Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode

NATIONAL JOURNAL CONTEST 2014

SELECTING OF HORTICULTURE PLANTS MOST EXACT TO CULTIVATION


USING TOPSIS METHODE

Shilvia Nurfauziah under guidance Dr. Bambang Ruswandi, M.Stat, Dr.Yanne Irene, M.Si
Faculty of Sains and Technology Matemathics
Islamic State University Syarif Hidayatullah Jakarta
e-mail: shilvia.nurfauziah10@mhs.uinjkt.ac.id, bambang.ruswandi@gmail.com,
yanne.irene@uinjkt.ac.id

ABSTRACT
This research was conducted with the curiosity about the criteria required for a selecting
of horticulture plants most exact for cultivation in every zone. One of methods to determine
the plants most exact to cultivation is using the Technique for Order Preference Similarity of
Ideal Solution (TOPSIS). In this thesis, using primary data obtained by interview and
questionnaire by the head of Sub Direktorat Pascapanen Tanaman Pohon dan Perdu Dirjen
Hortikultura Kementrian Pertanian RI. The result shows that the criteria taste liked by peoples
is the most important criteria of selecting the horticulture plants most exact to cultivation with
a weighting of 33%, then often to bear fruit with a weighting of 27%, amendable in narrow
area with a weighting of 19%, easy to care with a weighting of 11% and difficult to broken
with a weighting of 10%. The horticulture plants most exact to cultivation building on
TOPSIS methode is Mango’s plant with a weighting preference value of 0,62 then starfruit
with a weighting of 0,49, guava with a weighting of 0,40, crystal guava with a weighting
0,40, jackfruit with a weighting of 0,11, rambutan with a weighting of 0,07 and soursop with
a weighting of 0,04.

Keywords: Alternative, Criteria, Plants. TOPSIS , Horticulture, Preference

PRELIMINARY

The fulfillment of nutritional needs from horticultural crops in Indonesia is under of


sufficient value. The minimum requirement of horticultural crops for Indonesian is relatively
lower if compared to Japan or Netherlands. Indonesia only reached 100 grams of fruit and 200
grams of vegetables per capita per day, while the Netherlands reached 107 grams of fruits and
225 grams of vegetables per capita per day followed by Japan reached 111 grams of fruit and
347 grams vegetables per capita per day. To be able to meet these minimum requirements, the
consequences is should pressing horticulture product improvement as soon as possible.

The horticultural product fulfillment effort is inseparable from the government's role in
implementing the Food Diversification and Nutrition Project (DPG). The Implementation of
DPG is every family should emphasized to maximally utilize they area or their home yard for
cultivation of horticultural crops. Therefore, it is desirable a study that can give an idea of the
importance of various horticultural crops that require more attention to be cultivated in every

1
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

region. In relation to this research the method to used as a decision support is Technique for
Order Preference Similarity of Ideal Solution (TOPSIS).

THEORY

TOPSIS Algorithm

Baically, Topsis Algorithm followed this steps:


a. Spanning Matrix Decision
Matrix Decision 𝑋 in order to 𝑚 alternative to be evaluated in 𝑛 criteria.
𝑐1 𝑐2 𝑐3 ⋯ 𝑐𝑛

𝑎1 𝑥11 𝑥12 𝑥13 ⋯ 𝑥1𝑛


𝑎2 𝑥21 𝑥22 𝑥23 ⋯ 𝑥2𝑛
𝑋 = 𝑎3 𝑥31 𝑥32 𝑥33 ⋯ 𝑥3𝑛
⋮ ⋮ ⋮ ⋮ ⋱ ⋮
𝑎𝑚 𝑥𝑚1 𝑥𝑚2 𝑥𝑚3 ⋯ 𝑥𝑚𝑛

b. Determine Normalized Decision Matrix

𝑥 𝑖𝑗
𝑟𝑖𝑗 = , 𝑖 = 1,2, … , 𝑚; 𝑗 = 1,2, … , 𝑛.
𝑛 2
𝑖=1 𝑥 𝑖𝑗

c. Calculates a Weighted Normalized Decision Matrix

𝑦𝑖𝑗 = 𝑤𝑖 𝑟𝑖𝑗 , 𝑖 = 1,2, … , 𝑚; 𝑗 = 1,2, … , 𝑛

Where 𝑤𝑖 is weighted preference in every criteria has determined by decision maker


when 𝑊 = 𝑤1 , 𝑤2 , … , 𝑤𝑛 .
d. Determine the Ideal Positive Solution 𝐴+ and the Ideal Negative Solution 𝐴−

