Providing a multi-objective scheduling tasks by Using PSO algorithm for cost management in cloud computing

International Journal of Computer Applications Technology and Research
Volume 5– Issue 6, 368 - 375, 2016, ISSN:- 2319–8656
www.ijcat.com 368
Providing a multi-objective scheduling tasks by Using
PSO algorithm for cost management in cloud computing
Masrour Dowlatabadi,
Young Researchers and Elite
Club, Sabzevar
Branch, Islamic Azad University,
Sabzevar, Iran
Arman Kavosi Ghafi
Department of Computer
Software, Central Islamshahr
Branch, Islamic Azad
University, Iran
Rasoul Habibzadeh Shojaee,
Software, Central Islamshahr
University, Iran
Abdolreza Andalib
Software, Qeshm international
University, Qeshm, Iran
Abstract:
This article is intended to use the multi-PSO algorithm for scheduling tasks for cost management in cloud computing. This means that
any migration costs due to supply failure consider as a one objective and each task is a little particle and recognize by use of the
appropriate fitness schedule function (how the particles arrangement) that cost at least amount of total expense. In addition to, the weight
is granted to the each expenditure that reflects the importance of cost. The data which is used to simulate proposed method are series of
academic and research data that are prepared from the Internet and MATLAB software is used for simulation. We simulate two issues,
in the first issue, consider four task by four vehicles and divide tasks. In the second issue, make the issue more complicated and consider
six tasks by four vehicles. We write PSO's output for each two issues of various iterations. Finally, the particles dispersion and as well
as the output of the cost function were computed for each part.
Keywords: Cloud computing, Optimization particle algorithm, Apportion duties, MATLAB software

www.ijcat.com 369
1. INTRODUCTION
The basic concept of cloud computing dates back
to 1995, When John McCarthy stated: Cloud
computing may someday be organized as a public
industry [1]. Cloud computing, briefly Consist of
required computing, provide required fields and
presenting computer services by a cloud of
software, hardware and IT resources [2] [3].
Usually diagram of computer networks, the
figure of the cloud is used as a summery way to
hide the complex infrastructures that are inside
the network. Cloud word in the cloud computing
is extracted of these metaphor and is used to hide
the Internet and infrastructures in the cloud
computing [4]. By the cloud computing people
use the PC to achieve a network of computers that
are invisible status [5]. What makes this concept
a scalable and efficient is a Service-Oriented
Architecture (SOA) [6] [7]. Resources, are kept
on servers providers instead of to be installed on
a user's system.
By assume cloud computing, develop cloud
concept, in addition to servers cover
infrastructures of network. In mid-2008, Gartner
found that there was an opportunity in cloud
computing that was created for « Shaping the
relationship between the consumers of IT
services, between those who consume these
services and those who sell these services ». That
said, the cloud computing, will be the foundation
in the IT industry for the next 20 years [8].Cloud
computing is a new technology, which in recent
decades by various companies, most notably
Microsoft, Google and Amazon have been
released to the world. Cloud computing
technology is a great and safe storage space,
which is always different with processing and
computing power. In this regard, different
companies, offered different services with
various computing services [9]. These
calculations are based on TCP / IP (internet-
based) and include processors, massive
memories, quick data transfer networks and
architecture of reliable systems and without the
governing standard network protocols, cannot be
brought the entity to the technology [10].
Different types and levels of cloud services mean
it is important that, we are working with which
cloud computing systems [11]. A workflow, is an
automation of processes to work on the
documents, information, and tasks that passes
from one process to another and usually is
modeled in the form of a circular oriented graph
(DOG) [12]. An example of this oriented graph
can be seen in Figure 1. So that, each node shows
a task and each edge shows dependencies
between functional application tasks. Simply, the
Process of mapping tasks in a workflow
computing resources to perform tasks (To
maintain dependencies between them) is defined
workflows scheduling. The correct scheduling
can have a great effect on system performance.
The mapping of tasks to computing resources
belonging to the class of things that are well-
known NP-Complete [12] [13]. Useful
scheduling algorithm can meet user requirements
and improve resource efficiency. The technology
of cloud computing can be measured like
telephone and electricity companies and sharing
with customers [14]. Currently, Microsoft's
cloud-based service share with customers as
Microsoft Online service [15].Cloud computing
models resources, similar to a public industry
(electricity, telephone, etc.). As without the
preparation and storage of expensive tanks,
capacitors or transformers. Only when needed,
we use electricity and water and only pay based
on our consumption. In the case of software and
other resources, we just open the software (Eg,
tap) and only pay the amount that we consumed.
Without need to purchase and maintain expensive
resources for their implementation [16]. The
billing is based on the amount of minutes or hours
that tenants spent on a cloud system, transferred
data or filled storage space [17]. Sadri et al (1390)
in his study that called "secure message transfer
protocol in the cloud computing- based network".

