You are on page 1of 9

Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com
ScienceDirect
ScienceDirect
Available online at www.sciencedirect.com
Procedia Computer Science 00 (2022) 000–000
Procedia Computer Science 00 (2022) 000–000 www.elsevier.com/locate/procedia
ScienceDirect www.elsevier.com/locate/procedia
Procedia Computer Science 216 (2023) 571–579

7th International Conference on Computer Science and Computational Intelligence 2022


7th International Conference on Computer Science and Computational Intelligence 2022
Analysis of decision support using Elimination and Choice
Analysis ofReality
Expressing decision support using
(ELECTRE) Elimination
method and Choice
in determining best
Expressing Reality (ELECTRE) method in determining
candidate for Programmer position best
candidate for Programmer position
Nurainia,*
a
Nurainia,*
Computer Science Department, School of Computer Science, Bina Nusantara University, Jakarta 11480, Indonesia
a
Computer Science Department, School of Computer Science, Bina Nusantara University, Jakarta 11480, Indonesia

Abstract
Abstract
This research was conducted based on a case study at PT SIER in Indonesia. Programmer team at PT SIER had a very important
role
Thisin supporting
research was the company's
conducted operations.
based on a case Due
studytoatthe
PTincreasing number ofProgrammer
SIER in Indonesia. requests for team
application development
at PT SIER from
had a very various
important
departments to the the
role in supporting Programmer
company'steam, the company
operations. Due tointended to add number
the increasing one Programmer to for
of requests Programmer
applicationteam so that thefrom
development Programmer
various
team could carry
departments to theout their tasksteam,
Programmer faster.theTherefore,
company the company
intended to addopened a vacancy for
one Programmer one Programmer.
to Programmer team so Thenthatthe
thecompany
Programmergot
many candidates
team could carry for
outProgrammer position.
their tasks faster. After conducting
Therefore, the company selection,
opened thea company found
vacancy for onetop three candidates.
Programmer. Then But the company
the company got
had
manytocandidates
choose oneforofProgrammer
them because the company's
position. budget at selection,
After conducting that time was only to recruit
the company found one
top candidate. However,
three candidates. But the
the company
found
had to itchoose
difficult
onetoof choose one of the
them because them as the best
company's candidate.
budget To help
at that time this company
was only to recruit in
onedetermining the best candidate
candidate. However, the companyfor
Programmer position,
found it difficult this research
to choose one ofoffered
them as thethe
usebest
of Elimination
candidate. To andhelp
Choice
thisExpressing
company in Reality method the
determining or ELECTRE method
best candidate for
as a decision position,
Programmer support. Thethis ELECTRE method
research offered thewas
use used in this research
of Elimination becauseExpressing
and Choice this method was amethod
Reality method or to ELECTRE
make decision that
method
could choose support.
as a decision the bestThe option from alternatives
ELECTRE method was thatused
hadinseveral criteriabecause
this research or factors.
this The
methodresult
wasshowed
a methodthattoELECTRE Method
make decision that
could help in determining
choose the best option the best
fromcandidate for that
alternatives Programmer position
had several by incorporating
criteria or factors. The factors
resultorshowed
criteria that
of the candidate.Method
ELECTRE
© 2022
could TheinAuthors.
help Published
determining bycandidate
the best ELSEVIER for B.V.
Programmer position by incorporating factors or criteria of the candidate.
© 2022
© 2023
This The
is an Authors.
open
The Published
accessPublished
Authors. by ELSEVIER
article under
by Elsevier B.V.B.V.
the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
This is an open
Peer-review underaccess article under
responsibility the scientific
of the CC BY-NC-ND license
committee (https://creativecommons.org/licenses/by-nc-nd/4.0)
of the 7th International Conference on Computer Science and
This is an open access article under CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the 7th International Conference on Computer Science and
Computational
Peer-review Intelligence
under 2022 of the scientific committee of the 7th International Conference on Computer Science and
responsibility
Computational Intelligence 2022
Keywords: ELECTRE
Computational Method; 2022
Intelligence decision support; multi-criteria; decision making
Keywords: ELECTRE Method; decision support; multi-criteria; decision making

