Multilevel Feedback

Queue: Efficient Scheduling

and implementation
GROUP MEMBERS
KHURUM IQBAL
RAB NAWAZ
UBAIR BIN SULTAN
Introduction:

 In CPU scheduling various algorithms exist like

FCFS (First come first serve), SJF (Shortest job first), SRTF
(Shortest remaining time first), Priority Scheduling, Round Robin
(RR), MLQ (Multilevel queue), MLFQ (Multilevel feedback
queue) scheduling.

 Multilevel Feedback Queue (MLFQ)

algorithm allows the switching of processes between queues
depending on their burst time.
Problems:

 The definition of multilevel feedback-queue makes it the most general CPU

Scheduling Algorithm.
 Unfortunately, it is also the most complex algorithm, since defining the best
algorithm require some means by which select values for all the
parameters.
 Parameters: scheduling algorithm, when to upgrade or demote process
from one queue to another, which queue a process will enter when a
process need services.
Related work:

 A number of approaches have been made and researches

have been carried and cited in the field of process scheduling
using Multi level feedback-queue by having different scheduling
algorithms on different queues of MFQ.
 To improve the efficiency some have used concept of applying SJF prior to
RR by having static time quantum. But in this case starvation is possible for
some of the processes having larger burst time. So lower level queues kept
starving for CPU and resources.
 Proposed Algorithm:

Dynamic Time Quantum

Proposed algorithm (Cont..)
Proposed Algorithm(Cont…)

Static queue and Dynamic Queue)

Observation:

 So in this case average

waiting time and average
turn around time reduce as:
Conclusion:

 From all the above experiments it can be clearly

concluded that dynamically generating time quantum by the
method explained above and using SJF before RR helps us in
improving the CPU and resource utilization.

 By experimenting with various combinations of jobs and scheduling

policies, we have preferred Round Robin algorithm for fair use of CPU in
first queue and SJF prior to RR in next queues (which we called as SJFRR) to
reduce the average waiting time as well as turned around time by having
dynamic time quantum.
Thank you