Selecting Products for High-strength T-wise Testing of Software Product Line by Multi-Objective Method
Yekan Qian, Zhang Cheng, Futian Wang · 2018
Combinatorial interaction testing is a hot topic in software testing. This approach has been applied to model-based testing of Software Product Line (SPL). The strategy is to generate a set of product configurations to achieve high t-wise coverage for satisfying test objective. The number of valid combinations increases explosively with the increase of interaction strength. Therefore, how to choose the minimum number of product configuration to achieve maximum t-wise coverage is the complex problem in high-strength feature interaction. This paper introduces a multi-objective method to solve such problem. We use three objectives to decide product selection: t-wise coverage, number of products and similarity. Our method can generate a set of solutions quickly and the maximum t-wise coverage can be achieved with the smallest number of products.