基于多维QoS约束的网格任务负载均衡调度模型及算法

Grid task scheduling model and algorithm based on load balance and multiple QoS-constained

为了在网格任务调度过程中更好满足用户多维QoS需求和提高系统负载均衡性能,提出一种基于多维QoS约束的网格任务负载均衡优化算法.综合考虑任务执行费用、执行时间和系统负载均衡性能,构造了新的评价用户综合满意度的连续性效应函数和反映系统负载均衡的异构网格任务调度模型.通过对遗传算法算子进行重新设计,实现了目标优化.仿真实验结果表明,在同等条件下该算法与同类算法相比,在负载均衡性能、QoS总效应值、任务完成率方面具有较好的综合性能.

In order to meet the user＇s multiple QoS and improve load balance performance, a grid task scheduling algo- rithm based on load balance and multiple QoS- constained（LGGA） is proposed. Through taking into account the task execu- tion cost, execution time and load balance performance, constructed a aggregated total utility function and heterogeneous grid task scheduling mode. Optimization objective was achieved by way of the design of new evolutionary operators. Simula- tion results showed that the proposed algorithm is better than the compared ones in terms of load balance performance, total utility function values, task completion rates.