* Corresponding author.
nuraini@binus.ac.id
E-mail address:author.
* Corresponding
E-mail address: nuraini@binus.ac.id
1877-0509© 2022 The Authors. Published by ELSEVIER B.V.
This is an©open
1877-0509 2022access article under
The Authors. the CCbyBY-NC-ND
Published ELSEVIERlicense
B.V. (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under
This is an open responsibility
access of the
article under the scientific
CC BY-NC-NDcommittee of the
license 7th International Conference on Computer Science and
(https://creativecommons.org/licenses/by-nc-nd/4.0)
Computational Intelligence
Peer-review under 2022 of the scientific committee of the 7th International Conference on Computer Science and
responsibility
Computational Intelligence 2022
1877-0509 © 2023 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the 7th International Conference on Computer Science and Computational
Intelligence 2022
10.1016/j.procs.2022.12.171
572 Nuraini et al. / Procedia Computer Science 216 (2023) 571–579
566 Nuraini / Procedia Computer Science 00 (2022) 000–000

1. Introduction

Programmer team at PT SIER had a very important role in supporting the company's operations because they
were responsible for developing and maintaining applications used to support the company's operations. Due to the
increasing number of requests for application development from various departments to Programmer team, the
company intended to add one Programmer to Programmer team so that Programmer team could carry out their tasks
faster. Therefore, the company opened a vacancy for one Programmer. Then there were many applicants that applied
for this vacancy. So the company got many candidates for Programmer position. At first the company was able to
conduct selection to select the candidates without any difficulties. From that selection, the company found top three
candidates. But the company had to continue doing selection to choose one of them as the best candidate because the
company's budget at that time was only to recruit one Programmer. However, the company found it difficult to
determine who was the best of them because they seemed to be on the same level. Therefore they needed help in
selecting the best candidate.
Based on the problems faced by PT SIER, this research intended to offer using Elimination and Choice
Expressing Reality (ELECTRE) method as a decision support to determine the best candidate. The ELECTRE
method was used in this research because this method is a method to make decision that can choose the best option
of alternatives that have several criteria or factors.
ELECTRE method has been used by many researchers. Mesran Aan, Robbi Rahim and Garuda Ginting used
ELECTRE method to select best lecturer [1]. Eri Eriana, Ivan Pilhanudin, Raffi Maulana, Muhamad Edgar and Riki
Ridwan Margana used ELECTRE method to select housing areas [2]. While another research used ELECTRE
method to select supplier [3]. Fajar Masya, Hendra Prastiawan and Destriyani Putri used ELECTRE method in
development of Lecturer Evaluation System [4]. Then another research used ELECTRE method to select course in
the higher education [5]. S. Supraja and P. Kousalya used ELECTRE method to select best computer system [6].
H.Tezcan Uysal and Kemal Yavuz used ELECTRE method to select logistic centre location [7]. A research
conducted by Xiaohan Yu, Suojuan Zhang , Xianglin Liao and Xiuli Qi used ELECTRE method in prioritized
MCDM environment [8]. Then another research conducted by M. Gökhan Yücel and Ali Görener used ELECTRE
method to make decision for company acquisition [9]. And another research used ELECTRE method to make
decision for virtual network embedding [10].
Although other research have used widely the ELECTRE method but none has specifically used it to determine
the best candidate for Programmer position. Therefore, this research is the first research that specifically uses the
ELECTRE Method to determine the best candidate for Programmer position.

2. Literature Review and Methods

Elimination and Choice Expressing Reality or ELECTRE method is chosen to be used in this research because
that method can provide a reference as decision support to determine the best options. The basic of the ELECTRE
method is a procedure to eliminate options or alternatives that are less good to get the best alternative [1]. To run
ELECTRE method, it is necessary to prepare data or knowledge about the weights of all the factors used [1]. There
are eight steps in the ELECTRE method [1]. The following are the steps in the ELECTRE method.

First step: Forming Decision Matrix.


