ACM SIGSOFT Software Engineering Notes
Evolutionary algorithm for prioritized pairwise test data generation
Proceedings of the 14th annual conference on Genetic and evolutionary computation
| Hi-index | 0.00 |
Combinatorial interaction testing has become an established technique to systematically determine test sets for highly-configurable software systems. The generation of minimal test sets that fullfill the demanded coverage criteria is an NP-complete problem. Constraint handling and integrated test case prioritization, features necessary for practical use, further complicate the problem. We present a novel algorithm that exploits our observation that the combinatorial interaction testing problem with constraints can be modelled as a single propositional logic formula. Our test set calculation algorithm uses binary decision diagrams as efficient data structure for this formula. The algorithm supports constraints and prioritization. Our evaluation results prove its cost effectiveness. For many benchmark problems the algorithm calculates the best results compared to other greedy approaches.