GRID COMPUTING: STRATEGIC DECISION MAKING IN RESOURCE SELECTION

International Journal of Computer Science, Engineering and Applications (IJCSEA) Vol.2, No.6, December 2012
DOI : 10.5121/ijcsea.2012.2601 1
GRID COMPUTING: STRATEGIC DECISION MAKING
IN RESOURCE SELECTION
Shahriar Mohammadi1
, Sajad Homayoun2
, Ebrahim Torki Zadeh3
1
Associate Professor, IT Group, Department of Industrial Engineering, K. N. Toosi
University of Technology, Tehran, Iran
mohammadi@kntu.ac.ir
2
Corresponding Author, IT Group, Department of Industrial Engineering, K. N. Toosi
University of Technology, Tehran, Iran
sajadhomayoun@gmail.com
3
IT Group, Department of Industrial Engineering, K. N. Toosi University of Technology,
Tehran, Iran and Ahvaz Research Group, National Iranian Oil Products Distribution
Company (NIOPDC), Ahvaz, Iran
e.torkizadeh@gmail.com
ABSTRACT
The rapid development of computer networks around the world generated new areas especially in computer
instruction processing. In grid computing, instruction processing is performed by external processors
available to the system. An important topic in this area is task scheduling to available external resources.
However, we do not deal with this topic here. In this paper we intend to work on strategic decision making
on selecting the best alternative resources for processing instructions with respect to criteria in special
conditions. Where the criteria might be security, political, technical, cost, etc. Grid computing should be
determined with respect to the processing objectives of instructions of a program. This paper seeks a way
through combining Analytic Hierarchy Process (AHP) and Technique for Order of Preference by Similarity
to Ideal Solution (TOPSIS) to help us in ranking and selecting available resources according to
considerable criteria in allocating instructions to resources. Therefore, our findings will help technical
managers of organizations in choosing as well as ranking candidate alternatives for processing program
instructions.
KEYWORDS
Grid Computing, Resource Selection, Strategic Decision Making, AHP, TOPSIS.
1. INTRODUCTION
A. What is Grid Computing?
In general, grid is aimed at sharing resources in a dynamic and probably heterogeneous
environment. These resources would be accessible through a number of policies. This sharing is
mainly used for computational objectives in scientific applications. They can be in forms such as
CPU, hard disk, software, as well as sensors. If your paper is intended for a conference, please
contact your conference editor concerning acceptable word processor formats for your particular
conference.

2
In late 1990, computational grid networks emerged as substitute for current super computers to
solve special problems, and they required numerous numerical computations and more amount of
distributed data.
According to [1], grid computing is hardware and software infrastructure which offer a cheap,
distributable, coordinated and reliable access to powerful computational capabilities.
"When the network is as fast as the computer's internal links, the machine disintegrates across the
net into a set of special purpose appliances [2]".
B. What is the problem?
Usually a number of criteria affect the selection and due to the variety of criteria, decision maker
would be hesitant to decide on a specific problem. The problem here is choosing an objective for
processing instructions with respect to special criteria considered by an organization. They range
from political, security, processing speed, place of objective, etc. (if the case is considered from
another point of view, the problem of not choosing inappropriate alternatives would be the case
too). The important point is that the best decision is made in the most complex situations.
Suppose we plan to construct refineries in a country. To find location, we should extended
processing by relevant software. Here, we encounter two choices: First, we may go along with
processing with our own facilities and wait for months or even years to arrive at the results! Or
we may rely on grid computing. If the latter is selected, we should decide that taking security,
political, cost and other criteria into account, which nodes are more appropriate for processing our
instructions with no risk. Each criterion has its own importance, and this importance varies in
different organizations. For example, in our example the criterion of security is more important
than the cost criterion. The question is that “is it possible to process by any external system?”
Considering malicious intents for any country or organization, the answer to this would be
definitely “No” (see Figure. 1).
Figure 1. What the problem is
Related works are mentioned in section 2. section 3 outlines AHP method and section 4 is about
TOPSIS method. The proposed framework will be dealt with in section 5. And finally Section 6 is
devoted to conclusion.
2. RELATED WORKS
Most of authors have worked on grid computing issues form the technical perspective. Authors
like Bawa and Sharma [3], Vijaykumar and Banu [4] and Metke and Ekl [5] have proposed
Our Location
Location
A
Location
B
Location
C
Location
D
Our Location
Location
A
Location
B
Location
C
Location
D
a. All locations seem same b. Some locations are risky in real world

