Identifying Interactions for Combinatorial Testing using Data Flow Techniques

Abstract

Combinatorial (t-way) testing has been proved to be an effective technique that identifies faults caused by interactions among a small number of input parameters. However, the degree of interaction to be considered for testing is still an open issue. Although higher strength t-way testing improves fault detection, it leads to a considerable increase in number of interactions to be tested and hence the test set size. In this paper, an approach is proposed that attempts to reduce the number of interactions to be tested. The source code is transformed into a flow graph and data flow analysis is applied to it to identify the interactions that exist in the system. The initial results suggest that the approach is able to achieve a considerable reduction in the number of interactions to be tested.