中文摘要：

面向基于电池供电的嵌入式实时系统,设计效用获取的实时节能调度算法.在实时系统的研究中,通常采用同步机制来实现任务对资源的互斥共享访问.在这种情况下,调度算法旨在能量受限下获得最大的系统效用,同时满足实时任务的可调度及任务同步.提出了两步节能算法（two-step energy-efficient algorithm,TSEEA）,此算法包含两部分：静态算法和动态算法.静态算法部分在离线状态下,获得保守情况下任务的运行速度,并保证实时可调度及任务同步;动态算法部分在系统的运行过程中,综合考虑系统的静态信息、性能需求以及任务的动态行为,释放并回收空闲时间,有效调整任务运行速度,以求在有限能量供应前提下获取更大的系统效用.与其他有效算法相比,提出的算法保证只要供能充足,所有的任务均满足实时性要求,因此本算法可充分利用有限能量更多的效用,另外降低了时间复杂度.实验结果表明本算法在效用获取方面优于现有的其他算法.

英文摘要：

This paper presents an energy-efficient scheduling algorithm with accrual utility in battery-powered embedded real-time systems.The real-time tasks considered here synchronize to access the shared resources in a mutually exclusive manner.Under these constraints,the goal of a scheduling algorithm is to yield more utility within a supply of limited energy,while satisfying the timeliness and task synchronization requirements.We propose a two-step energy-efficient algorithm（TSEEA）,which consists of two phases： a static algorithm,which runs offline and achieves the static execution speeds of all tasks under conservative conditions;and a dynamic one,which strives to releasereclaim slack times and to effectively tune the executing speeds of candidate task in a timely manner by synthesizing the static information and dynamic behaviors of performance demands at runtime.Compared with other algorithms,it is guaranteed that any task can meet its deadline constraint by our approach if the system energy supply is sufficient.Consequently,our algorithm can fully exploit the limited energy supply while yielding a high utility to the system.Further,our algorithm tries to reduce the time complexity.The experiments validate our analytical results and demonstrate that the proposed algorithm outperforms other existing algorithms in terms of accrual utility.