www.ijcat.com 370
A brief description of a secure messaging system
design under internal network according to the
cloud computing model with utilizing Rayndl
algorithms [18]. Ashoorion (1390) studied
messages safety by computing in smart
platforms.
He indicated that mobile agents architecture has
overcome the many of limitations of architecture
on client and server [19]. Delavar Ghorbannia
and Arian (1390), provide a combined scheduling
method based on genetic algorithm in cloud
computing environment [20]. Matin Far (1391)
were studied methods, challenges and tools to
implement fault tolerance in cloud computing. In
this research, the fault tolerant techniques in
forms of cloud computing, fault tolerance
policies, challenges and the tools used to
implement fault tolerance in cloud computing
were discussed[21]. Maleki et al (1391) in his
study of the CloudSim’s simulation model and
it’s evaluation, introduced CloudSim’s
simulation tools, that, enables modeling and
simulation of cloud computing systems and
environments to the preparation of applications
[22]. Reza Zadeh (1391) in his study investigated
"the tolerability of a bug in the cloud computing".
In these calculations some bug’s tolerance
models were presented. In these models
redundancy has been used to improve the
reliability [11]. Shams and Ghiyasi (1392),
provided combined scheduling method based on
genetic algorithm in cloud computing
environment. In this study, they selected the
distributed scheduling reliable algorithm in cloud
computing (RSDC) due to its simplicity and has
been investigated the reliability on the chosen
algorithm [23]. Slvarany and Soodha (2010),
offered a fee-based algorithms for efficient
mapping of tasks to available resources in the
cloud. The scheduling algorithm also measures
the cost of using the resource and performance of
analysis [24]. Raphael et al. (2011) is to minimize
power consumption in data centers and improve
their weight balance at the same time. In this
regard, they stated a scheduling algorithm [25].
Moun and Tou (2011), provided an efficient
method for scheduling virtual machine to virtual
machine management. The proposed method has
been called EVMS. Based on EVMS, scheduler
system knows the current status of virtual
machines in the cloud [26]. Huang et al. (2012)
provided an efficient energy scheduling
algorithm (ESS) to reduce energy consumption.
The results show that, ESS has ability to reduce
energy consumption significantly by taking
SLA's [27]. Huh et al. (2012), provided a task
migration algorithm based virtual execution time
that limited the maximum difference in virtual
execution time between tasks. [28]. Peng et al.
(2011), delivered an energy-efficient scheduling
method based private cloud. In the paper
presented three schedule criteria proposed and
provided a combined of energy-efficient
scheduling method [29]. The results show that,
the efficiency of resources in the enhanced
algorithm improved [30].
2. Optimization Particles Algorithm
The swarm particles optimization algorithm is
one of the most important algorithms in the field
of swarm intelligence. The algorithm introduced
by Kennedy and Eberhart in 1995 and by
inspiration of the social behavior of animals such
as fish and birds live together in small and large
groups, have been designed. Other names of this
algorithm are the swarm particle algorithm and
birds algorithm. PSO, has similarities with
evolutionary computation techniques, like
genetic algorithms (GA). However, unlike GA,
PSO has not evolutionary functions, such as
transit and crossover. For successful
implementation of PSO, one of the key actions is,
discover solution of the problem within the PSO

