混合测试用例自动生成算法

A Method of Hybrid Test Cases Auto Generation

采用传统智能优化算法进行测试用例自动生成已经取得一定研究成果,但是还存在算法效率不高的问题。花朵授粉算法是新发现的一种群体智能优化算法,将该算法应用到测试用例自动生成方面具有寻优精度高、可操作性强等优点,但会产生早熟现象,无法跳出局部最优解,从而造成收敛速度慢、计算效率低等问题。对此,提出了一种混合测试用例自动生成算法,将群体爬山算法思想混合并融入花朵授粉算法中,保留了寻优精度高等优点,同时提高了标准花朵授粉算法的脱困能力和运算效率。实验结果表明,该算法在测试用例自动生成上精度较高,同时在收敛速度和计算效率方面比标准花朵授粉算法、粒子群算法有较大提高。

The automatic generation of test cases using traditional intelligent optimization algorithm has obtained some research results,butthe algorithm’s efficiency is low. Flower pollination algorithm is a new heuristic optimization algorithm,and its application to automaticgeneration of test cases has the advantages of high precision of optimization and high operability,but it will produce premature phenome-non and cannot get out of local optimal solution,resulting in slow convergence and low computational efficiency. Therefore,we proposea hybrid algorithm of automatic generation of test cases,which combines the idea of colony mountain climbing into the flower pollinationalgorithm. This algorithm preserves the advantages of high precision and improves relief capability and computational efficiency of thestandard flower pollination algorithm. The experiment shows that the proposed algorithm has higher precision in automatic generation oftest cases and has a greater convergence speed and computational efficiency than standard flower pollination algorithm and particle swarmoptimization algorithm.