Using a Fault Hierarchy to Improve the Efficiency of DNF Logic Mutation Testing

  • Authors:
  • Garrett Kent Kaminski;Paul Ammann

  • Affiliations:
  • -;-

  • Venue:
  • ICST '09 Proceedings of the 2009 International Conference on Software Testing Verification and Validation
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Mutation testing is a technique for generating high quality test data. However, logic mutation testing is currently inefficient for three reasons. One, the same mutant is generated more than once. Two, mutants are generated that are guaranteed to be killed by a test that kills some other generated mutant. Three, mutants that when killed are guaranteed to kill many other mutants are not generated as valuable mutation operators are missing. This paper improves logic mutation testing by 1) extending a logic fault hierarchy to include existing logic mutation operators, 2) introducing new logic mutation operators based on existing faults in the hierarchy, 3) introducing new logic mutation operators having no corresponding faults in the hierarchy and extending the hierarchy to include them, and 4) addressing the precise effects of equivalent mutants on the fault hierarchy. An empirical study using minimal DNF predicates in avionics software showed that a new logic mutation testing approach generates fewer mutants, detects more faults, and outperforms an existing logic criterion.