基于遗传算法的网格资源任务调度策略研究

网格是当前高性能计算领域的一个研究热点。首先分析了网格任务的特点,给出了网格任务模型,在此基础上提出了一种层次结构的网格资源管理任务调度系统模型。采用了遗传算法作为调度策略,给出了调度策略的每一步具体实现方案。此模型可为网格的资源管理提供一个统一的高层管理框架,能实现全局最优调度,为网格调度提供了一种有效的方法和途径。

Research on the Strategy of Grids Resource Management Scheduling Based on Genetic Algorithm

Grid is one of the research hotspots in high performance computing field.Efficient scheduling of grid applications in grid environment reveals several challenges due to its high heterogeneity,dynamic behavior,and space shared utilization.The characteristics of gird tasks are studied,and a common model of grid tasks is built,a hierarchy grid resource management scheduling model is put forth.An advanced genetic algorithm(GA) is used in scheduling strategy.A concrete plan to put every step of the scheduling strategy into practice is given.In this way,this model can provide a uniform upper-level management framework for grid resource management and it also can realize the global optimum scheduling.It is an effective approach for grid scheduling.