基于测试状态表的云测试平台任务调度算法  

Task scheduling algorithm of cloud test platform based on test state table

作　　者：朱信杰 王志[1] 金东生 ZHU Xinjie;WANG Zhi;JIN Dongsheng(th Department of Basic,North China Computing Technology Institute,Beijing 100083,China)

机构地区：[1]华北计算技术研究所基础四部,北京100083

出　　处：《电子设计工程》2021年第3期141-147,152,共8页Electronic Design Engineering

摘　　要：针对基于遗传算法的云测试平台任务调度研究中限制处理机数量、忽略任务间依赖、忽略处理机可以并行执行多个任务等问题,提出了一种基于测试状态表的云测试平台任务调度算法,依据测试任务的特点设计了调度算法的映射编码方法,提出了根据测试状态表计算任务完成时间的方法,根据云测试中的重要评价指标,改进了适应度函数、负载均衡度的计算公式。实验分析发现引入测试状态表可以得到更短的任务完成时间,也为处理机数量、任务间依赖和并行执行任务等问题提供了解决方案。Aiming at the problems of limiting the number of processors,ignoring inter-task dependencies,and ignoring the situation that processors can execute multiple tasks in parallel in the current research on task scheduling of cloud test platforms based on genetic algorithms,a task scheduling algorithm of cloud test platform based on the test state table is proposed.According to the characteristics of the test task,the mapping coding method of the scheduling algorithm is designed,and the method of calculating the task completion time according to the test status table is proposed.According to the important evaluation indicators in the cloud test,the fitness function and load balance of the calculation formula are improved.Experimental analysis found that the introduction of the test state table can get shorter task completion time,and also provides solutions for the number of processors,inter-task dependency,and parallel execution of tasks.

关 键 词：云测试 任务调度 遗传算法 测试状态表 适应度函数 

分 类 号：TN06[电子电信—物理电子学]