摘要
通过将种群划分为多个子种群,对每个子种群执行遗传操作,p个核心并行执行遗传算法搜索测试路径,以加速测试用例的生成;通过在处理核心之间迁移子种群的个体,使得交叉变异后得到的种群个体变得丰富,算法能更好地全局寻优,可以寻找到较多能够覆盖全部路径的测试用例.实验结果表明,与基于串行遗传算法生成测试用例相比,多核并行遗传算法并行生成测试用例能够生成较多覆盖全都路径的测试用例且运行时间少.
A population is partitioned into several sub-populations ,the genetic operations for each sub-population are executed ,and parallel genetic algorithms are executed by p cores to search the test paths to speed up generating test cases .The individualities for population are enriched by migrating thee individualities for sub-populations among cores ,the optimization solutions are searched globally better ,and more test cases for covering all the test paths are found .The experiment results show that compared with the sequential generating test cases method using genetic algorithm ,the multi-core parallel generating test cases method using parallel genetic algorithm can obtain more generated test cases and require less execution time .
出处
《微电子学与计算机》
CSCD
北大核心
2013年第11期149-153,共5页
Microelectronics & Computer
基金
广西自然科学基金项目(2011GXNSFA018152)
广西自然科学基金(桂科基0728033)
关键词
测试用例生成
并行遗传算法
多核系统
线程级并行
generation test cases
parallel genetic algorithm
multi-core systems
thread-level parallelism