www.ijcat.com 371
particles. Correctly affects its performance and its
possible implementation (Abraham et al., 2006).
Kennedy and Eberhart (1995), demonstrated that
PSO of each particle represents a possible answer
that moves randomly in problem space. Change
the location of each particle in the search space is
influenced by its knowledge and neighbors, so the
position of the swarm particle affects the way of
particle search. As a result of modeling this
behavior is the search process in which particles
move to the appropriate areas. Particles are taught
in groups, based on acquired knowledge, they go
to the best of their neighbors. PSO based on the
assumption that in every moment of every
particle in the search space adjusts its location
according to its best location, which has ever been
and the best place in the neighboring. The overall
design effective implement of PSO is difficult.
Also, check out the dynamic environment for
real-time applications such as fast convergence
feature PSO to follow the target variable
(minimum or maximum fitness function) are
necessary.
3. The proposed algorithm and simulation
If you want to run your program in parallel and
distributed in the cloud. We should know the
properties of the environment. As we become
more familiar with the characteristics of this
environment, we will be able to use it better and
more correct and avoid wasting resources.
Parallelization of a program can be done at
different levels. From the hardware to a single or
multiple systems or cluster. Cloud computing
architecture has a three-layer architecture which
in the infrastructure layers, we are faced with
resource management, the base layer is for
development of applicable programs on the cloud
resources and the software layer provided
applications for final users. If we want to improve
the management of resources, doing things and
scheduling, we must go into the infrastructure
layer. If we work on the development of a special
issue generally, we should go into the substrate
layer, and if we want to focus on how to provide
final service for users, we have compiled
software layer. Here we focused on the
substrate layer, because we are going to develop
a way to deal with cloud PSO algorithm. Our
Intended use is in the field of scheduling tasks
that can be as a final service is available to users
and run and optimization is done in the substrate
layer.
To be able to run the program in the field of PSO
algorithm, it is necessary to provide the substrate
layer for programs implement and development.
The proposed method evaluation have been
carried out by using two tests. we simulate PSO
algorithm for the cloud network with 4 and 4 tools
.Data by the network as follows:
Figure 1. The task dependency graph
Duty as a TDG (task dependency graph) are
shown in the following graph. The graph can
be converted to the tasks matrix.

www.ijcat.com 372
Table 1. Tasks Matrix
Task⋕1234
10001
20013
30100
41300
Table 4.2 Price matrix task dependencies
Data
Center⋕
1234
10124
21031
32302
44120
The cost function is defined as follows:
Minimize Total costs= f(x)= TM + TD + TO +TE
+TC
TC=∑ 𝑪𝒄 𝒋𝒊 + 𝑫𝒄𝒌𝒍
Where 1≤ j.i≤ T , 1≤ k, l≤ R
Fourth
machin
e
Third
machin
e
Second
machin
e
First
machin
e
2143
Wor
k
Cc, Dc are matrix elements tasks and dependent
costs.
In this part, we optimize the cost function by
using pso algorithm and the way of performance
of each task for each car is determined to have
the lowest cost. To convert this issue to a format
which PSO could run it, we shoul define the
limitation exactly to get correct results, So, in the
next step, at least each device have a work, not in
such way that many work are done by a device
and another device does not have any work. For
this part, we specify constraint for costs that is
acceptable answer which every works will be
divided equally between all equipment and
correct answers should be also determined in the
cost function.
The objective function described in this section.
The calculation of the objective function is
described below. Imagine we want to shift the
task of machine number 2 to machine number 3,
rows and columns of matrix hould be moved in a
way which Each task is assigned to the correct
machine. In this particular case, row 2 to row 3 is
replaced and column 2 and the third column are
obtained at the end of a diagonal matrix in this
case which confirms the correctness of
operations.
After this step, we calculate the cost for each
machine which is the product of two tasks and
dependent costs matrixes. We have used PSO for
aforementioned example, The results after 20
times as follows which final composition shows

