强实时环境下调度非周期任务的时限寻优方法

Deadline Optimizing Approach for Scheduling Aperiodic Tasks in Hard Real-Time Environment

文章提出了强实时环境下调度弱时限非周期任务的时限寻优方法（ＤＯＡ），该方法在保证周期任务和偶发性任务满足时限要求的前提下，使非周期任务的响应时间达到最优。它还可根据实时应用的需要对算法的执行性能和计算复杂度进行折衷调整。仿真实验表明，ＤＯＡ与现有的动态调度算法相比，使非周期任务响应时间更短，同时它收敛快，额外开销小，计算复杂度低，实现方便，因此是强实时环境下对周期任务与非周期任务进行混合调度的一种较好的方法。

This paper proposes Deadline Optimizing Approach (DOA)for scheduling soft aperiodic tasks in hard realtime environment. The approach optimizes the response time of aperiodic task,still guaranteeing that periodic tasks and sporadic tasks meet their timing constraints. DOA allows a designer to compromise between performance and complexity according to real-time applications requirements. The simulation experiments have shown that DOA makes aperiodic response time shorter than that scheduled under existing dynamic algorithm. Moreover,it is simple to implement and converges rapidly,it has little overhead and low computing complexity,therefore DOA is a good approach for hybrid scheduling of periodic tasks and aperiodic tasks in hard real-time environment.