摘要
在分析现有启发式调度算法的优缺点的基础上,综合考虑负载平衡和使尽可能多的任务调度到更适合它的机器上,提出了基于动态适应度(DFD)的调度算法,任务的DFD描述了最早完成时间所在机器对任务的适应程度,本算法以任务的DFD为优先级,优先调度DFD高的任务。与参考文献中其他算法进行比较表明,本算法能产生更好的调度结果。
After analyzing the advantages and disadvantages of the scheduling heuristics in the literature, in order to get better load balance and schedule more tasks to its more suitable machines at the same time, a new scheduling algorithm based on Dynamic Fit Degree （DFD） was put forward. The suitability of the machine with minimum time for finishing the task was described by task＇s DFD schema. The proposed algorithm used DFD as the priority, and the task with biggest DFD was scheduled first in each step. The simulation results compared with other scheduling algorithms show that the proposed algorithm produces better results.
出处
《计算机应用》
CSCD
北大核心
2006年第12期3001-3003,共3页
journal of Computer Applications
关键词
异构计算系统
独立任务调度
动态适应度
最早完成时间
heterogeneous computing systems
independent tasks scheduling
dynamic fit degree
minimum completion time