www.ijcat.com 373
that the work assignement to the machine
according to the cost function is as follows:
Figure 2 PSO output for the first case
Figure 3 PSO output for the second case
Sixt
h
mac
hine
Fift
h
mac
hine
Four
th
mac
hine
Thir
d
mac
hine
Sec
ond
mac
hine
First
mac
hine
422143
W
or
k
There is a condition in this case that must be met
for the following:
Each machine can accept a maximum of two
works
All machines must have partners to carry out the
work.
The objective function can be expressed as
before, except that dependent costs matrix is 4 x
4, When it would be expressed in terms of
works, it needs to become a 6 x 6 matrix. For this
conversion, It is necessary to repeat a column of
the matrix twice and Or in other words, the rows
and columns of the matrix which allocated to a
machine that can do two things at the same time ,
is repeated twice and The general matrix is
obtained. In this case, the simulation was
conducted and the following results were
obtained. Pso which have been used exerted
initial population of 100 particles and we
determined the number of repetitions on PSO, the
amount of 30 was considered for repetitions.
4.Conclusion
According to the simulation results and adapting
to the reality of tasks division can be concluded
that PSO can properly diagnose the tasks
division in cloud computing according to cost
functions and graphs and for the division of tasks
provides a good answer.
5. Refrence
[1] S.Pandey,L.Wu,S.M.Guru,andR.Buyya, A
Particle Swarm Optimization-based Heuristic for
Scheduling Workflow Applications in Cloud
Computing Environments , peoc. On th IEEE
International Conference on Advanced
Information Networking and Applications, 2010 .

www.ijcat.com 374
[2] Junjie Peng; Zhou Lei; Wu Zhang, An
Energy-efficient Scheduling Approach Based on
Private Clouds, Journal of Information &
Computational Science 8: 4 (2011).
[3] M. Damshenas, et al., "Forensics
investigation challenges in cloud computing
environments," in Cyber Security, Cyber Warfare
and Digital Forensic (CyberSec), 2012
International Conference on, 2012, pp. 190-194.
[4] Y. Song, H. Wang, Y. Li, B. Feng, Y. Sun,"
Multi-Tiered On-Demand resource scheduling
for VM-Based data center", in: Proceedings of the
9th IEEE/ACM International Symposium on
Cluster Computing and the Grid , Shanghai,
China,pp. 148–155 Object Management
Group,2009
[5] M. Armbrust, A. Fox, R. Griffith, A.D.
Joseph, R. Katz, A. Konwinski,"A view of cloud
computing, Commun". ACM New York, USA,pp
50–58,2009.
[6] G. Dhiman, K. Mihic, T. Rosing, "A system
for online power prediction in virtualized
environments using gaussian mixture models",
in: Proceedings of the 47th ACM/IEEE Design
Automation Conference,Anaheim, USA, pp.
807–812,2010.
[7] Ramgovind S, Eloff MM, Smith E, 2010,
"The Management Of Security in Cloud
Computing", IEEE.
[8] C. Clark, K. Fraser, S. Hand, J.G. Hansen, E.
Jul, C. Limpach, I. Pratt and A. Warfield, "Live
Migration of Virtual Machines", In Proc. of the
second USENIX Symposium on Networked.
[9] David Burford, CFA, MBA, MCP Cloud
Computing: A Brief Introduction February 20,
2010 .
[10] Gong, Ch., Liu, J., Zhang, Q., Chen, H., and
Gong, Zh. 2010. The Characteristics of Cloud
Computing. 39th International Conference on
Parallel Processing Workshops.
[11] Saynsky, Barry (1390), compete reference of
cloud computing, Translated by Navid Farokhi,
Computer Sciences Publication, First Version,
Fall 1390
[12] T. Wei-Tek, et al., "Service-Oriented Cloud
Computing Architecture," in Information
Technology: New Generations (ITNG), 2010
Seventh International Conference on, 2010, pp.
684-689.
[13] Rajkumar Buyya , Chee Shin Yeo , Srikumar
Venugopal,James Broberg, and Ivona
Brandic."Cloud Computing and Emerging IT
Platforms:Vision , Hype , and Reality for
Delivering Computing as the 5 th Utility".Future
Generation Computer Systems, ,
Amesterdam,The Netherlands,pp 599-616, June
2009.
[14] M. Armbrust, A. Fox, R. Griffith, A.D.
Joseph, R. Katz, A. Konwinski,"A view of cloud
computing, Commun". ACM New York, USA,pp
50–58,2009.
[15] Ghazizadeh, "Cloud Computing Benefits
and Architecture in E-Learning," in Wireless,
Mobile and Ubiquitous Technology in Education
(WMUTE), 2012 IEEE Seventh International
Conference on, 2012, pp. 199-201.
[16] Nurmi, D., Wolski, R., Grzegorczyk, C.,
Obertelli, G., Soman, S., Youseff, L., and
Zagorodnov, D. 2008. Eucalyptus opensource
cloudcomputing system. In CCA08: Cloud
Computing and Its Applications.
[17] K. R. Joshi, et al., "Dependability in the
cloud: Challenges and opportunities," in
Dependable Systems & Networks, 2009. DSN
'09. IEEE/IFIP International Conference on,
2009, pp. 103-104.

