首页 | 官方网站   微博 | 高级检索  
     

异构网格环境下任务分配的自主计算方法
引用本文:刘 涛,曾国荪,吴长俊.异构网格环境下任务分配的自主计算方法[J].通信学报,2006,27(11):139-143.
作者姓名:刘 涛  曾国荪  吴长俊
作者单位:同济大学,计算机科学与技术系,上海,201804;国家高性能计算机工程技术中心同济分中心,上海,201804
基金项目:国家自然科学基金;教育部科学技术研究项目;上海市高校网格技术E-研究院基金
摘    要:首先对网格环境的异构性进行描述,定义了关键计算域并给出其划分算法。刻画了基于多agent系统进行任务分配的机制,分析了在自主agent不同移动策略下网格各区域任务分布的不同。采用控制理论的方法加速agent的合理分布,从而可以充分利用关键计算域的异构特征。实例研究表明,提出的任务分配机制具有自配置和自优化特征,能够动态适应异构网格环境的要求。

关 键 词:异构网格  控制理论  自主计算  任务分配
文章编号:1000-436X(2006)11-0139-15
收稿时间:2006-07-20
修稿时间:2006-07-202006-10-06

Autonomic computing approach for task distribution in heterogeneous computational grid
LIU Tao,ZENG Guo-sun,WU Chang-jun.Autonomic computing approach for task distribution in heterogeneous computational grid[J].Journal on Communications,2006,27(11):139-143.
Authors:LIU Tao  ZENG Guo-sun  WU Chang-jun
Abstract:Firstly, heterogeneity of gird was featured and key computing field of grid was defined, in addition, its partition algorithm was presented. The task distribution mechanism based on multi-agent systems was described and diverse tasks distribution in areas of grid was analyzed with autonomous agents moving by different policies. Control theory methods were used to accelerate the appropriate distribution of agents in order to take advantage of the heterogeneity of key computing fields. Experiments demonstrate that the task distribution mechanism proposed can self-configure and self-optimize, also it can adapt to the heterogeneity of grid environment dynamically.
Keywords:heterogeneous grid  control theory  autonomic computing  task distribution
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《通信学报》浏览原始摘要信息
点击此处可从《通信学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号