An Improved Round Robin CPU Scheduling Algorithm with Varying Time Quantum

International Journal of Computer Science, Engineering and Applications (IJCSEA) Vol.4, No.4, August 2014
DOI : 10.5121/ijcsea.2014.4401 1
AN IMPROVED ROUND ROBIN CPU SCHEDULING
ALGORITHM WITH VARYING TIME QUANTUM
Manish Kumar Mishra1
and Dr. Faizur Rashid2
1
Department of Information Technology, Haramaya University, Dire Dawa, Ethiopia
2
Department of Computer Science, Haramaya University, Dire Dawa, Ethiopia
ABSTRACT
Process management is one of the important tasks performed by the operating system. The performance of
the system depends on the CPU scheduling algorithms. The main aim of the CPU scheduling algorithms is
to minimize waiting time, turnaround time, response time and context switching and maximizing CPU
utilization. First-Come-First-Served (FCFS) Round Robin (RR), Shortest Job First (SJF) and, Priority
Scheduling are some popular CPU scheduling algorithms. In time shared systems, Round Robin CPU
scheduling is the preferred choice. In Round Robin CPU scheduling, performance of the system depends on
the choice of the optimal time quantum. This paper presents an improved Round Robin CPU scheduling
algorithm coined enhancing CPU performance using the features of Shortest Job First and Round Robin
scheduling with varying time quantum. The proposed algorithm is experimentally proven better than
conventional RR. The simulation results show that the waiting time and turnaround time have been reduced
in the proposed algorithm compared to traditional RR.
KEYWORDS
CPU Scheduling, Round Robin Scheduling, Improved Round Robin Scheduling, Burst Time, Waiting Time,
Turnaround Time
1. INTRODUCTION
Improper use of CPU can reduce the efficiency of the system in multiprogramming computing
systems. In multiprogramming systems, multiple processes are being kept in memory for
maximum utilization of CPU [1]. CPU utilization can be maximized by switching CPU among
waiting processes in the memory and running some process all the time [2]. The main aim of the
CPU scheduling algorithms is to minimize waiting time, turnaround time, response time and
context switching and maximizing CPU utilization. This study focuses on improving the
effectiveness of Round Robin CPU scheduling algorithm.
1.1. Performance parameters
The processes that have been submitted to the system and waiting for the processor time are put
in a queue called ready queue. The CPU should be busy as much as possible to use it effectively.
Whenever CPU becomes idle, a waiting process from the ready queue is selected and CPU
allocated to that process [2]. The time for which a process uses the CPU is known as burst time.
Arrival Time is the time at which a process joins the ready queue. The total time taken by a
process from the time of submission to the time of completion of the process is the turnaround
time. The total time spent by the process waiting in the ready queue is termed as waiting time of

