摘要： 有效的资源调度算法提高了任务的执行时间，对优化资源的使用起着非常重要的作用。在网格计算环境下，需要用统计预测的方法对任务的执行时间进行估计。该文提出了任务执行时间的组合预测方法，以任务过去执行时间的观察值为基础，用多种预测方法对任务的执行时间进行估计，用多种预测方法得出的估计值进行组合预测，给出任务执行时间的估计值。实验表明，组合预测要优于单一模型
预测。

关键词: 网格计算, 执行时间, 组合预测

Abstract: Effective resource scheduling algorithm improves the execution time of a given task, it plays an important role in optimizing resource, the need arises in estimating execution time of tasks for statistical prediction methods. This paper proposes the method of combining forecasts used to estimate the execution time of tasks. The forecasts are gotten by some prediction methods based upon past observations of the task executions. The estimating execution time of task is gotten by combining these forecasts. The theory and experimental results show that the result from combination forecast outperforms one from single model.

Key words: Grid computation, Execution time, Combination forecasts

中图分类号: 

• TP301.6