www.ijcat.com 375
[18] Buyya, R., Yeo, C.S., Venugopal, S.,
Broberg, J., and Brandic, I., 2009. Cloud
computing and emerging IT platforms: Vision,
hype, and reality for delivering computing as the
5th utility. Future Generation Computer Systems,
(2009), 25 (6), pp. 599-616.
[19]C. Clark, K. Fraser, S. Hand, J.G. Hansen, E.
Jul, C. Limpach, I. Pratt and A. Warfield, "Live
Migration of Virtual Machines", In Proc. of the
second USENIX Symposium on Networked
Systems Design and Implementation (NSDI),
Boston, MA, USA, May 2005.
[20]Chunming Rong, Son T. Nguyen, Martin
Gilje Jaatun, 2012, "Beyond Lighting: Asurvey
on security challenges in cloud computing",
Elsevier.
[21] D. Zissis and D. Lekkas, "Securing e-
Government and e-Voting with an open cloud
computing architecture," Government
Information Quarterly, vol. 28, pp. 239-251,
2011.
[22] David Burford, CFA, MBA, MCP Cloud
Computing: A Brief Introduction February 20,
2010 .
[23] Amrhein, D., Forget Defining Cloud
Computing.Available[Online]:http://ibm.ulitzer.
com/node/10188.
[24] Ramgovind S, Eloff MM, Smith E, 2010,
"The Management Of Security in Cloud
Computing", IEEE.
[25] M. T. B. Aun ,et al., "Cloud enabled spam
filtering services: Challenges and opportunities,"
in Sustainable Utilization and Development in
Engineering and Technology (STUDENT), 2011
IEEE Conference on, 2011, pp. 63-68.
[26] Gong, Ch., Liu, J., Zhang, Q., Chen, H., and
Gong, Zh. 2010. The Characteristics of Cloud
Computing. 39th International Conference on
Parallel Processing Workshops
[27] P. Mell and T. Grance. (2009, Oct.). The
NIST definition of cloud com-puting, NIST
special publication 800-145, National Institute of
Standards and Technology. Available:
http://www.csrc.nist.gov/publications/nistpubs/8
00-145/SP800-145.pdf.
[28] S. Chungman, et al., "Implementation of
Cloud Computing Environment for Discrete
Event System Simulation using Service Oriented
Architecture," in Embedded and Ubiquitous
Computing (EUC), 2010 IEEE/IFIP 8th
International Conference on, 2010, pp. 359-362.
[29] Ivanka Menken , Gerard Blokdijk."Cloud
Computing Virtualization Specialist Complete
Certification Kit".2008.
[30] S. Marston, et al., "Cloud computing — The
business perspective," Decision Support
Systems, vol. 51, pp. 176-189, 2011

Providing a multi-objective scheduling tasks by Using PSO algorithm for cost management in cloud computing

Recommended

More Related Content

What's hot (16)

Viewers also liked (16)

Similar to Providing a multi-objective scheduling tasks by Using PSO algorithm for cost management in cloud computing (20)

More from Editor IJCATR (20)

Recently uploaded (20)

Providing a multi-objective scheduling tasks by Using PSO algorithm for cost management in cloud computing