Decision Matrix is a matrix showing the factors value for each alternative. Each row in Decision Matrix represents
the value of all factors for an alternative. While each column in the Decision Matrix shows the value of all
alternatives for a factor. The values in the Decision Matrix are determined by the Decision Maker.

Second step: Normalization of Decision Matrix.


Normalization of decision matrix is processed by using the following formula:
Nuraini et al. / Procedia Computer Science 216 (2023) 571–579 573
Nuraini / Procedia Computer Science 00 (2022) 000–000 567

(1)
rij means the value of Normalization of Decision Matrix in row i and column j.
xij means the value of Decision Matrix in row i and column j.
m means the number of rows (the number of alternatives)

Third step: Weighting the Normalized Decision Matrix.


Decision Maker should firstly determine the weights of each factor and the addition of these weights should be one
as the following formula:

(2)
n means the number of factors.
j means the order of factors.

Weighting the Normalized Decision Matrix is processed by using the following formula:

vij=wjrij (3)
vij means the value of Weighted Normalization of Decision Matrix in row i and column j.
rij means the value of Normalization of Decision Matrix in row i and column j.
wj means weight of factor in column j.

Each column of the decision matrix is multiplied by the weight (wj) of related factor.

Fourth step: Determining index set of concordance and discordance.


Index set of concordance and discordance is determined by using the following formula:
𝐶𝐶𝑘𝑘𝑘𝑘 = 𝑗𝑗, 𝑦𝑦𝑘𝑘𝑘𝑘 ≥ 𝑦𝑦𝑙𝑙𝑙𝑙 , 𝑓𝑓𝑓𝑓𝑓𝑓 𝑗𝑗 = 1,2,3, … . 𝑛𝑛 (4)
𝐷𝐷𝑘𝑘𝑘𝑘 = 𝑗𝑗, 𝑦𝑦𝑘𝑘𝑘𝑘< 𝑦𝑦𝑙𝑙𝑙𝑙, 𝑓𝑓𝑓𝑓𝑓𝑓 𝑗𝑗 = 1,2,3, … . 𝑛𝑛 (5)
𝐶𝐶𝑘𝑘𝑘𝑘 means concordance index in row k and column l.
D𝑘𝑘𝑘𝑘 means discordance index in row k and column l.
𝑦𝑦𝑘𝑘𝑘𝑘 means the value of Weighted Normalization of Decision Matrix in row k and column j.
𝑦𝑦𝑙𝑙𝑙𝑙 means the value of Weighted Normalization of Decision Matrix in row l and column j.
n means the number of column.

Fifth step: Forming concordance and discordance matrix.


Concordance and discordance matrix is formed by determining the value of the elements in the set of concordance
and discordance index. The element value in the concordance matrix is calculated by adding the weights included in
the set of concordance index using the formula below.

(6)
ckl means the value of Concordance Matrix in row k and column l.
𝐶𝐶𝑘𝑘𝑘𝑘‡ƒ•‹†‡š•‡–‘ˆconcordance in row k and column l.
574 Nuraini et al. / Procedia Computer Science 216 (2023) 571–579
568 Nuraini / Procedia Computer Science 00 (2022) 000–000

The element value in the discordance matrix is calculated by dividing the maximum difference between the values
of the factors included in the set of discordance index with the highest difference between the values of all existing
factors based on the values in the Weighted Normalization of Decision Matrix using the formula below.

(7)
dkl means the value of Discordance Matrix in row k and column l.
Sixth step: Determining dominant concordance matrix and dominant discordance matrix.
Dominant concordance matrix and dominant discordance matrix is built with the help of a threshold value.
Threshold value c for dominant concordance matrix is obtained by using the following formula:

(8)

Then matrix F as dominant concordance matrix is determined by using the following formula:

𝑖𝑖𝑖𝑖 𝑐𝑐𝑘𝑘𝑘𝑘 ≥ 𝑐𝑐 then fkl=1 else 𝑖𝑖𝑖𝑖 𝑐𝑐𝑘𝑘𝑘𝑘 < 𝑐𝑐 then fkl=0 (9)

fkl means the value of dominant concordance matrix in row k and column l.

