摘要： 任务调度是网格研究的核心问题之一，在研究网格任务调度问题的基础上，利用模糊聚类思想提出将网格任务与资源进行混合模糊聚类的网格独立任务调度算法，该算法将最适合的资源分配给与之相适应的任务，即尽量将任务调度到恰好满足其需求的资源上执行，从而把综合能力大大超过当前任务的资源“预留”给将来的任务使用，算法具有良好的性能和负载均衡效果，为网格任务调度提供一种新的思路。

关键词: 网格, 任务调度, 模糊聚类, 任务资源混合聚类, 负载均衡

Abstract: Task scheduling is one of the key problems in the field of grid computing research area. Based on analyzing research works done in grid task scheduling field, this paper gives out a novel gird independent task scheduling algorithm based on hybrid clustering of grid tasks and resources. By using the method of fuzzy clustering of task and resources, this algorithm tries to schedule the task to the grid resource which exactly satisfies the needs of the task, and tries to avoid allocating the tasks to grid resources that extremely exceed the task’s needs. In this way, the algorithm reserves the competent grid resources for future use when they are needed by complex grid tasks, while efficiently satisfy the needs of current tasks. This novel algorithm has high efficiency and sound load balancing capability, and provides a new way for grid tasks scheduling.

Key words: grid, task scheduling, fuzzy clustering, task &, resource hybrid clustering, load balancing

中图分类号: 

• TP311