𝐴+ = maxi 𝑦𝑖𝑗 |𝑗 ∈ 𝐽 , (mini 𝑦𝑖𝑗 |𝑗 ∈ 𝐽∗ )|𝑖 = 1,2,3, … , 𝑚


= {𝑦𝑗+|𝑗 = 1,2, … , 𝑛}
= 𝑦1+, 𝑦2+, 𝑦3+, … , 𝑦𝑛+
𝐴− = mini 𝑦𝑖𝑗 |𝑗 ∈ 𝐽 , (maxi 𝑦𝑖𝑗 |𝑗 ∈ 𝐽∗ )|𝑖 = 1,2,3, … , 𝑚
= {𝑦𝑗−|𝑗 = 1,2, … , 𝑛}
= 𝑦1−, 𝑦2−, 𝑦3−, … , 𝑦𝑛−

2
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

Where:

𝐽 = {𝑗 = 1,2,3, … , 𝑛 and 𝐽 is benefit criteria}


𝐽∗ = {𝑗 = 1,2,3, … , 𝑛 and 𝐽∗ is cost criteria}
𝑦𝑖𝑗 = Element of Weighted Normalized Decision Matrix
𝑦𝑗+ = Element of Ideal Positive Solustion
𝑦𝑗− = Element of Ideal Negative Solustion

e. Counting of Separation
Separation denoted 𝑆 where 𝑆𝑖+ is distance between value of alternative with ideal
positive solution and 𝑆𝑖− distance between value of alternative with ideal negative
solution. 𝑆𝑖+ dan 𝑆𝑖−defined:

𝑆𝑖+ = 𝑛
𝑗 =1(𝑦𝑖𝑗 − 𝑦𝑗+)2 , 𝑖 = 1,2,3, … , 𝑚

𝑆𝑖− = 𝑛
𝑗 =1(𝑦𝑖𝑗 − 𝑦𝑗−)2 , 𝑖 = 1,2,3, … , 𝑚

Where:
𝑦𝑖𝑗 = elemen dari matriks keputusan yang ternormalisasi terbobot 𝑌.
Others is followed before.

f. Counting Preference Value


Preference Value denoted 𝐶𝑖+ if selecting of maxium value dan 𝐶𝑖− if selecting minimum
value, defined:
𝑆𝑖−
𝐶𝑖+ = , 𝐶𝑖+ ∈ 0,1 ∀ 𝑖 = 1,2,3, … , 𝑚
𝑆𝑖+ +𝑆𝑖−

𝑆𝑖+
𝐶𝑖− = ,𝐶𝑖− ∈ 0,1 ∀ 𝑖 = 1,2,3, … , 𝑚
𝑆𝑖+ +𝑆𝑖−

g. Rangking Of Alternative
If using the equation 𝐶𝑖+ then the alternative is sorted from the largest value to the
smallest value. The alternative with the largest 𝐶𝑖+ is the best solution because it has the
closest preference value to the positive ideal solution. Conversely, if using the equation
𝐶𝑖− then the alternative is sorted from the smallest value to the largest value. An

3
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

alternative with a smallest value of 𝐶𝑖− is the best solution because it has the furthest
preference value against the ideal ideal solution.

METHODOLOGY

FLOWCHART

4
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

RESULT AND DISCUSSION

Decision Matrix

Tabel 1 Decision Matrix


Criteria
Alternative Often to Dificult to Amendable in
Taste Easy to Care
Bear Fruit Broken Narrow
Mango 0,41 0,12 0,15 0,13 0,11
Rambutan 0,04 0,05 0,08 0,11 0,11
Guava 0,18 0,15 0,18 0,22 0,18
Crystal Guava 0,14 0,19 0,27 0,19 0,20
Starfruit 0,09 0,38 0,16 0,18 0,22
Jackfruit 0,07 0,07 0,08 0,09 0,11
Soursop 0,06 0,03 0,08 0,07 0,08

Base the table we have information that every alternative have weight in order to every
criteria.

