节点文献

基于多Agent的动态数据并行计算方法研究

Study on Parallel Computing Method of Dynamic Data Based on Multi-Agent

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 刘颖马斌王长涛黄宽李海峰

【Author】 LIU Ying;MA Bin;WANG Chang-tao;HUANG Kuan;LI Hai-feng;Information and Control Engineering Faculty,Shenyang Jianzhu University;

【机构】 沈阳建筑大学信息与控制工程学院

【摘要】 为实现动态数据的快速处理和计算,提出一种基于多Agent和分布式并行计算的动态数据处理方法。该方法以智能Agent作为计算和控制的最小单元,构建有向无环的动态网络拓扑结构,采用基于信任度的合同网模型,并与阈值相结合,实现动态数据处理和快速求解,体现系统的智能性和实时性。实验结果表明,该计算方式与传统计算方式相比,系统的计算速率显著提高,实现了动态数据的实时性处理。

【Abstract】 A method of dynamic data processing based on multi-Agent and distributed parallel computing is proposed to achieve rapid processing and calculation of the dynamic data.According to this method,the intelligent Agent is regarded as the minimum unit of calculation and control,and a dynamic network topology with directed acyclic structure is constructed.The contract net model based on trust is adopted and combined with the threshold.Thus,the dynamic data processing and fast solution can be achieved.Meanwhile,the intelligence and real-time performance of the system is embodied.The experimental results show that,compared with the traditional calculation,the calculation rate of the system is increased significantly and real time processing of dynamic data is implemented.

【基金】 住房和城乡建设部科学计划项目(2012-K8-33)
  • 【文献出处】 测控技术 ,Measurement & Control Technology , 编辑部邮箱 ,2014年07期
  • 【分类号】TP338.6
  • 【被引频次】4
  • 【下载频次】76
节点文献中: 

本文链接的文献网络图示:

本文的引文网络