2
the process. The CPU scheduling algorithms focus on reducing the waiting time, turnaround time
and context switches by scheduling the processes from the ready in an effective manner.
1.2. CPU Scheduling Algorithms
CPU scheduling algorithms are used to allocate the CPU to the processes waiting in the ready
queue. First-Come-First-Served (FCFS) Round Robin (RR), Shortest Job First (SJF) and, Priority
Scheduling are some popular CPU scheduling algorithms. In FCFS CPU scheduling algorithm,
the process that arrives first in the ready queue is served first. The average waiting time in this
scheduling is quite long [2]. In SJF CPU scheduling algorithm, the process with shortest CPU
burst time executes first from the ready queue. In SJF average waiting time decreases. CPU is
allocated to the processes based on their priority in Priority scheduling algorithm. The process
with highest priority gets executed first and then the second highest and so on. Each process from
the ready queue is given a fixed time quantum in RR CPU scheduling algorithm. In this paper we
have proposed a new algorithm that uses the features of SJF and RR with varying time quantum
to reduce the waiting time and turnaround time.
2. RELATED WORK
In the past years, several CPU scheduling algorithms have been introduced for improving the
system performance. A fixed time quantum is allocated to the process waiting in the ready queue,
only in first cycle and then SJF is used to select next process in An Improved Round Robin
Scheduling Algorithm for CPU Scheduling [1]. Time quantum is continuously adjusted according
to the burst time of the processes in Self-Adjustment Time Quantum in Round Robin Algorithm
[3]. The job mix order for the algorithm in [3] is used in Dynamic Quantum with Readjusted
Round Robin Scheduling Algorithm [4]. Robust quantum time value has been proposed in [5]
after arranging the process in the ascending order and taking the average of minimum and
maximum burst time of the processes in the ready queue. A new weighting technique is
introduced for CPU Schedulers in Burst Round Robin (BRR) [6]. In this approach, the processes
having smaller CPU time are given more time, so that these processes can be cleared from the
ready queue in a short time span. Debashree Nayak et. al. [7] did the similar work as [3] [4]. The
optimal time quantum is assigned to each process after every cycle of execution. Optimal time
quantum is the average of highest CPU burst time and the median. In [8] a new CPU scheduling
algorithm is presented. In this approach, the running processes are scheduled based on three
parameters of CPU. These parameters are burst time, I/O service time, and priority of processes.
A new fare-share scheduling with weighted time slice [9] assigns a weight to each process and the
process having the least burst time is assigned the largest weight. The time quantum is calculated
dynamically, using weighted time slice method and then the processes are executed. Algorithm in
[10] calculates the original time slice suited to the burst time of each processes and then dynamic
ITS (Intelligent Time Slice) is found out in conjunction with the SRTN algorithm [2]. The
scheduling algorithm proposed in [11] uses two processors, one processor is dedicated to execute
CPU-intensive processes only and the other processor is dedicated to executed I/O-intensive
process. This gives better result in a two processor environment than [4]. A New Round Robin
based Scheduling Algorithm for Operating Systems: Dynamic Quantum Using the Mean
Average [12] introduce a new concept of calculating the average burst time of the processes from
the ready queue after every cycle and allocated as dynamic time quantum. The processes are
scheduled for execution by giving importance to both the user priority and shortest burst time
priority in Fair Priority Round Robin with Dynamic Time Quantum [13]. In this approach, the
individual time quantum for each process is decided by a factor based on both user priority and
the burst time priority. The CPU is allocated to the first process from the ready queue for a time
interval of up to one time quantum in An Improved Round Robin (IRR) CPU Scheduling

