Dynamic approach to minimize overhead and response time in scheduling periodic real-time tasks

In real-time systems, a task or a set of tasks needs to be executed and completed successfully within a predefined time. Those systems require a scheduling technique or a set of scheduling methods to distribute the given task or the set of tasks among different processors or on a processor. In this paper, a new novel scheduling approach to minimize the overhead from context switching between several periodic tasks is presented. This method speeds up a required response time while ensuring that all tasks meet their deadline times and there is no deadline miss occurred. It is a dynamic-priority technique that works either on a uniprocessor or several processors. In particular, it is proposed to be applied on multiprocessor environments since many applications run on several processors. Various examples are presented within this paper to demonstrate its optimality and efficiency. In addition, several comparison experiments with an earlier version of this approach were performed to demonstrate its efficiency and effectiveness too. Those experiments showed that this novel approach sped up the execution time from 15% to nearly around 46%. In addition, it proved that it reduced the number of a context switch between tasks from 12% to around 50% as shown from simulation tests. Furthermore, this approach delivered all tasks/jobs successfully and ensured there was no deadline miss happened.