3
models to make high the security of system by self-protection capability. Kavitha and
Sankaranarayanan [6] have proposed a computational grid architecture for sharing of resources
based on user QoS and trust, but their considered criteria are not flexible and users can not add
their desired criteria to the framework.
An important issue is most of authors worked on criteria like security, etc. from the technical
perspective and we are going to work on grid resource selection from the managers view. Our
framework tries to help the managers when they want to block insecure resources from the list or
add one according to their organization strategies and policies. We are going to design a flexible
framework to rank resources.
3. AHP
Analytical hierarchical process (AHP) is an MADM method used for decision making and
selecting an alternative out of many alternatives based on the related [7, 8]. This method is
created in 1980 by Saaty [5]. The Analytical Hierarchical Process reflects natural behaviour and
thought of human being. This technique analyzes complicated issues based on their mutual
interaction and changes them into simple form and then starts solving them.
AHP follows the following steps:
• Step 1. Modelling AHP Structure
In this step, the objectives of decision making are ordered as a hierarchy of decision elements
which are related to each other [10]. The elements include criteria and decision alternatives
(Figure 2).
Figure. 2. AHP structure
Step 2. Establish a pair-wise comparison decision matrix
According to each criterion, a comparison is made between various alternatives. A judgment is
also made on the importance of decision criterion with respect to pair-wise comparisons.
In AHP, elements in each level are compared with their related elements in higher level in pair-
wise and their weight is calculated. These weights are called mean weight. These weights are
combined to arrive at the ultimate weight for each alternative [10].
First, there is n × n matrix (n represents the number of criteria) and manager determines relative
importance of criteria. Then, the decision matrix will be n × n matrix where again n represents
the number of criteria.
.
C1
Goal
... Cm
A1 ... An
Goal
Criteria
Alternative

4
Standard Preference table (Table 1) shows how to put the selected numbers. The quantities are
assigned in the matrix based on Eq. 2.
= ⋮ ⋮ (10)
= , = 1, > (2)
Table 1. Standard preference table
• Step 3. Calculation of Criteria Weight
The weight of each criterion is calculated through normalization of each line or column of matrix
A. There are a number of methods for calculating the matrix’ weight such as Least squares
method, Logarithmic least squares method, Eigenvector method, Approximation method, to name
a few.
= (3)
The can be acquired. If the is equal to n and the rank of matrix A is n, A is consistent.
In this case, the relative criteria can be discussed. The weight of each criterion will be calculated
by normalizing any of the rows or columns of matrix A [11].
• Step 4. Test Consistency
AHP matrixes should be compatible. In other words, if from the manager’s viewpoint, the
comparative quantity of Criterion C1 compared with C2 and C3 is respectively 5 and 7, then C2
compared with C3 should be higher, i.e. it should be 3. Here two parameters are involved:
consistency index (CI), and consistency ratio (CR) that are defined as below:
= , (4)
= , (5)
where RI is random index. For different count of criteria, it has different value demonstrated in
Table 2. If CR is less than 0.10, the result can be acceptable and matrix A is sufficient
consistency. Otherwise, we have to return to step 1 and repeat again.
ValueDefinition
1Equal importance
3Weak importance
5Essential importance
7Demonstrated importance
9Extreme importance
2, 4, 6, 8Intermediate values

