Scheduling Heterogeneous Multiprocessor Real-Time Systems with Mixed Sets of Task

Zhuo Cheng, Jinyun Xue, Haitao Zhang, Zhen You, Qimin Hu, Yuto Lim

Abstract: Real-time system is playing an important role in our society. Currently, almost all the practical real-time systems are equipped within multiple processors, compared to uniprocessor systems, scheduling on multiprocessor systems is far from wellstudied. This becomes unfortunate since current progress in developing heterogeneous multiprocessor systems is a long way ahead of research efforts to determine the best scheduling policies. Moreover, considering many real-time systems for industrial applications containing mixed sets of task including both time-and event-driven tasks, it endows further challenges for scheduling. To address such challenges, continuing our previous work, a Real-time scheduling methodology based on Satisfiability Modulo Theories (RSMT) is proposed. In RSMT, the problem of scheduling is treated as a satisfiability problem. The key work is to formalize the satisfiability problem using first-order logical formulas. After formalization, a desired schedule can be generated. To show the performance of RSMT, it is first applied to time-driven systems, heterogeneous multiprocessor systems have been considered. Then, in order to handle mixed sets of task, through a novel method of combining RSMT and online scheduling algorithm, RSMT also shows capability to design scheduling for such systems. Through a case study and simulation experiments, the usage of RSMT is shown.