您的位置：首页 > 论文页面
基于双种群遗传算法的测试用例优先级排序
发表时间：2023-06-30 浏览量：487 下载量：76
| 全部作者： | 朱亚南，刘峰 |
| 作者单位： | 北京交通大学计算机与信息技术学院 |
| 摘 要： | 本研究将双种群遗传算法引入测试用例排序中以解决单一种群中过早收敛和最终解质量不稳定等问题，通过设置多样性较高的初始解，并在两个进化种群中使用不同的控制参数来协同进化，达到扩大解搜索空间的目的，以降低算法陷入局部最优的风险；同时使用引入权重因子的平均方法覆盖率作为适应度函数，利用 Boltzmann选择法实现不同进化阶段选择压力的自适应变化，期望加快算法后期收敛速度。最后在具有真实故障的数据集Defects4J上进行对比验证，结果表明：本文算法在平均故障检测率（average percentage of fault detection，APFD）方面优于单一种群遗传算法，且这种性能的提升在统计学上是显著的。 |
| 关 键 词： | 计算机软件；测试用例优先级排序（TCP）；双种群遗传算法；自适应选择压力 |
| Title： | Test case prioritization based on dual population genetic algorithm |
| Author： | ZHU Yanan, LIU Feng |
| Organization： | School of Computer and Information Technology, University of Beijing Jiaotong |
| Abstract： | In this paper, the dualpopulation genetic algorithm was introduced into the test case prioritization to solve the problems of premature convergence and unstable final solution quality in a single population. By setting an initial solution with high diversity and using different control parameters in the two evolutionary populations to co-evolution, the solution search space could be expanded to reduce the risk of the algorithm falling into local optimum. At the same time, the average method coverage rate with the weight factor was used as the fitness function. The Boltzmann selection method was used to realize the self-adaptation of the selection pressure at different evolution stages. It was expected to speed up the convergence speed of the algorithm in the later stage. Finally, a comparative verification was carried out on the dataset Defects4J with real faults. The results show that the algorithm in this paper is superior to the single population genetic algorithm in terms of average percentage of fault detection (APFD), and this performance improvement in the statistical is notable. |
| Key words： | computer software; test case prioritization (TCP); dual population genetic algorithm; adaptive selection pressure |
| 发表期数： | 2023年6月第2期 |
| 引用格式： | 朱亚南，刘峰. 基于双种群遗传算法的测试用例优先级排序[J]. 中国科技论文在线精品论文，2023，16（2）：223-232. |
1
评论数 0请您登录
暂无评论
下载全文
多维论文