A M-level parallel task scheduling

Task scheduling is the challenging issues of Parallel Processing. Optimal scheduling can reduce computation time of the processor. In scheduling, tasks are not independent of each other. So, Directed Acyclic Graph is used to show the interdependency among tasks. In this paper, we introduce a new scheduling measurement called Middle-level (m-level) that improves the performance of scheduling. Our algorithm has two step approaches: m-level and sufferage value. It reduces the idle time and efficient utilization of the processor time in comparison to other scheduling algorithms. It gives optimal or close to optimal solution. Communication cost is considered to be a part of real time systems.