Dynamic Optimal Control for Aperiodic Soft Real-Time Systems

This paper studies the aperiodic soft Real-Time Systems (RTS) with random task arrival and unpredictable task execution time, and proposes a dynamic optimal control model based on the queuing theory. Firstly, we construct a dynamic optimal control model for the FCFS (First Come First Served) scheduling strategy; then, we present a more useful model for DM (Deadline Monotonic) scheduling algorithm. This model can statistically guarantee the schedulability of soft RTS, at the same time it can guarantee the QoS (Quality of Service) of incoming tasks and improve the system throughput. Finally, an example shows that the model is capable of improving the systems performance and it is an efficient workload management method.