基于离线利用率分析的混合关键级任务调度优化研究

现主流的混合关键级调度算法在系统高关键级状态下时主要通过抛弃低关键级任务来保证高关键级任务的执行，进而保证系统的正确性。此方法常常导致低关键级任务无法执行但系统资源却过剩的问题发生，故基于该问题提出复合型SDU（schedule depend on utilization）调度算法。该方法根据任务集对系统资源需求情况的不同进行利用率区间的划分，通过对各个区间实际使用情况的分析，设计相应的子算法进行调度，并提出了SDU算法对应的可调度性判据。仿真实验结果表明，相较于混合关键级任务调度领域主流的EDF-VD（earliest deadline first-virtual deadline）算法，所提SDU算法可将系统对任务集的调度率提升30%，并在相同情况下将系统对低关键级任务的执行率提升165%，证明了该算法可以极大地提高系统资源使用率，并保证系统服务完整性。

Research on scheduling optimization of mixed-criticality tasks based on offline utilization analysis

The current mainstream mixed-criticality scheduling algorithm guarantees the execution of high-criticality tasks by discarding low-critical tasks when the system is in a high-critical state, thereby ensuring the correctness of the system. Since this method often leads to the problem that low-critical tasks cannot execute, but the system resources are excessive. Based on this problem, this paper proposed a composite SDU scheduling algorithm which divided the utilization interval according to the resource requirements of the task set. According to the analysis of actual usage of each interval, it designed the corresponding algorithm for scheduling, and proposed the schedulability criterion corresponding to the SDU algorithm. The simulation results show that, comparing with the mainstream algorithm EDF-VD in the field of mixed-critical task scheduling, the SDU algorithm can improve the system''s scheduling rate by 30% and under the same conditions, the execution rate of low-critical tasks is increased by 165%, which greatly improves the utilization rate of system resources and ensures the integrity of system services.