Normalize Decision Matrix

Tabel 2 Normalize Decision Matrix


Kriteria
Often to
Alternative Dificult to Amendable in
Taste Bear Easy to Care
Broken Narrow
Fruit
Mango 0,84 0,25 0,37 0,34 0,27
Rambutan 0,09 0,11 0,18 0,28 0,27
Guava 0,37 0,32 0,44 0,54 0,45
Crystal Guava 0,28 0,41 0,64 0,48 0,49
Starfruit 0,19 0,79 0,39 0,44 0,55
Jackfruit 0,14 0,16 0,19 0,23 0,27
Soursop 0,13 0,06 0,18 0,19 0,19

Weighted Normalize Decision Matrix

Tabel 3 Weighted Normalize Decision Matrix


Kriteria
Often to
Alternative Dificult to Amendable in
Taste Bear Easy to Care
Broken Narrow
Fruit
Mango 0,28 0,07 0,04 0,04 0,05
Rambutan 0,03 0,03 0,02 0,03 0,05
Guava 0,12 0,09 0,04 0,06 0,09
Crystal Guava 0,09 0,11 0,06 0,05 0,09
Starfruit 0,06 0,21 0,04 0,05 0,10
Jackfruit 0,05 0,04 0,02 0,03 0,05
Soursop 0,04 0,02 0,02 0,02 0,04

5
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

Base the table we have information that every alternative created new weight in order to every
criteria. With the every weight is 𝑊 = (0,33 0,27 0,10 0,11 0,19).

Ideal Positive and Negative Solution


Tabel 4 Ideal Positive Solution
Often to Dificult
Amendable in
Criteria Taste Bear to Easy to Care
Narrow
Fruit Broken
Max Value
0,28 0,21 0,06 0,06 0,10
(𝑨+ )

Tabel 4 Ideal Negative Solution


Often to Dificult
Amendable in
Criteria Taste Bear to Easy to Care
Narrow
Fruit Broken
Min Value
0,03 0,02 0,02 0,02 0,04
(𝑨− )

Separation

Tabel 5 Separation of Ideal Positif Solution


alternative (𝑺+
𝒊 )
Mango 0,16
Rambutan 0,32
Guava 0,20
Crystal Guava 0,21
Starfruit 0,22
Jackfruit 0,30
Soursop 0,32

Base the table we have that mango is closest distance from ideal positive solution,
otherwise soursop and rambutan have furthest distance from ideal positive solution.

Tabel 6 Separation of Ideal Negatif Solution


Alternative (𝑺−
𝒊 )
Mango 0,26
Rambutan 0,02
Guava 0,14
Crystal Guava 0,14
Starfruit 0,21
Jackfruit 0,04
Soursop 0,01

Base the table we have that mango is furthest distance from ideal negetive solution, otherwise
soursop have closest distance from ideal negative solution.

6
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

Preference Value

Tabel 7 Preference Value


Alternative (𝑪+
𝒊 )
Mango 0,62
Rambutan 0,07
Guava 0,40
Crystal Guava 0,40
Starfruit 0,49
Jackfruit 0,11
Soursop 0,04

Base the table founded that each alternative gets a preference value, where the

alternative has the greatest preference value is the best alternative to be cultivated and the

alternative has the smallest preference value is the worst alternative to be cultivated. It seen

from the above table that mango is the biggest preference value, in other words mango is the

best solution and soursop is the lowest preference value, it means soursop is the worst

solution.

CONCLUSION

The horticultural crop has the closest distance to ideal positive solution is a mango
with a distance of 0.16 and has the furthest distance with ideal positive solution is a soursop
with a distance of 0.32. While the horticultural crops have the furthest distance with the ideal
negative solution negative is mango with a distance of 0.26 and has the closest distance with
the ideal negetive solution is soursop with a distance of 0.01.

The largest preference value obtained by mango with a value of 0.62, followed by starfruit
with a value of 0.49, guava with a value of 0.40, guava crystal with a value of 0.40, jackfruit
with a value of 0.11, rambutan with a value of 0.07 and soursop with a value of 0.04. So that
mango is the best horticultural crop that is right for cultivation.