Threshold value d for dominant discordance matrix is obtained by using the following formula:

(10)

Then matrix G as dominant discordance matrix is determined by using the following formula:

𝑖𝑖𝑖𝑖 d𝑘𝑘𝑘𝑘 ≥ d then gkl=1 else 𝑖𝑖𝑖𝑖 d𝑘𝑘𝑘𝑘 < d then gkl=0 (11)

gkl means the value of dominant discordance matrix in row k and column l.

Seventh step: Determining aggregate dominant matrix.


Aggregate dominant matrix or total dominant matrix or matrix E is determined by calculating multiplication
between the matrix elements of matrix F with the matrix elements of matrix G using the following formula:

𝑒𝑒𝑘𝑘𝑘𝑘 = 𝑓𝑓𝑘𝑘𝑘𝑘 𝑥𝑥𝑥𝑥𝑘𝑘𝑘𝑘 (12)

𝑒𝑒𝑘𝑘𝑘𝑘means the value of aggregate dominant matrix in row k and column l.

Eighth step: Elimination of less good alternatives


Total dominant matrix or matrix E shows the preferred order of each option. The line in matrix E that has the lowest
total value can be eliminated.

3. Result and Discussion

The initial step was forming Decision Matrix. The steps for forming the Decision Matrix can be seen in the
flowchart in Fig. 1.
Nuraini et al. / Procedia Computer Science 216 (2023) 571–579 575
Nuraini / Procedia Computer Science 00 (2022) 000–000 569

Fig. 1. Flowchart to form Decision Matrix

Decision Maker determined the list of candidates include Candidate 1 (C1), Candidate 2 (C2) and Candidate 3
(C3). Decision Maker determined the factors that would be judged from each candidate include Expected Salary
(F1), Education (F2), Programming Skill (F3), Experience (F4), Accuracy (F5), Attitude (F6), Communication Skill
(F7) and Analytical Skill (F8).
Decision Maker determined the value of each factor for each candidate and the result was displayed in a Decision
Matrix as shown in Table 1.
Table 1. Decision Matrix

The second step was normalization of decision matrix based on equation (1). Then the result was displayed in a
Normalization of Decision Matrix table as shown below.

Table 2. Normalization of Decision Matrix

The third step was determining the weight of the matrix. Before processing this step, Decision Maker firstly
determined the weights of each factor by using equation (2) then the result was displayed in table as shown below.
576 Nuraini et al. / Procedia Computer Science 216 (2023) 571–579
570 Nuraini / Procedia Computer Science 00 (2022) 000–000

Table 3. List of Factor Weight

After determining the weights of each factor, weighting Normalization of Decision Matrix can be processed by
multiplying the value of each cell in the matrix by the weight of the related factor based on equation (3). Then the
result was displayed in Weighted Normalization of Decision Matrix table as shown below.

Table 4. Weighted Normalization of Decision Matrix

The fourth step was determining index set of concordance and discordance. By using equation (4), the index set
of concordance could be obtained as in table as shown below.

Table 5. The index set of concordance

Index set of discordance could synchronize with index that was not contained in the index set of concordance.
Then by using equation (5), the index set of discordance was obtained as in table as shown below.

Table 6. The index set of discordance

The fifth step was forming concordance and discordance matrix. The formation of the concordance matrix was
processed by determining the elements value of index set of concordance by using equation (6) and the result as
follows:
c11 = 0, c12 = 0.5487, c13 = 0.6561, c21 = 0.7500, c22 = 0, c23 = 0.6561, c31 = 0.7500, c32 = 0.7539 and c33 = 0.

Then the concordance matrix could be obtained as shown below:


Nuraini/ Procedia
Nuraini et al. / Procedia
ComputerComputer
Science Science 216000–000
00 (2022) (2023) 571–579 577
571

The formation of the discordance matrix was processed by determining the elements value of the index set of
discordance by using equation (7).

