A Randomized Algorithm To Schedule Real-Time Task Graphs To Satisfy A Multi-Criteria Objective Function

The objective of previous scheduling algorithms for real-time applications is to execute real-time tasks in a timely fashion. But these algorithms do not consider other performance metrics such as system reliability and inter-task communication. These factors may degrade any additional performance gained by using cluster computing. In this paper, we develop an approach to schedule real-time task graphs that need to satisfy multi-criteria simultaneously. We propose an objective function that guides the scheduling process to satisfy given real-time constraints, to maximize reliability and to minimize inter-communication costs. We use two different searching techniques, a heuristic search technique and a random search technique, that use the objective function. We simulate both techniques to study and compare their merits. Results show that the random search technique achieves better performance.