A Test Case Generation Method of Combinatorial Testing based on τ-way Testing with Adaptive Random Testing

Abstract

Combinatorial testing is an effective software testing technique, which has gained wide attention on industry and academic. It detects faults triggered by the interactions among parameters relevant to software through selection of a reasonably sized set, which consists of the combination of the values of these parameters. However, as the complexity of software system increases, the time cost increases greatly, which leads how to efficiently generate the smallest coverage array under the given input parameter model to become the major sticking points in some scenarios. In order to address this issue, by analyzing existing generation algorithms, it is found that these algorithms are based on the complete input parameter model constructed in the first step of combinatorial testing. This paper proposes a test case generation method of combinatorial testing based on $\tau$-way testing and adaptive random testing which test cases can be generated partially using $\tau$-way strategy and partially using adaptive random testing by splitting the input parameter model, so as to achieve a balance between effectiveness and efficiency in a specific scenario. To this end, experimental results show that the proposed method has better faults detection ability and the computational overhead of test case generation on subject Tcas program.