Generating Optimal Test Set for Neighbor Factors Combinatorial Testing

Combinatorial testing is a specification-based testing method, which can detect the faults triggered by interaction of factors. For one kind of software in which the interactions only exist between neighbor factors, this paper proposes the concept of neighbor factors combinatorial testing, presents the covering array generation algorithms for neighbor factors pair-wise (N=2) coverage, neighbor factors N-way (Nges2) coverage and variable strength neighbor factors coverage, and proves that the covering arrays generated by these three algorithms are optimal. Finally we analyze an application scenario, which shows that this approach is very practical