固定优先级混合关键偶发任务能耗感知算法

Fixed Priority Mixed-Criticality Sporadic Tasks Energy-Aware Algorithm

混合关键系统是将不同关键层次的应用或组件集成到同一个共享平台.由于受尺寸、重量与体积的限制,能耗对于混合关键系统而言尤其重要.能耗感知调度算法是解决混合关键系统能耗问题的关键,现有的能耗感知算法主要基于动态优先级策略且空闲时间利用率低.针对固定优先级混合关键系统偶发任务能耗感知问题,提出节能效果更好的固定优先级混合关键调度(fixed priority mixed criticality schedule,FPMCS)算法.首先,提出关键层次单调速率策略(criticality rate monotonic scheme,CRMS)调度混合关键偶发任务,分析该策略的调度可行性,且计算出能耗感知速度.其次,利用高关键层次任务预留的空闲时间,通过事件触发的方法动态更新混合关键偶发任务集的利用率来回收偶发任务到达时间不确定产生的空闲时间.再次,利用混合关键偶发任务集的利用率决定任务的执行速度以达到降低能耗的目的.最后,通过理论分析和实验验证FPMCS算法是可行的;仿真实验表明:所提出的FPMCS算法比现有的方法可以节约大约33.21%的能耗.

Mixed-criticality systems integrate different criticality levels applications and components into a common shared platform.Energy consumption is very important for mixed-criticality systems due to size,weight and volume constraints.Energy-aware scheduling algorithm is the effective method to solve the energy consumption problem of mixed-criticality systems.Existing energy-aware algorithms based on dynamic priority schemes have lower slack time utilization.Fixed priority mixed-criticality scheduling(FPMCS)algorithm is proposed to solve the energy consumption problem of mixed-criticality systems.Firstly,a criticality rate monotonic scheme(CRMS)is proposed to schedule mixed-criticality sporadic tasks.In addition,the scheduling feasibility of CRMS is analyzed and the energy-aware speed is computed.Secondly,the slack time reserved for higher criticality level task is used to re-compute the utilization of higher criticality level task.The dynamically update the utilization of mixed-criticality sporadic tasks method through the method of event triggering is proposed to reclaim slack time generated from the random arrival of sporadic tasks.Thirdly,the speed of tasks is determined by the mixed-criticality sporadic tasks utilization to save energy.Finally,FPMCS algorithm is verified to be feasible by theoretical analysis and experiments.The experimental results show that the proposed FPMCS algorithm can save about 33.21%of energy consumption than existing algorithms.