A novel approach for combinatorial test case generation using multi objective optimization

Combinatorial testing is a promising technique for testing highly-configurable systems. Software systems become larger and more complex every day, and due to the time and cost limitations, it is infeasible to test everything in a software with a large configuration space, or a graphical user interface with many settings and events. Combinatorial testing generates an interaction test suite to discover faults caused by parameter interactions using a systematic sampling method. Generation of an optimal test suite in combinatorial testing despite the necessity of determining the execution order of test cases and considering constraints is a challenging process. In this paper, we consider the combinatorial testing as a multiobjective optimization problem and propose a novel approach to generate combinatorial test cases by considering all metrics simultaneously. The experimental results showed the effectiveness of the proposed approach in generating test suites with reduced size and increased priority while the higher priority test cases are generated earlier and the higher priority parameter-values appear more frequently.