Evolutionary generation of test data for multiple paths coverage with faults detection

The aim of software testing is to find faults in the program under test. Generating test data which can reveal faults is the core issue. Although existing methods of path-oriented testing can generate test data which traverse target paths, they cannot guarantee that the data find the faults in the program. In this paper, we transform the problem into a multi-objective optimization problem with constrains and propose a method of evolutionary generation of test data for multiple paths coverage with faults detection. First, we establish the mathematical model of this problem and then a strategy based on multi-objective genetic algorithms is given. Finally we apply the proposed method in some programs under test and the experimental results validate that our method can find specified faults effectively. Compared with other methods of test data generation for multiple paths coverage, our method has greater advantage in faults detection and testing efficiency.