Bi-Criteria Models for All-Uses Test Suite Reduction

  • Authors:

  • Jennifer Black;Emanuel Melachrinoudis;David Kaeli

  • Affiliations:

  • Northeastern University;Northeastern University;Northeastern University

  • Venue:
  • Proceedings of the 26th International Conference on Software Engineering
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Using bi-criteria decision making analysis, a new modelfor test suite minimization has been developed that pursuestwo objectives: minimizing a test suite with regard to a particularlevel of coverage while simultaneously maximizingerror detection rates. This new representation makes it possibleto achieve significant reductions in test suite size withoutexperiencing a decrease in error detection rates. Usingthe all-uses interprocedural data flow testing criterion, twobinary integer linear programming models were evaluated,one a single-objective model, the other a weighted-sums bicriteriamodel. The applicability of the bi-criteria model toregression test suite maintenance was also evaluated. Thedata show that minimization based solely on definition-useassociation coverage may have a negative impact on the errordetection rate as compared to minimization performedwith a bi-criteria model that also takes into account theability of test cases to reveal error. Results obtained withthe bi-criteria model also indicate that test suites minimizedwith respect to a collection of program faults are effectiveat revealing subsequent program faults.