5
Table 2. The relationship between value and count of criterion
4. TOPSIS
In 1981, TOPSIS model was proposed by Hwang and Yoon [12] for ranking. Based on this
theory, the best alternative bears two features: it is the nearest one to the positive-ideal, and it is
the farthest on to the negative-ideal [13].
The positive-ideal solution minimizes the cost criteria and maximizes the benefit criteria. It is
consisted of all best values attainable from the criteria. At the same time, the negative-ideal
solution is a solution that maximizes the cost criteria and minimizes the benefit criteria, which has
all worst values attainable from the criteria [14, 15].
The TOPSIS steps are followed [16].
• Step 1. Constructing a decision matrix
If the number of criteria and alternatives are n and m respectively, there will be a decision matrix
with m lines and n columns. See Table 3.
Table 3. Decision Matrix
In Table 3, there are quantities which represents efficiency rate of each alternative with respect
to the related criterion.
• Step 2. Calculation of normalized decision matrix
The normalized quantities are calculated as below.
=
∑
= 1, 2, … , ; = 1,2, … , (6)
• Step 3. Calculate the weighted normalized decision matrix
The matrix is from multiplying the normalized decision matrix by its associated weights as:
= × , = 1, 2, … , ; = 1, 2, … , , (7)
where w is the weight of the jth
attribute or criterion, and ∑ w = 1.
n 1 2 3 4 5 6 7 8 9
RI 0 0 0.58 0.90 1.12 1.24 1.32 1.41 1.45
Criterion n...Criterion 1
...Alternative 1
............
...Alternative m

6
• Step 4. Determine the positive-ideal and negative-ideal solutions
•
∗
= { ∗
, ∗
, … , ∗ } = ∈ ′
, ∈ ′′
, = 1, 2, … , ; = 1, 2, … , (8)
= { , , … , } = ∈ ′
, ∈ ′′
, = 1, 2, … , ; = 1, 2, … , (9)
where ′ is associated with benefit criteria, and ′′ is associated with cost criteria.
• Step 5. Using n-dimensional Euclidean distance
The closeness of each alternative to the positive-ideal is calculated as below:
∗
= ∑ ( , ∗
) = 1, 2, … , . (10)
Distance to the negative-ideal is also calculated as below:
= ∑ ( , ) = 1, 2, … , . (11)
• Step 6. Calculation of Closeness to Ideal condition
It is defined for ith
alternative :
∗
= ∗ . (12)
• Step 7. Ranking
Ranking is made based on ∗
. This quantity is between 0-1 for each alternative. The ultimate
ranking will be descending ∗
and ∗
will be the higher one [17].
5. PROPOSED FRAMEWORK
The usefulness of a combined AHP and TOPSIS is well documented [18], but we use it to solve a
different problem. We designed a framework and it is shown in Figure 3. Let us review our
suggested framework in an example.

7
Figure 3. Proposed model.
Imagine we intend to choose a processor for finding the best location for a refinery.
• Stage 1.
In this stage, objectives and strategies of the manager will be identified, and the evaluation
criteria will be selected through those objectives. The manager has determined these criteria:
1) Cost: Cost per each clock or instruction.
2) Security: confidentiality and security of executing instructions.
3) Location: country or region.
4) Process speed: execution time of each instruction
5) RTT (Round Trip Time): Time of trip and return between source and destination.
Let us we intend to select an alternative out of the alternatives: A, B, C.
Interview with management
Criteria identification
Stage 1
Evaluation of alternatives
Ranking alternatives
Stage 3
Y
Acquiring criteria weights
Constructing hierarchical
structure
Forming pair-wise comparison
matrices
Approving
criteria
weights
Stage 2
N
Interview with management
Criteria identification
Stage 1
Evaluation of alternatives
Ranking alternatives
Stage 3
Y
Acquiring criteria weights
Constructing hierarchical
structure
Forming pair-wise comparison
matrices
Approving
criteria
weights
Stage 2
N

