Abstract
In real-time systems, scheduling algorithms are used in control situations where it is a crucial or a critical to complete a task successfully within a specific time interval. Many scheduling techniques consider scheduling tasks according to their Worst-Case Execution Time (WCET) or average execution time while neglecting a change in their probability distributions. In real-time applications such as multimedia, Using either WCET or the average value to schedule several tasks is impractical and inappropriate and could cause a catastrophic result. The previous studies show that the multimedia real-time applications such as Audio or Video statistically has a great variation in their execution times which means scheduling them according to the WCET or the average execution time is insufficient and unwanted results may occur. In this paper, a new effective and efficient dynamic method to schedule periodic real-time tasks is presented based on using a dynamic moving average approach. Dynamic moving average refers to a change in a probability distribution being used when a task is added or removed. The objective is to develop a method that guarantees the delivering of all tasks to meet their timing constraints and also to minimize the overhead occurring from context switching between different tasks. Furthermore, enhancing the response time minimization is desired. Our intensive experiments on a developed simulation performance evaluation indicate that the developed method is capable of handling all tasks to meet their deadline times, achieving around an average of 24% to 49% reduction in the overhead and the response time enhancing by average of 50%.