Fault class prioritization in Boolean expressions

  • Authors:

  • Ziyuan Wang;Zhenyu Chen;Tsong-Yueh Chen;Baowen Xu

  • Affiliations:

  • Nanjing University, Nanjing, China;Nanjing University, Nanjing, China;Swinburne University of Technology, Victoria, Australia;Nanjing University, Nanjing, China

  • Venue:
  • Proceedings of the 27th Annual ACM Symposium on Applied Computing
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

A recent study has classified faults in Boolean expressions into ten classes and has proved that there are five key fault classes, namely CCF, CDF, ORF, ENF and ASF, such that if a test suite can kill all faulty versions of these five core fault classes, if can kill all faulty versions of all fault classes. In order to generate more effective test suites, we should prioritize these five fault classes further, such that test cases with stronger fault detection capability could be generated as early as possible. Such a process is referred to as the fault class prioritization. Based on the observation in the fault class hierarchy, we divide the five fault classes into two groups {CCF, CDF} and {ORF, ENF, ASF}. Two strategies of fault class prioritization are proposed to generate test cases efficiently. We design experiments using TCAS Boolean expressions and some randomly generated Boolean expressions. The experimental results suggest that if we generate test cases for CCF and CDF firstly, the final test suite always have a higher efficiency of killing faults.