Reducing the Number of Higher-Order Mutants with the Aid of Data Flow

Authors

Ahmed S. Ghiduk

Abstract

Higher-order mutants are created by injecting two or more mutations into the original program, while first-order mutants are generated by seeding single faults in the original program. Mutant generation is a key stage of mutation testing which is computationally very expensive, especially in the case of higher-order mutants. Although many mutation testing techniques have been developed to construct the first-order mutants, a very small number of techniques have been presented to generate the higher-order mutants because of the exponential growth of the number of higher-order mutants, and the coupling effect between higher-order and first-order mutants. To overcome the exponential explosion in the number of higher-order mutants considered, this paper introduces a new technique for generating a reduced set of higher-order mutants. The proposed technique utilizes a data-flow analysis to decrease the number of mutation points through the program under test and consequently reduce the number of higher-order mutants. In this technique only positions of defs and uses are considered as locations to seed the mutation. The generated set of higher-order mutants consists of a reduced number of mutants, which reduces the costs of higher-order mutation testing. In addition, the proposed technique can generate the higher-order mutants directly without generating the first-order mutants or by combining two or more first-order mutants. A set of experiments are conducted to evaluate the effectiveness of the proposed technique. The results of the conducted experiments are presented and compared with the results of the related work. These results showed that the proposed technique is more effective than the earlier techniques in generating higher-order mutants without affecting the efficiency of mutation testing.

Keywords

combinatorial interaction testing, pair-wise testing, covering array, artificial bee colony