MIPOG - An Efficient t-Way Minimization Strategy for Combinatorial Testing

Abstract

This paper presents a study comparing different techniques to achieve minimal test suites in combinatorial testing. Considering high interaction strength is not without difficulties. When the number of parameter coverage increases, the size of t-way test sets also increases exponentially, hence, resulting into combinatorial explosion problem. Addressing these aforementioned issues, a new strategy capable of supporting high interaction strength, called Modified IPOG (MIPOG) is proposed. Similar to its predecessor IPOG (In Parameter Order General), MIPOG adopts the horizontal and vertical extensions in order to construct the desired test set. However, unlike IPOG, MIPOG optimizes both the horizontal and vertical extensions resulting into a smaller size solution than that of IPOG, (i.e., with the test size ratio ≤ 1). In fact, MIPOG, in most cases, surpasses some IPOG variants (IPOD, IPOF1, and IPOF2) as well as other existing strategies (Jenny, TVG, TConfig, and ITCH), as far as the test size is concerned with an acceptable execution time. Additionally, MIPOG has also contributed to enhance many known CA and MCA that exist in the literature.