3
Algorithm [14]. After completion of process’s time quantum, the remaining CPU burst time of
the currently running process is compared with the time quantum. If the burst time of the
currently running process is less than one time quantum, the CPU is again allocated to the same
process so that it can finish execution and removed from the queue. It reduces the waiting time of
the process in the ready queue and hence improve the performance. An Additional Improvement
in Round Robin (AAIRR) CPU Scheduling Algorithm [15] proposed an improvement in the
conventional RR and IRR [14]. This approach is similar to the IRR but it picks the process from
the ready queue whose remaining burst time is shortest. The algorithm proposed in [16] is again
an improvement of IRR [14] which uses two queues, ARRIVE queue and REQUEST queue. It is
giving significance performance improvement compared to IRR.
3. IRRVQ CPU SCHEDULING ALGORITHM
The improved Round Robin CPU scheduling algorithm with varying time quantum (IRRVQ)
combines the features of SJF and RR scheduling algorithms with varying time quantum. Initially
the processes in the ready queue are arranged in the ascending order of their remaining burst time.
CPU is allocated to the processes using RR scheduling with time quantum value equal to the burst
time of first process in the ready queue. After each cycle processes in the ready queue are
arranged in the ascending order of their remaining burst time and CPU is allocated to the
processes using RR scheduling with time quantum value equal to the burst time of first process in
the ready queue.
Following is the proposed IRRVQ CPU scheduling algorithm
1. Make a ready queue RQUEUE of the Processes submitted for execution.
2. DO steps 3 to 9 WHILE queue RQUEUE becomes empty.
3. Arrange the processes in the ready queue REQUEST in the ascending order of their
remaining burst time.
4. Set the time quantum value equal to the burst time of first process in the ready queue
RQUEUE.
5. Pick the first process from the ready queue RQUEUE and allocate CPU to this
process for a time interval of up to 1 time quantum.
6. Remove the currently running process from the ready queue RQUEUE, since it has
finished execution and the remaining burst time is zero.
7. REPEAT steps 8 and 9 UNTIL all processes in the ready queue gets the CPU time
interval up to 1 time quantum.
8. Pick the next process from the ready queue RQUEUE, and allocate CPU for a time
interval of up to 1 time quantum.
9. IF the currently running process has finished execution and the remaining CPU burst
time of the currently running process is zero, remove it from the ready queue ELSE
remove the currently running process from the ready queue RQUEUE and put it at
the tail of the ready queue.
3.1. Illustration
A ready queue with four processes P1, P2, P3 and P4 has been considered for illustration purpose.
The processes are arriving at time 0 with burst time 12, 8, 21 and 15 respectively. The processes
P1, P2, P3 and P4 are arranged in the ascending order of their burst time in the ready queue which
gives the sequence P2, P1, P4 and P3. The time quantum value is set equal to the burst time of
first process in the ready queue i.e. 8. CPU is allocated to the processes P2, P1, P4 and P3 from
the ready queue for a time quantum of 8 milliseconds (ms). After first cycle, the remaining burst

4
time for P2, P1, P4 and P3 are 0, 4, 7 and 13 respectively. The process P2 has finished execution,
so it is removed from the ready queue. The processes P1, P3 and P4 are arranged in the ascending
order of their remaining burst time in the ready queue which gives the sequence P1, P4 and P3.
The time quantum value is set equal to the burst time of first process in the ready queue i.e. 4.
CPU is allocated to the processes P1, P4 and P3 from the ready queue for a time quantum of 4
ms. After second cycle, the remaining burst time for P1, P4 and P3 are 0, 3 and 9 respectively.
The process P1 has finished execution, so it is removed from the ready queue. The processes P3
and P4 are arranged in the ascending order of their remaining burst time in the ready queue which
gives the sequence P4 and P3. The time quantum value is set equal to the burst time of first
process in the ready queue i.e. 3. CPU is allocated to the processes P4 and P3 from the ready
queue for a time quantum of 3 ms. After third cycle, the remaining burst time for P4 and P3 are 0
and 6 respectively. The process P4 has finished execution, so it is removed from the ready queue.
Now only process P3 is in the ready queue, so CPU is allocated to P3 for a time quantum of 6 ms
which is the burst time of first process in the ready queue. The waiting time is 24 ms for P1, 0 ms
for P2, 35 ms for P3 and 32 ms for P4. The average waiting time is 22.75 ms. With the same set
of process with same arrival and CPU burst times, the average waiting time is 25.75 ms for time
quantum 8, 27.75 ms for time quantum 4, 29.25 ms for time quantum 3 and 28.75 ms for time
quantum 6 in RR. The average turnaround time is 41.75 in IRRVQ while average turnaround time
is 39.75 for time quantum 8, 41.75 for time quantum 4, 43.25 for time quantum 3 and 42.75 for
time quantum 6 in RR.
4. EXPERIMENTAL ANALYSIS
4.1. Assumptions
For performance evaluation, it has been assumed that all the processes are having equal priority in
a single processor environment. The number of processes and their burst time are known before
submitting the processes for the execution. The context switching overhead incurred in switching
from one process to another has been considered zero. The overhead of arranging the ready queue
processes in ascending order has also been considered zero. All processes are CPU bound. No
processes are I/O bound. The time quantum is taken in milliseconds (ms).
4.2. Experiments Performed
Two different cases have been taken for performance evaluation of our proposed IRRVQ
algorithm. In the case 1, CPU burst time is in random orders and processes arrival time is
assumed zero. In the case 2, CPU burst time is in random orders and processes arrival time is
assumed non zero. The CPU burst time in ascending or descending orders have not been
considered since it gives the same result as the CPU burst time in random orders.
4.2.1. CASE 1 – Zero Arrival Time
In this case, CPU burst time is in random orders and processes arrival time is assumed zero. A
ready queue with five processes P1, P2, P3, P4, and P5 has been considered as shown in table 1.

