网格环境下基于免疫克隆策略的资源分析和任务调度

网格将多种资源组织在一个统一的框架下,为各种复杂的计算任务提供资源.因此计算任务在各种资源之间的调度成为了一个关键的问题.将遗传算法和免疫克隆算法结合起来,充分发挥各自的优势,先通过克隆操作来产生一组新的个体,然后再独立地对所产生的各个体进行克隆变异,克隆交叉操作,从而求得问题的最优解.实验结果表明,这种并行遗传免疫克隆算法全局寻优能力强,能有效地实现资源的负载均衡,可以成功地应用于网格任务调度中.

Resource allocation and task scheduling based on immune clonal strategy in grid environment

Gr id is a new comput ing f ramew ork to col lect remote resour ces for computer tasks. T ask scheduling in these hetero geneous reso urces is a crit ical pr oblem. T he immune clonal algorithm w as combined with g enetic alg orithm to fully ex ert respectiv e advantages. A new g roup o f individuals was generated through clonal oper at ion, and then all the g ener ated indiv iduals w ere cross-cloned respect ively. Ex periment s show that the parallel genet ic immune clo nal alg orithm has a go od g lobal search capacity , and can be applied to the opt imizat ion of task scheduling in grid enviro nment .