Rearch And Application Of An Effective Evaluation Mechanism To Guide Test Case Generation

Test case’s generation is an essential step in the testing of software to increase the automation degree of testing. In recent years, people try to use intelligent search algorithm to resolve this problem and the result is favorable. Currently, most researches focus on improving the course by optimizing the flow of algorithm or selection of parameter. Different from those researches, this paper’s research is based on how to improve the evaluation mechanism.Now, the most popular method of evaluation is Korel’s branch function method. The shortcoming of this method is that it can’t tell how much the test case meets the branch condition; it can only tell how much the test case doesn’t meet the branch condition. For this shortcoming, this paper brings forward a new evaluation mechanism. Compare to the traditional one, the new mechanism can tell how much the test case meets the branch condition.In order to verify the improving is effective, this paper designs an experiment. First of all, we design the structure of a system used to generating test cases and select three algorithms containing genetic algorithm, particle swarm optimization and estimation of distribution algorithm. Then, we code the program of experiment system using the two evaluation mechanism and select the mark of comparison. At last, we take the experiment and analyses the result of the experiment. The experimental data show that the new method compared with the traditional one has been improved in efficiency and other indicators of result data.