Abstract: To address the task schedule problem in grid resource management, an immunity schedule algorithm for grid task is put forward. The algorithm abides by the immune principles of clonal selection and affinity maturation, and seeks the all-around excellent result to the schedule problem for grid task. The settings of group generation and parameters of the algorithm are discussed, which can influence the performance of the algorithm. Experimental results of simulation demonstrate that the algorithm, compared with the conventional algorithm of grid task schedule, has the virtues of rapid task schedule, short-time resource allocation and steady function.

Key words: grid, artificial immunity, task schedule

摘要： 针对网格资源管理的任务调度问题，提出一种网格任务免疫调度算法。算法遵循克隆选择、亲和度成熟2个免疫原理，求解网格任务调度问题的全局最优解。讨论种群代数设置和算法参数的设置对该算法性能的影响。仿真实验结果表明，与传统的网格任务调度算法相比，该算法具有任务调度速度快、资源分配时间短、运行稳定等优点。

关键词: 网格, 人工免疫, 任务调度

CLC Number: 

• TP391