嵌入式系统动态策略任务调度算法

Dynamic strategy task scheduling algorithms on embedded system

针对嵌入式实时系统任务调度问题,讨论综合截止期和关键度两种特征参数的任务调度策略。引入动态策略的概念,采用任务分组优先级分段的方式,通过判断函数的判断动态选择调度策略,利用可达截止期优先降低任务错失率,关键保证重要任务的调度成功率,提出一种动态截止期-关键度调度算法。理论分析和仿真实验表明,在非过载的情况下该算法可以达到与可达截止期优先相当的调度性能;在过载的情况下,相比截止期优先和固定优先级算法具有更好的调度性能,能够有效地保证关键实时任务的成功调度。

Aiming at embedded real-time system task scheduling problems, the scheduling strategy of integrating two characteristic parameters （deadlines and criticalness）, was studied. A dynamic deadlinecriticalness algorithm was presented. The algorithm does the following： groups tasks; partitions priorities; changes the scheduling strategy dynamically based on the judgment function; employs a strategy of reaching the earliest deadline first （EDF） to lower the task missed deadline percentage; and uses the criticalness to guarantee the task scheduling success ratio. Theoretical analysis and experimental simulation show that this algorithm achieves the same performance as EDF under non-overload conditions. The algorithm can improve the scheduling success ratio of critical tasks compared with the classical EDF and fixed priority algorithms under overload conditions.