8
• Stage 2.
We build the hierarchical structure of decision elements (Figure 4).
Figure 4. Hierarchical evaluation structure and criteria weight of Z site
We introduce pair-comparison matrix for five criteria and ask the manager to put weight based
on priorities (Table 4).
Table 4. Priorities of criteria
We build matrix for each criterion, and determine alternatives’ status towards each other (Table 5,
Table 6, Table 7, Table 8, Table 9).
Table 5. Security criteria pair-wise comparison
Rank Destination Resources
SecurityCost Location Process Speed RTT
A B C
Goal
Criteria
Alternatives
RTT: Round Trip Time
RTT
Process
Speed
LocationSecurityCostCriteria
11/31/31/91Cost
99719Security
7311/73Location
311/31/93Process
Speed
11/31/71/91RTT
Incon.= 0.08
CBASecurity
1/31/71A
31B
1C
Incon. =0.01

9
Table 6. Cost criteria pair-wise comparison
Table 7. Location criteria pair-wise comparison
Table 8. Process speed criteria pair-wise comparison
Table 9. Process speed criteria pair-wise comparison
Then we calculate the weight of matrixes. For example, we used Expert Choice software
version 11, and arrived at Figure 5
Figure 5. Ultimate Weight calculated by Expert Choice 11.
CBACost
731A
51B
1C
Incon =0.06
CBALocation
31/51A
71B
1C
Incon=0.06
CBALocation
331A
1/31B
1C
Incon=0.13
CBALocation
751A
31B
1C
Incon=0.06

10
• Stage 3.
Once stages 1 and 2 are finished, we arrived at weights of each criterion and alternative. In this
stage, we evaluated alternatives and ultimate ranking. Table 10 and Table 11 is formed by AHP
and it is used as input for TOPSIS model. Then it is normalized. The TOPSIS model gave us
Table 13 as final results according to Table 12.
Table 10. Calculated weights by AHP
Table 11. Ultimate weight for criteria by AHP
Table 12. Status for each alternative
Table 13. Positive-Ideal and Negative-Ideal distance for each alternative
The final weight is calculated by Eq. 12 and is showed in Table 14.
RTT
Process
Speed
0.7310.5840.1880.0880.649A
0.1880.1350.7310.6690.279B
0.0810.2810.0810.2430.072C
WeightCriteria
0.046Cost
0.647Security
0.178Location
0.089Process Speed
0.040RTT
RTT
Process
Speed
0.02920.05190.03340.05690.0298A
0.00750.01200.13010.43280.0128B
0.00320.02500.01440.15720.0033C
-*
0.00330.1506A
0.15470.0023B
0.01010.0913C

