CAJ | 학술논문

利用遗传算法进行测试数据自动生成是近年来的研究热点，其有效性高度依赖于适应度函数的选取和初始种群的筛选。该文探索将遗传算法应用到IMX(Integrated Management X-software)系统测试数据自动生成以提高其回归测试的质量，将IMX系统专业测试人员手动生成的测试数据作为基础测试数据，并提出一种基于测试路径对目标路径覆盖率的初始种群筛选标准。在三角形程序和IMX系统平台上的实验表明，所提方法在寻找测试数据时所用的时间和迭代次数较少，且生成的测试数据具有较好的多样性。
이용유전산법진행측시수거자동생성시근년래적연구열점，기유효성고도의뢰우괄응도함수적선취화초시충군적사선。해문탐색장유전산법응용도IMX(Integrated Management X-software)계통측시수거자동생성이제고기회귀측시적질량，장IMX계통전업측시인원수동생성적측시수거작위기출측시수거，병제출일충기우측시로경대목표로경복개솔적초시충군사선표준。재삼각형정서화IMX계통평태상적실험표명，소제방법재심조측시수거시소용적시간화질대차수교소，차생성적측시수거구유교호적다양성。
Using genetic algorithms to generate test data automatically is becoming a hot topic in recent years, the method on effectively generating data is highly dependent on choosing the proper fitness function and the selecting standard. The genetic algorithm is used on Integrated Management X-software (IMX) system to help it improve the quality of regression test. Those basic test data used in this paper are taken from the data that generated by professional testers in IMX, and an initial population selecting standard is proposed based on the coverage. Experiments on IMX and triangle program show that the proposed algorithm is more effective than others, for example, with less time and iteration the method can find the testing data correctly, especially on data variety.