5
Table 1. Processes with their arrival and burst time (Case 1).
Process Arrival Time Burst Time
P1 0 15
P2 0 32
P3 0 10
P4 0 26
P5 0 20
The comparison result of RR and proposed IRRVQ is shown in Table 2. Figure 1, Figure 2 and
Figure 3 show the Gantt chart representation of RR with time quantum 10, 5 and 6 respectively.
Figure 4 shows the Gantt chart representation of IRRVQ.
Table 2. Comparison of RR and IRRVQ.
Algorithm Time Quantum
(TQ)
Average Waiting
Time (ms)
Average Turnaround
Time (ms)
RR 10, 5, 6 54.2, 56.2, 60.2 74.8, 76.8, 80.8
IRRVQ 10, 5, 5, 6, 6 46.2 66.8
Figure 1. Gantt chart representation of RR with TQ = 10

6
Figure 4. Gantt chart representation of IRRVQ
4.2.2. CASE 2 – Non - Zero Arrival Time
In this case, CPU burst time is in random orders and processes arrival time is assumed non zero.
A ready queue with five processes P1, P2, P3, P4, and P5 has been considered as shown in table
3.
Table 3. Processes with their arrival and burst time (Case 2).
Process Arrival Time Burst Time
P1 0 7
P2 4 25
P3 10 5
P4 15 36
P5 17 18
The comparison result of RR and proposed IRRVQ is shown in Table 4. Figure 5 and Figure 6
show the Gantt chart representation of RR with time quantum 7 and 11 respectively. Figure 7
shows the Gantt chart representation of IRRVQ.
Table 4. Comparison of RR and IRRVQ
Algorithm Time Quantum
(TQ)
Average Waiting
Time (ms)
Average Turnaround
Time (ms)
RR 7, 11 25.6, 27.0 43.8, 54.4
IRRVQ 7, 11, 7, 11 19.4 37.6

7
Figure 7. Gantt chart representation of IRRVQ
5. CONCLUSIONS
One of the important tasks of the operating system is the allocation of CPU to the processes
waiting for execution. Many CPU scheduling algorithms have been presented with some
advantages and disadvantages. An improved round robin CPU scheduling algorithm with varying
time quantum proposed in this paper giving better performance than conventional RR algorithm.
The waiting time and turnaround time have been reduced in the proposed IRRVQ scheduling
algorithm and hence the system performance has been improved. Simulation results also prove
the correctness of the theoretical results. The proposed algorithm can be integrated to improve the
performance of the systems.
REFERENCES
[1] Rakesh Kumar Yadav, Abhishek K Mishra, Navin Prakash & Himanshu Sharma, (2010) “An
Improved Round Robin Schedduling Algorithm for CPU Scheduling”, International Journal on
Computer Science and Engineering, Vol. 2, No. 4, pp 1064-1066.
[2] A. Silberschatz, P. B. Galvin & G. Gagne, (2005) Operating System Concepts, John Wiley and Sons
Inc, pp 157-167.
[3] Rami J. Matarneh, (2009) “Self-Adjustment Time Quantum in Round Robin Algorithm Depending on
Burst Time of Now Running Processes”, American J. of Applied Sciences, Vol. 6, No. 10, pp1831-
1837.
[4] H.S. Behera, R. Mohanty & Debashree Nayak, (2010) “A New Proposed Dynamic Quantum with Re-
Adjusted Round Robin Scheduling Algorithm and Its Performance Analysis”, International Journal of
Computer Applications, Vol. 5, No. 5, pp 10-15.
[5] M Lavanya & S. Saravanan, (2013) “Robust Quantum Based Low-power Switching Technique to
improve System Performance”, International Journal of Engineering and Technology, Vol. 5, No. 4,
pp 3634-3638.
[6] Helmy, T. & A. Dekdouk, (2007) “Burst Round Robin as a Proportional-share Scheduling
Algorithm”, IEEEGCC, http://eprints.kfupm.edu.sa/1462/.
[7] Debashree Nayak, Sanjeev Kumar Malla & Debashree Debadarshini, (2012) “Improved Round Robin
Scheduling using Dynamic Time Quantum”, International Journal of Computer Applications, Vol. 38,
No. 5, pp 34-38.