7
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

REFERENCE

[1] Olson, D.L. 2004. Comparison of Weights in TOPSIS Models. USA:


University of Nebraska.
[2] Dou, H., S.Kadarsah, Ariawan. 2004. Development of Decision Support System Based
on Category Management Concept to Increase Sales Performance of a Category
as a Business Intelegence Tool. In International symposium on Competitive
Intelligence by University of Aix. Jakarta, Indonesia. Juni 23-24.
[3] Figueira, J., S. Greco, and M Ehrgott.eds. 2005. Multiple Criteria Decision Analysis:
State of the Art Surveys. USA: Springer.
[4] Tzeng, G.H., J. Huang. 2011. Multiple Attribute Decision Making Methods and
Application. USA: CRC Press.
[5] Kahraman, C. 2008. Fuzzy Multi-Criteria Decision Making Theory and Application
with Recent Developments. Turkey: Istanbul Technical University.
[6] Anton, Howard.1994. Elementary Linear Algebra. USA: Drexel University.

8
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

PROOFMENT METHODE :

a. PROOF THAT 𝝀 ≇ 𝒏 !

ANSWER :

Assume that weighted matrix 𝑊 is Pairwise Comparison Matriks (PCM)

where 𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛

𝑤 𝑤1 𝑤1
𝑤1 1 𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛
⋮ ⋮
𝑤𝑖 𝑤 𝑤𝑖 𝑤𝑖
𝑊= 𝑖 ⋯ ⋯
𝑤1 𝑤𝑗 𝑤𝑛
⋮ ⋮ ⋮ ⋮
𝑤𝑛 𝑤𝑛 ⋯
𝑤𝑛 𝑤𝑛
𝑤1 𝑤𝑗 ⋯ 𝑤𝑛

Assume that 𝐴 is Pairwise Comparison Matriks (PCM) dan 𝜆 is eigen value from 𝐴

correlated with eigen value 𝑤 …………… (a)

𝑤1

where 𝑤 = 𝑤𝑖 .

𝑤𝑛

Because 𝐴 is PCM, we can write 𝐴 as :

𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛

𝑤 1 𝑤1 ′ 𝑤 𝑤1 ′ 𝑤1 ′
1
⋯ 𝑤𝑗 ⋯ 𝑤𝑛

⋮ ⋮
𝑤𝑖 ′ 𝑤𝑖 ′ 𝑤𝑖 ′
𝐴 = 𝑤𝑖 𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛

⋮ ⋮ ⋮

𝑤𝑛 𝑤𝑛 ′ 𝑤𝑛 ′ 𝑤𝑛 ′
𝑤 1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛

9
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

1
Assume 𝐴 is consistence PCM, as like 𝑎𝑖𝑗 entry 𝑖𝑗 from 𝐴 that 𝑎𝑖𝑗 ≅ 𝑎 or
𝑗𝑖

𝑤𝑖 ′ 1
𝑤𝑗 ≅ 𝑤 𝑗 ′ .
𝑤𝑖

Will proofed 𝜆 ≅ 𝑛 where 𝑛 is dimension from matrix 𝐴.

𝑤𝑖 ′ 1 𝑤𝑖 ′ 𝑤𝑖 𝑤 𝑖 ′𝑤 𝑗 ′
Because 𝑤𝑗 ≅ 𝑤 𝑗 ′ ⟷ 𝑤𝑗 = 𝑤𝑗 ′
↔ 𝑤𝑖𝑤𝑗
=1
𝑤𝑖

Attention:

𝑤1 ′ 𝑤1 ′ 𝑤1 ′
𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛 𝑤1 𝑛𝑤1 ′ 𝑤1 ′
⋮ ⋮ ⋮

𝑤𝑖 ′ 𝑤𝑖 ′ 𝑤𝑖 ′ 𝑤𝑗 𝑛𝑤𝑗 ′ 𝑤𝑗 ′
𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛 .= =𝑛 .
⋮ ⋮ ⋮
⋮ ⋮ ⋮
𝑤𝑛 ′ 𝑤𝑛 ′ 𝑤𝑛 ′
𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛 𝑤𝑛 𝑛𝑤𝑛 ′ 𝑤𝑛 ′

