncertainty,and probabilistic reasoning is thus applied to accommodate such uniqueness.This paper proposed a new probabilistic reasoning method to diagnose multiple-fault software by using variant probabilistic graphs FCG and their inference.Distinguished from the BARINEL method and the classical Bayesian network,FCG features Bayesian and Noisy-or inference from undirected graph consisting of candidate faults and their correlation which can be set up from statement-level control and data dependencies.Experiments were conducted on programs ranging from Siemens suite to larger ones like space and grep.The experimental results validate the effectiveness of the present approach in handling programs no matter with single fault and with multiple faults,and especially it is more accurate than competitors such as LOUPE,Tarantula,Ochiai and even BARINEL.

Key words: Multiple-fault,Fault correlation and uncertainty,Probabilistic reasoning,Control and data dependencies,Variant probabilistic graph

[1] JONES J A,HARROLD M J.Empirical evaluation of the tarantula automatic fault-localization technique[C]∥Proceedings of the ACM International Conference on Automated Software Engineering.New York,USA,2005:273-282.
[2] MASRI W,ABOU-ASS R,EL-GHALI M,et al.An empiricalstudy of the factors that reduce the effectiveness of coverage-based fault localization [C]∥Proceedings of the 2nd International Workshop on Defects in Large Software Systems.Chicago,Illinois,USA,2009:1-5.
[3] BAAH G K,PODGURSKI A,HARROLD M J.The Probabilistic Program Dependence Graph and Its Application to Fault Diagnosis[J].IEEE Transactions on Software Engineering,2010,36:528-545.
[4] FENG M,GUPTA R.Learning universal probabilistic models for fault localization [C]∥Proceedings of ACM SIGPLAN-SIGSOFT Workshop on Program Analysis for Software Tools and Engineering.Toronto,Ontario,Canada,2010:81-88.
[5] ABREU R,ZOETEWEIJ P,GOLSTEIJN R,et al.A practicalevaluation of spectrum-based fault localization [J].Journal of Systems and Software,2009,82(11):1780-1792.
[6] ABREU R,ZOETEWEIJ P,VAN Gemund A J C.A NewBayesian Approach to Multiple Intermittent Fault Diagnosis[C]∥Proceedings of International Joint Conference on Artifical Intelligence.San Francisco,2009:653-658.
[7] DIGIUSEPPE N,JONES J A.On the influence of multiple faults on coverage-based fault localization[C]∥Proceedings of the International Symposium on Software Testing and Analysis.ACM,2011:210-220.
[8] YU K,LIN M,GAO Q,et al.Locating faults using multiplespectra-specific models[C]∥Proceedings of ACM Symposium on Applied Computing.2011:1404-1410.
[9] PEARL J.Probabilistic reasoning in intelligent systems:net-works of plausible inference [M].Morgan Kaufmann Publishers,1988.
[10] KOLLER D,Friedman N.Probabilistic Graphical Models:Principles and Techniques [M].The MIT Press,2009.
[11] BAO X A,XIE X M,ZHANG N,et al.Optimized Software Testing Strategy Based on the Defect Correlation Markov Model [J].Journal of Software,2015,26(1):14-25.(in Chinese) 包晓安,谢晓鸣,张娜,等.基于缺陷关联度的Markov模型软件优化测试策略[J].软件学报,2015,6(1):14-25.
[12] WOTAWA F.Fault Localization Based on Dynamic Slicing and Hitting-Set Computation[C]∥The 10th International Confe-rence on Quality Software,2010.2010:161-170.
[13] WEN W Z,LI B X,SUN X B,et al.Technique of Software Fault Localization Based on Hierarchical Slicing Spectrum [J].Journal of Software,2013,24(5):977-992.(in Chinese) 文万志,李必信,孙小兵,等.一种基于层次切片谱的软件错误定位技术[J].软件学报,2013,4(5):977-992.

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!