A Task Scheduling Algorithm for Minimizing Worst- Case Execution Time of Real-Time Embedded Systems

Modern real-time embedded systems always consist of a set of tasks which have strict timing constrains. Thus, a feasible task scheduling to approach is required to satisfy timing to constrain and minimize the total execution time of the tasks on an embedded system. In this paper, we propose a novel approach by integrating cache locking and task scheduling for minimizing Worst-Case Execution Time (WCET) of a real-time embedded system. The task model considered in this paper is preemptive multi-tasks with corresponding release times and deadlines. Our approach can assign an appropriate space of cache to each task, and achieve a feasible schedule. We implement our approach and compare with non-preemptive No-cache method and static analysis method (STA) by using SNU real-time benchmark suit. The experimental results show that our approach can achieve better performance.