Application of fuzzy logic to real-time scheduling

Although a lot of research has been done in the sphere of real-time systems and scheduling algorithms for real-time tasks, implementation of these real-time tasks still poses a lot of problems due to the unpredictability of the tasks involved and due to the lack of complete task knowledge prior to the execution process. In this paper, presented is a fuzzy logic controlled dynamic real-time scheduling algorithm to try and achieve better overall system performance and utilization for systems dealing with soft real-time tasks, where fuzzy logic uses the system's deadline miss ratio (DMR) and system throughput to calculate a value which provides a system performance metric used to drive the system to the desired level of performance by automatically adjusting to a predetermined set point.