Then the discordance matrix could be obtained as follows:

The sixth step was calculating threshold value before determining dominant concordance matrix and dominant
discordance matrix. By using equation (8), value threshold c was obtained as follows:

By using equation (10), value threshold d was obtained as follows:

Then matrix F as dominant concordance matrix was determined by using equation (9) as follows:

Then matrix G as dominant discordance matrix was determined by using equation (11) as follows:
578
572 Nuraini
Nuraini et al. / Computer
/ Procedia Procedia Computer
Science 00Science
(2022)216 (2023) 571–579
000–000

The seventh step was determining matrix E as total dominant matrix. By using equation (12), matrix E was
obtained as follows:

Each line of total dominant matrix or matrix E represent each alternative of Decision Matrix as follows:

Then elimination of less good alternatives as the final step could be processed. To do this, total value of each line
in E matrix as total value of the related candidate was calculated as shown in a table as shown below.

Table 7. Total value of each Candidate

From the last result above, it could be seen that candidate C1 had the lowest total value, so candidate C1 could be
eliminated. Then it could also be seen that the total value of candidate C2 is lower than the total value of candidate
C3, so candidate C2 could also be eliminated because there was only one programmer vacancy. The result showed
that candidate C3 had the highest total value. Therefore candidate C3 could be decided as the best candidate.

4. Conclusion

This research show that Elimination and Choice Expressing Reality (ELECTRE) method can be used as a
decision support in determining the best candidate for Programmer position because this method can generate
reference about the best candidate by incorporating the factors or criteria of the candidate.
For further research, researcher may try to determine the most powerful free video editing software using the
ELECTRE method, where the results of this research can be a recommendation for YouTube video creators.

References

[1] Aan, Mesran, Robbi Rahim, and Garuda Ginting. (2017) “Implementation of Elimination and Choice Expressing Reality (ELECTRE)
method in selecting the best lecturer (case study STMIK BUDI DARMA).” International Journal of Engineering Research & Technology6
(2): 141–144.
[2] Eriana, Eri, Ivan Pilhanudin, Raffi Maulana, Muhamad Edgar, and Riki Ridwan Margana. (2020) “Decision Support System for selecting
housing areas using ELECTRE method.” PalArch’s Journal of Archaeology of Egypt / Egyptology17 (10): 1047–1060.
Nuraini et al. / Procedia Computer Science 216 (2023) 571–579 579
Nuraini / Procedia Computer Science 00 (2022) 000–000 573

[3] Fei, Liguo, Jun Xia, Yuqiang Feng, and Luning Liu. (2019) “An ELECTRE-Based Multiple Criteria Decision Making Method for Supplier
Selection Using Dempster-Shafer Theory.” IEEE Access7: 84701–84716.
[4] Masya, Fajar, Hendra Prastiawan, and Destriyani Putri. (2017) “Design and implementation of Lecturer Evaluation System using ELECTRE
method in web-based application.” International Research Journal of Computer Science4 (5): 242–250.
[5] Sudarma, Made, Anak Agung Kompiang Oka Sudana, and Irwansyah Cahya. (2015) “Decision Support System for the Selection of Courses
in the Higher Education using the Method of Elimination Et Choix Tranduit La Realite.” International Journal of Electrical and Computer
Engineering5 (1): 129–135.
[6] Supraja, S., and P. Kousalya. (2016) “ELECTRE Method for the Selection of Best Computer System.” Indian Journal of Science and
Technology9 (39): 1–5.
[7] Uysal, H.Tezcan, and Kemal Yavuz. (2014) “Selection of Logistics Centre Location via ELECTRE Method: A Case Study in Turkey.”
International Journal of Business and Social Science5 (9): 276–289.
[8] Yu, Xiaohan, Suojuan Zhang , Xianglin Liao, and Xiuli Qi. (2018) “ELECTRE methods in prioritized MCDM environment.” Information
Sciences Journal424: 301–316.
[9] Yücel, M. Gökhan, and Ali Görener. (2016) “Decision Making for Company Acquisition by ELECTRE Method.” International Journal of
Supply Chain Management5 (1): 75–83.
[10] Zhang, Peiying, Haipeng Yao, Chao Qiu, and Yunjie Liu. (2018) “Virtual Network Embedding Using Node Multiple Metrics Based on
Simplified ELECTRE Method.” IEEE Access6: 37314–37327.

You might also like