摘要
针对不确定任务特征 ,提出应用模糊理论进行动态抢占调度 .用语言模糊集来描述任务的不确定特征和不同的优先级等级 ,利用最大隶属度原理确定任务的优先级等级 ,采用优先调度高优先级等级任务的调度策略提高重要任务的调度成功率 ,实现具有不确定任务特征的抢占调度 .与传统的EDF和LSF算法相比较 ,仿真表明 ,所提算法能够提高重要任务的调度成功率 ,并降低重要任务的截止期错失率 ;同时 ,任务间的平均切换次数大大小于LSF的平均切换次数 ,而与EDF保持相当 .该方法可应用于计算机控制系统的控制任务调度 ,并借鉴于其它具有不确定任务特征或具有有限优先级等级的实时调度问题研究中 .
Aiming to uncertain task characteristics, it is presented to apply fuzzy theory to dynamic preemptive scheduling. Making use of linguistic fuzzy sets for describing uncertain characteristics and different priority levels, the principle of maximum membership function for determining task priority level, and the policy of scheduling tasks with high priority lever first for increasing the scheduling success ratio of critical tasks, it realizes the preemptive scheduling of tasks with uncertain characteristics. Compared to the classical EDF and LSF, the simulations show that the proposed approach can improve the scheduling success ratio of critical tasks and reduce their mean missed deadline rate, meanwhile, mean switching number among tasks gets decreased greatly compared to LSF and is close to EDF. The proposed method can be applied to control task scheduling in computer-controlled systems, and used for reference in real-time scheduling research with uncertain characteristic or limited priority level.
出处
《计算机学报》
EI
CSCD
北大核心
2004年第6期812-818,共7页
Chinese Journal of Computers
基金
国家自然科学基金 ( 60 3 740 5 8
60 3 73 0 5 5 )
国家"八六三"高技术研究发展计划基金 ( 2 0 0 1AA413 0 2 0 )资助
