云环境下的一种并行任务划分方法研究

Algorithm for task partition in application grid

并行任务划分一直是高性能计算的研究重点。结合地震资料数据处理的应用云环境,以任务运行时间估计模型作为优化目标函数,提出了一种改进的粒子群优化算法,用以解决地震资料任务划分问题。仿真实验证明,改进后的算法增强了全局搜索能力,提高了收敛速度和收敛精度,有效提高了云环境下任务的执行效率。

The partition of parallel tasks has been the focus of grid computing. Combined with grid applications of seismic data processing, an improved particle swarm optimization （PSO） algorithm is proposed to solve the problem of task partition, with the expected time of task model as the optimizing target function. The comparative simulation results verify that the optimized PSO algorithm has improved global search capability, convergence rate and precision.