𝑤1 ′ 𝑤1
⋮ ⋮

𝑤𝑗 ′ 𝑤𝑗
𝑛 .= 𝜆 . ( base premis (a) )
⋮ ⋮

𝑤𝑛 ′ 𝑤𝑛

than

𝑛𝑤𝑖′ = 𝜆𝑤𝑖 ……………. (i)

𝑛𝑤𝑗′ = 𝜆𝑤𝑗 ……………..(ii)

From equation (i) and (ii) we have

𝑛2 𝑤𝑖′ 𝑤𝑗′ = 𝜆2 𝑤𝑖 𝑤𝑗

10
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

𝑛2 𝑤𝑖 ′𝑤𝑗 ′
= ≅1
𝜆2 𝑤𝑖 ′𝑤𝑗

Than we have 𝑛2 ≅ 𝜆2 if only if 𝑛 ≅ 𝜆.


1
Next,if 𝑛 ≇ 𝜆 than 𝑎𝑖𝑗 ≇ 𝑎 and PCM 𝐴 not consistence. ∎
𝑗𝑖

b. PROOF ANALITYCAL EIGEN VALUE

Answer :

Assume 𝑊𝑥 = 𝑛𝑥 with scalar 𝑛 is eigen value with corresponding to eigen vector 𝑥

where matrix W:

𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛
𝑤1 𝑤 𝑤1 𝑤1
1
𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛
⋮ ⋮ ⋮ ⋮
𝑤𝑖 𝑤𝑖 𝑤𝑖 𝑤𝑖
𝑊= 𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛
⋮ ⋮ ⋮ ⋮
𝑤𝑛 𝑤𝑛 ⋯
𝑤𝑛 𝑤𝑛
𝑤1 𝑤𝑗 ⋯ 𝑤𝑛

Attention :

a. Normalizing PCM

𝑤1 𝑤1 𝑤1
𝑤1 𝑤𝑗 𝑤𝑛
𝑛 ⋯ 𝑛 ⋯ 𝑛
𝑊
𝑘=1 𝑘 𝑊
𝑘=1 𝑘 𝑊
𝑘=1 𝑘 𝑤1 𝑤1 𝑤1
𝑤1 𝑤𝑗 𝑤𝑛 𝑛
⋯ 𝑛 ⋯ 𝑛
⋮ ⋮ 𝑘=1 𝑊𝑘 𝑘=1 𝑊𝑘 𝑘=1 𝑊𝑘

𝑤𝑖 𝑤𝑖 𝑤𝑖 ⋮ ⋮ ⋮
𝑤1 𝑤𝑗 𝑤𝑛 𝑤𝑖 𝑤𝑖 𝑤𝑖
⋯ 𝑛 ⋯ 𝑛 = ⋯ 𝑛 ⋯ 𝑛
𝑛
𝑘=1 𝑊𝑘
𝑊
𝑘=1 𝑘 𝑊
𝑘=1 𝑘
𝑛
𝑘=1 𝑊𝑘 𝑘=1 𝑊𝑘 𝑘=1 𝑊𝑘
𝑤1 𝑤𝑗 𝑤𝑛
⋮ ⋮ ⋮
⋮ ⋮ ⋮ 𝑤𝑛 𝑤𝑛 𝑤𝑛
𝑤𝑛 𝑤𝑛 𝑤𝑛 𝑛 ⋯ 𝑛 ⋯ 𝑛
𝑤1 𝑤𝑗 𝑤𝑛 𝑘=1 𝑊𝑘 𝑘=1 𝑊𝑘 𝑘=1 𝑊𝑘
𝑛 ⋯ 𝑛 ⋯ 𝑛
𝑊
𝑘=1 𝑘 𝑊
𝑘=1 𝑘 𝑊
𝑘=1 𝑘
𝑤1 𝑤𝑗 𝑤𝑛

