Hard Real-Time Constraints in Implementing the Myopic Scheduling Algorithm

This paper presents an investigation into the constraints in implementing the Myopic scheduling algorithm. The Myopic algorithm is a hard real-time (RT) task scheduling algorithm for multiprocessor systems. It selects a suitable task based on a heuristic function, H from a subset (feasibility check window, K ) of all ready tasks. On the other hand, the original heuristic scheduling algorithm chooses the task with the least H value from all N ready tasks. Performance of the Myopic algorithm significantly depends on the chosen heuristic function and the size of the feasibility check window since it considers only K tasks from N tasks (where K ≤N ). This research investigates the impact of scheduling non-periodic hard RT tasks using the Myopic algorithm and evaluates the performance for different parameters to demonstrate the merits and constraints of the algorithm. The effects of the feasibility check window size, K , choice of heuristic function, H , the worst case processing time of tasks, p T on the performance of the Myopic algorithm under various loads are investigated. Finally, the performance of the algorithm is evaluated as task completion ratio, presented and discussed through a set of experiments.