An approach for improving stability of path-oriented test data generation

摘要

Path coverage testing is an important method for software structural testing. The application of evolutionary algorithms to path test can not only reduce the cost, but ensure the reliability of software quality. The research on applying genetic algorithm to test data generation of path coverage has gained substantial achievements in recent years. Because the accuracy of searching is not high enough to meet the requirements for generating path test data, the results are unstable in generic algorithm. In this paper, we propose multivariate optimization algorithm to solve this problem. Experimental results show that multivariate optimization algorithm achieves better searching precision and lower volatility in the search process, which provides a new method for the path-oriented Test data generation with higher quality.