The Research of Path-Oriented Test Data Generation Based on a Mixed Ant Colony System Algorithm and Genetic Algorithm

It is very practical significance to seek an effective path-oriented test data automatic generation method. The genetic algorithm, ant colony algorithm is commonly used to generate test data, and the both can improve the efficiency of test data generation. But, for both algorithms, there was a little limitation to target path in path testing for being prone to local optimal solution. Some researchers have combined the genetic algorithm and ant colony algorithm to generate the test data path, in which the result was better. At the same time, they found hybrid ant colony algorithm was still subject to the limitation of global search ability of ant colony algorithm. The Ant colony system algorithm is improved based on the ant colony algorithm. It is proved that it is more suitable for global search. In the present study, we propose to combine the ant colony system algorithm and genetic algorithm (ACSGA) to generate path-oriented software testing data. Classical triangle discrimination problem in path-oriented software testing is chose as a simulation experiment to verify ACSGA. The results show that the generation efficiency of target path has been improved apparently.