非传统安全关键系统中模糊分类调度模型  

Real-time scheduling model of fuzzy itemized tasks for non-traditional safety critical system

作　　者：彭礼强[1] 尹俊文[2] 汪飞[1,3] 

机构地区：[1]湖南师范大学数学与计算机学院,长沙410081 [2]国防科技大学计算机学院,长沙410074 [3]湖南师范大学数学与计算机科学学院高性能计算与随机信息处理省部共建教育部重点实验室,长沙410081

出　　处：《计算机工程与应用》2011年第27期55-59,共5页Computer Engineering and Applications

基　　金：湖南省高校科技创新团队支持计划;湖南省科技计划项目(No.2010GK3024);湖南师范大学青年基金资助项目(No.53112-1876)

摘　　要：物联网及无线传感网的发展将使越来越多的应用成为安全关键系统,而这类应用在性质上区别于医疗、核能和航空航天等领域的传统安全关键系统。研究了非传统安全关键系统的调度特征,发现这类应用与传统安全关键系统存在较大差异,特别是在任务分类上,若直接套用已有的调度模型将产生诸多问题。在特征分析的基础上建立了满足特征需求的实时调度模型并给出了相应的调度算法。模型采取模糊分类的方式划分任务,算法分两个阶段实现了任务按敏感度优先的原则进行调度。最后对算法进行了调度可行性分析,并在长沙市非税系统中进行了实验,结果表明其性能及失效率较MUF更优。The development of IoT and WSN may take more applications into safety-critical systems.But those applications have different characters with traditional safety-critical systems,such as medical systems,nuclear power systems and aeronautical systems.Study on scheduling characters of non-traditional safety critical systems is carried on in this paper.And great differences are founded between the non-traditional and the traditional,especially on itemization of tasks.It will cause many problems while applying existing scheduling models to those applications directly.Scheduling characters of non-traditional safety-critical systems are set up.A real-time scheduling model and MSF algorithm to meet those characters,where tasks are fuzzy itemized and the algorithm includes two steps to schedule tasks ordered by sensitity.Finally,a feasibility analysis and a test on the non-taxable system of Changsha city are taken.The result shows that MSF holds better performance and less failure number of tasks than MUF.

关 键 词：安全关键系统 调度特征 模糊分类 敏感度优先 实时调度模型 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]