一种实时系统中的多任务可预测调度算法

An Algorithm for Multitask Predictable Scheduling in Real-time Systems

通过研究实时系统中可能出现的任务类型,提出了一种实时系统任务模型和基于这个模型的多任务调度算法,分析了该算法下所有周期任务和非周期任务可调度的条件。该算法结合了DMS(Deadline Monotonic scheduling)和DD(Deadline Driven)算法,在保证系统中周期任务执行的前提下,也使非周期任务获得了较好的响应时间。

By studying task types of real-time system, a task model for real-time system and a multitask scheduling algorithm based on this model are proposed. And the conditions that guarantee periodic and aperiodic tasks to be schedulable are given. The algorithm adopts the DMS and the DD tactics, shortens the response time of aperiodic tasks and still guarantees that periodic tasks meet their timing constraints.