A Strategy for Grid Based T-Way Test Data Generation

Although desirable as an important activity for ensuring quality assurances and enhancing reliability, complete and exhaustive software testing is next to impossible due to resources as well as timing constraints. While earlier work has indicated that pairwise testing (i.e. based on 2-way interaction of variables) can be effective to detect most faults in a typical software system, a counter argument suggests such conclusion cannot be generalized to all software system faults. In some system, faults may also be caused by more than two parameters. As the number of parameter interaction coverage (i.e. the strength) increases, the number of t-way test set also increases exponentially. As such, for large system with many parameters, considering higher order t-way test set can lead toward combinatorial explosion problem (i.e. too many data set to consider). We consider this problem for t-way generation of test set using the Grid strategy. Building and complementing from earlier work in In-Parameter-Order-General (or IPOG) and its modification (or MIPOG), we present the Grid MIPOG strategy (G_MIPOG). Experimental results demonstrate that G_MIPOG scales well against the sequential strategies IPOG and MIPOG with the increase of the computers as computational nodes.