基于PSODE的网格任务调度算法  

Grid job scheduling algorithm based on PSODE

作　　者：于泓博[1] 李会[1] 韩德丽[1] 

机构地区：[1]齐齐哈尔大学通信与电子工程学院,黑龙江齐齐哈尔161006

出　　处：《计算机工程与设计》2012年第4期1447-1451,共5页Computer Engineering and Design

基　　金：黑龙江省教育厅科学技术研究基金项目(11541399)

摘　　要：针对网格计算中任务在各个资源之间的调度问题,提出了一种网格环境下PSODE的任务调度算法。该算法实现了计算资源、存储资源、带宽资源、数据资源的利用率最高化和代价最低化。对基本粒子群算法和差分进化算法进行了分析,通过构造算法函数、适应值函数和权重公式,建立了粒子群差分混合算法并对其进行优化,介绍了算法的实现过程。实验结果表明,该算法与其它调度算法比较,具有良好的性能。A kind of task scheduling algorithms PSODE （particle swarm optimization＇s differential evolution algorithm）, is pro- posed in grid environment for the problem of task scheduling among the various resources in grid computing. The algorithm has realized maximum utilization and cost minimization of computing resources, storage resources, bandwidth resources as well as data resources. First of all, particle swarm optimization and differential evolution algorithm are analyzed, then particle swarm＇s differential hybrid algorithm is established and its optimization by constructing the arithmetic functions, the fitness function and weight formula. Finally the algorithm＇s implementation process is introduced. Through experimentation, it is proved that the al- gorithm has good performance compared with other scheduling algorithms.

关 键 词：网格计算 任务调度 粒子群算法 差分混合算法 构造函数 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]