本文研究移动网格环境下的依赖性任务调度问题,提出预调度加重调度的动态调度策略.预调度利用改进的HEFT算法将所有任务映射到当前可用资源,目标是总完成时间最短,同时考虑了资源的能量约束条件.任务运行过程中监测资源的各种异常行为,当满足重调度触发条件时进行重调度,从而实现对动态环境的自适应性.通过实例分析表明了调度策略的有效性.
Dependent task scheduling problem in mobile grid environment is studied and dynamic scheduling strategy is presented which includes pre--scheduling and re-scheduling. Pre-scheduling maps all tasks to currently available resources using modified HEFT algorithm, with goal of shortest completion time, taking into account the energy constraint of resources. During the running of tasks, resources will be monitored. If resource behavior is abnormal and meets rescheduling conditions, then re-scheduling is triggered to achieve adaption to dynamic environment. Experiment results show the effectiveness of presented scheduling strategy.