Fuzzy based multilevel queue scheduling algorithm

Many scheduling techniques has been designed for scheduling of processes in the multiprogramming system, one of these techniques is multilevel queue CPU scheduling technique where the ready queue is divided into multiple sub-queues. Processes are assigned to different sub queues depending on their classification. In traditional multilevel queue scheduling approach each queue is assigned a priority level. Processes from the higher level queues always gets priority over the processes in lower level queues which can cause starvation for processes in lower queues. In this paper we propose a new fuzzy based multilevel queue CPU scheduling algorithm. In our proposed work ready queue is divided into two sub-queues which contain I/O bound and CPU bound processes. CPU time is allocated dynamically to each queue. We have designed two fuzzy inference systems. One is to dynamically allocate the CPU time to two queues. Another FIS is designed to calculate the time quantum to schedule the I/O bound processes. Our proposed work improves the starvation problem and the average response time.