8
[8] Mehdi Neshat, Mehdi Sargolzaei, Adel Najaran & Ali Adeli, (2012) “The New method of Adaptive
CPU Scheduling using Fonseca and Fleming’s Genetic Algorithm”, Journal of Theoretical and
Applied Information Technology, Vol. 37, No. 1, pp 1-16.
[9] H.S. Behera, Rakesh Mohanty, Jajnaseni Panda, Dipanwita Thakur & Subasini Sahoo, (2011)
“Experimental Analysis of a New Fare-Share Scheduling Algorithm with Waited Time Slice for Real
Time Systems”, Journal of Global Research in Computer Science, Vol. 2, No. 2, pp 54-60.
[10] H.S. Behera, Simpi Patel & Bijaylaxmi Panda, (2011) “A new dynamic Round-robin and SRTN
algorithm using variable original time slice and dynamic intelligent time slice for soft real time
system”. International Journal of Computer Applications, Vol. 2, No.1, pp 54-60.
[11] H.S. Behera, Jajnaseni Panda, Dipanwita Thakur & Subasini Sahoo, (2011) “A New Proposed Two
Processor Based CPU Scheduling Algorithm with Varying Time quantum for Real Time Systems”,
Journal of Global Research in Computer Science, Vol. 2, No. 4, pp 81-87.
[12] Abbas Noon, Ali Kalakech, and Seifedine Kadry, (2011) “A New Round Robin based Scheduling
Algorithm for Operating Systems: Dynamic Quantum Using the Mean Average”, International
Journal of Computer Science Issues, Vol. 8, Issue 3, No. 1, pp 224-229.
[13] M.K. Srivastav, Sanjay Pandey, Indresh Gahoi & Neelesh Kumar Namdev, (2012) “Fair Priority
Round Robin with Dynamic Time Quantum”, International Journal of Modern Engineering Research,
Vol. 2, Issue 3, pp 876-881.
[14] Manish Kumar Mishra & Abdul Kadir Khan, (2012) “An Improved Round Robin CPU Scheduling
Algorithm”, Journal of Global Research in Computer Science, Vol. 3, No. 6, pp 64-69.
[15] Abdulrazak Abdulrahim, Salisu Aliyu, Ahmad M Mustapha & Saleh E. Abdullahi, (2014) “An
Additional Improvement in Round Robin (AAIRR) CPU Scheduling Algorithm”, International
Journal of Advanced Research in Computer Science and Software Engineering, Vol. 4, Issue 2, pp
601-610.
[16] Abdulrazak Abdulrahim, Saleh E. Abdullahi & Junaidu B. Sahalu, (2014) “A New Improved Round
Robin (NIRR) CPU Scheduling Algorithm”, International Journal of Computer Applications, Vol. 90,
No. 4, pp 27-33.

An Improved Round Robin CPU Scheduling Algorithm with Varying Time Quantum

More Related Content

What's hot (16)

Similar to An Improved Round Robin CPU Scheduling Algorithm with Varying Time Quantum (20)

Recently uploaded (20)

An Improved Round Robin CPU Scheduling Algorithm with Varying Time Quantum