11
Table 14. Final weights from TOPSIS
Now the ultimate ranking reflects descending order of alternatives’ weights, where in our
example they are A, C, and B.
We reserve the alternatives with the second rank, or we may divide the program into parts and use
the resources in parallel.
6. CONCLUSIONS
Grid computing helps us to process computer instructions faster and it is very useful in processing
programs with huge amount of instructions. Sometimes we have to avoid deputing instructions to
potential inappropriate available resources for the reasons such as cost, security, governmental
policies, etc. In this article, a method is proposed for decision making and choosing one or some
resources out of existing alternatives with respect to the special conditions of an organization.
This method helps decision makers (managers) in complicated conditions where a number of
criteria such as security, cost, processing speed, etc. are involved in choosing an alternative. We
developed a framework based on AHP and TOPSIS and described it through an example. AHP is
used to put weight to criteria and alternatives. We assigned its output as input to TOPSIS and
finally we arrived at the ultimate ranking leading to an optimum choice.
REFERENCES
[1] Foster, I. (2002), “What is the Grid”, Daily News and Information for the Global Grid Community J.,
Vol. 1, No. 6.
[2] Barroso & Jose Morales (2005), “From Computer Networks to the Computer on Net, the
Convergence of Internet, Broadband and Telephone Networks in the IEEE 802 Standards,” IEEE
Global Communication Newsletter, page 2.
[3] Kumar, R., Bawa & Sharma, G. (2012), “Reliable Resource Selection in Grid Environment”. In
International Journal of Grid Computing and Applications, Vol. 3, No. 1.
[4] Vijayakumar, V. & Banu, R.S.D.W. (2009), “Secured Resource Selection in Grid Computing: Trust
and Reputation Sentient Scheme”, Advances in Communication and Networking, Vol. 27, pp. 169-
183.
[5] Metke, A.R. & Ekl, R.L. (2010), “Security Technology for Smart Grid Networks”, Smart Grid, IEEE
Transactions on , Vol. 1, No. 1, pp. 99-107.
[6] G.Kavitha & V.Sankaranarayanan. (2011), “Resource Selection in Computational Grid Based on User
QoS and Trust”, IJCSNS International Journal of Computer Science and Network Security, Vol. 11,
No. 3, pp. 214-221.
[7] Gumus, A. T. (2009), “Evaluation of hazardous waste transportation ﬁrms by using a two-step fuzzy-
AHP and TOPSIS methodology,” Expert Systems with Applications J., Vol. 36, pp. 4067–4074.
[8] Lin, M. C., Wang, C. C., Chen, M. S., & Chang, C. A. (2008), “Using AHP and TOPSIS approaches
in customer-driven product design process,” Computers in Industry J., Vol. 59, No. 1, pp. 17–31.
[9] Saaty, T. L. (1980), “The analytic hierarchy process,” New York: McGraw-Hill.
[10] Albayrak, E., & Erensal, Y. C. (2004), “Using analytic hierarchy process (AHP) to improve human
performance. An application of multiple criteria decision making problem,” Intelligent Manufacturing
J., Vol. 15, pp. 491–503.
Final weight
0.1539A
0.1570B
0.1014C

12
[11] Wang, J. J., & Yang, D. L. (2007), “Using a hybrid multi-criteria decision aid method for information
systems outsourcing,” Computers & Operation Research J., Vol. 34, pp. 3691–3700.
[12] Hwang, C. L., & Yoon, K. (1981), “Multiple attribute decision making: Methods and applications. A
state of the art survey”, New York: Springer-Verlag.
[13] Ertugrul, I., & Karakasoglu, N. (2007), “Performance evaluation of Turkish cement ﬁrms with fuzzy
analytic hierarchy process and TOPSIS methods,” Expert Systems with Applications J., Vol. 36, No.
1, pp. 702–715.
[14] Wang, Y. J. (2008), “Applying FMCDM to evaluate financial performance of domestic airlines in
Taiwan,” Expert Systems with Applications J., Vol. 34, pp. 1837–1845.
[15] Wang, Y. M., & Elhag, T. M. S. (2006), “Fuzzy TOPSIS method based on alpha level sets with an
application to bridge risk assessment,” Expert Systems with Applications J., vol. 31, pp. 309–319.
[16] Shyur, H. J., & Shih, H. S. (2006), “A hybrid MCDM model for strategic vendor selection,”
Mathematical and Computer Modeling J., Vol. 44, pp. 749–761.
[17] Opricovic, S., & Tzeng, G. H. (2007), “Extended VIKOR method in comparison with outranking
methods,” Operational Research Euro. J., Vol. 178, No. 2, pp. 514–529.
[18] Xiaobing Yu, Shunsheng Guo, Jun Guo & Xiaorong Huang (2011), “Rank B2C e-commerce websites
in e-alliance based on AHP and fuzzy TOPSIS”, Expert Systems with Applications J., Vol. 38, pp.
3550-3557.

GRID COMPUTING: STRATEGIC DECISION MAKING IN RESOURCE SELECTION

More Related Content

What's hot (19)

Similar to GRID COMPUTING: STRATEGIC DECISION MAKING IN RESOURCE SELECTION (20)

Recently uploaded (20)

GRID COMPUTING: STRATEGIC DECISION MAKING IN RESOURCE SELECTION