Abstract: New job schedule algorithm TD_Max-min based on trust-driven is presented for computing grid. Using the same samples and under the same conditions, TD_Max-min algorithm with current schedule algorithms TD_Min-min & TD_Sufferage are compared with numerical experiments. Experimental results show that trust-driven schedule algorithms are better than traditional time-driven ones, especially when the number of tasks with strong trusting relation requirement is more than that with weak trusting or without trusting relation requirement.

Key words: grid computing, job schedule, grid security, trust model, trust-driven

摘要： 利用信任机制的概念，对传统的Max-min调度算法进行改进，提出一种信任驱动的网格任务调度算法——TD_Max-min。在相同的假定条件设置下，将其与现有的TD_Min-min算法和TD_Sufferage算法进行仿真和比较研究。结果表明，信任驱动的网格调度算法不仅优于传统的基于时间的调度算法，而且当具有强信任关系的任务数量大于弱信任关系及无信任关系的任务数量时，TD_Max-min算法的调度性能明显优于TD_Min-min算法。

关键词: 网格计算, 任务调度, 网格安全, 信任模型, 信任驱动

CLC Number: 

• TP391