基于PVM的数据融合并行任务调度  

A Parallel Task Scheduling of Data Fusion Based on PVM

作　　者：万文福[1] 王小非[1] 夏学知[1] 涂葵[1] 

机构地区：[1]武汉数字工程研究所,武汉430074

出　　处：《舰船电子工程》2006年第3期5-8,12,共5页Ship Electronic Engineering

基　　金：国防预研基金(编号:41306050103)资助项目

摘　　要：根据多信息源、多平台数据融合算法特点,将数据融合与并行处理技术相结合,设计了基于PVM集群的多任务并行调度,提出了两层调度和主动报告的并行任务调度策略,并给出了数据融合并行任务调度映射和软件实现流程,在集群计算机系统上实现了数据融合的实时大容量并行处理。在系统仿真测试中获得了较高的加速比和系统效率,有较好的实用价值和应用前景。Connecting the data fusion algorithms with technology of parallel processing, this paper designs a parallel task schedule algorithrns based on Cluster of PVM（ Parallel Virtual Machine）, and proposes two - tier scheduling and active report strategies on the basis of the analyses.of the feature of multi - source and multi - plat data fusion algorithms, and the mapping methods and software of parallel task schedule are given on cluster computing system according to the algorithm. The higher speedup ration and system efficiency is acquired by some parallel algorithms in the system tested, the results indicate that the data fusion system is helpful.

关 键 词：数据融合 并行任务 任务划分 消息传递 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]