网格任务调度算法的研究

Research on Task Scheduling Algorithm in Grid Computing

提出了一种基于遗传算法和Min-Min的网格调度算法,该算法主要分为资源搜索和任务分配两部分。首先,利用遗传算法从网格的可用资源中快速地搜索出最优的或是近最优的资源;然后从任务队列中取一组任务,利用Min-Min算法把任务分配给已搜索到的资源上去执行。仿真实验结果证明:对于大规模资源的调度该算法是有效的,能够减少了任务执行时间,改善网格任务调度的效率。

In this paper, we propose a new scbeduling algorithm based on genetic algorithm and Min-Min, which including two parts ; resource search and task allocation. At first, genetic algorithm is used to search for the optimal or ,schrboptimal resources in the available computing resources, then a set of tasks is pick up, finally, Min-Min algorithm is applied to schedule tasks to the optimal available resources. The simulation results show that for the scheduling of large-scale resources, the proposed algorithm is efficient. It reduces the task execution time and improves tbe efficiency of grid computing.