Improving the performance of fair share scheduling algorithm using fuzzy logic

In Fair Share scheduling shares of CPU time for each group of processes is fixed and then accordingly processes from each group are allocated CPU for their execution. We have proposed a FIS using which CPU share of group of Processes will be decided dynamically. FIS has also been constructed for scheduling within the group to minimize the overheads. For scheduling within the group Fair Share Scheduling algorithm uses Round Robin scheduling. If the process do not complete its execution within the time quantum, the process is put at the end of ready queue and process switch occurs where state of the running process is put onto stack and the state of the next process is taken from the stack and its execution is resumed. If the time required for the running process is slightly more than time quantum by a fraction, even then process is preempted and context switch occurs. This results into more waiting time for that process and more overheads due to unnecessary context switch. In this paper an algorithm using fuzzy logic has also been proposed that tries to remove these two problems.