11
Selecting of Horticulture Plants Most Exact for Cultivation using TOPSIS Methode
NATIONAL JOURNAL CONTEST 2014

b. Calculate Average of PCM

𝑤1 𝑤1 𝑤1 𝑤1
𝑤1
𝑛 + ⋯ + 𝑛 + ⋯ + 𝑛 𝑛 𝑛 𝑊 𝑛
𝑘=1 𝑊 𝑘
𝑘 =1 𝑊 𝑘 𝑘 =1 𝑊 𝑘 𝑘 =1 𝑊 𝑘 𝑘=1 𝑘
⋮ ⋮ ⋮ ⋮ ⋮
1 𝑤𝑖 𝑤𝑖 𝑤𝑖 1 𝑤𝑖 𝑤𝑖
+ ⋯ + 𝑛 + ⋯ + 𝑛 = 𝑛 𝑛 𝑊 = 𝑛
𝑘 =1 𝑊 𝑘 𝑘 =1 𝑊 𝑘
𝑛
𝑛 𝑘 =1 𝑊 𝑘 𝑛 𝑘=1 𝑘 𝑘=1 𝑊 𝑘
⋮ ⋮ ⋮ ⋮ ⋮
𝑤𝑛 + ⋯ 𝑤𝑛 + 𝑤𝑛 𝑤𝑛 𝑤𝑛
𝑛 + 𝑛 ⋯ + 𝑛 𝑛 𝑛 𝑊
𝑘 =1 𝑊 𝑘 𝑘 =1 𝑊 𝑘 𝑘 =1 𝑊 𝑘 𝑘=1 𝑘 𝑛
𝑘=1 𝑊 𝑘

c. Calculate Weight Sum Vector (PCM 𝐴 × 𝑎𝑣𝑒𝑟𝑎𝑔𝑒)

𝑤1 𝑤12 𝑤1 𝑤𝑗 𝑤1 𝑤𝑛
𝑤1 ⋯ 𝑛 ⋯
𝑤1 ⋯ 𝑤1 𝑛 𝑤1 𝑛𝑘=1 𝑊𝑘 𝑤𝑗 𝑘=1 𝑊𝑘 𝑤𝑛 𝑛
𝑘=1 𝑊𝑘
𝑤1 𝑤𝑗 ⋯ 𝑤𝑛 𝑘=1 𝑊𝑘
⋮ ⋮ ⋮ ⋮ ⋮
⋮ ⋮ 𝑤𝑖 𝑤𝑗 𝑤𝑖 𝑤𝑛
𝑤𝑖 𝑤𝑖 𝑤𝑖 𝑤𝑖 𝑤𝑖 𝑤1
𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛 𝑛 = ⋯ 𝑛 ⋯ 𝑛
𝑘=1 𝑊𝑘 𝑤1 𝑛𝑘=1 𝑊𝑘 𝑤𝑗 𝑘=1 𝑊𝑘 𝑤𝑛 𝑘=1 𝑊𝑘
⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮
𝑤𝑛 𝑤𝑛 𝑤𝑛 𝑤𝑛 𝑤𝑛 𝑤1 𝑤𝑛 𝑤𝑗
𝑤1 ⋯ 𝑤𝑗 ⋯ 𝑤𝑛 𝑤𝑛2
𝑛 ⋯ ⋯
𝑘=1 𝑊𝑘 𝑤1 𝑛𝑘=1 𝑊𝑘 𝑤𝑗 𝑛
𝑘=1 𝑊𝑘 𝑤𝑛 𝑛
𝑘=1 𝑊𝑘

𝑤1
𝑛 𝑛
𝑘=1 𝑊𝑘

𝑤𝑖
= 𝑛 𝑛
𝑘=1 𝑊𝑘

𝑤𝑛
𝑛 𝑛
𝑘=1 𝑊𝑘

d. Calculate Consistency Vector ( Weight Sum Vector . average inverse)


𝑛
𝑤1 𝑘=1 𝑊𝑘 𝑛
𝑛 𝑛 .
𝑘=1 𝑊𝑘 𝑤1
⋮ ⋮
𝑛
𝑤𝑖 𝑘=1 𝑊𝑘
𝑛 𝑛 . = 𝑛
𝑘=1 𝑊𝑘 𝑤𝑖
⋮ ⋮
𝑛
𝑤𝑛 𝑘=1 𝑊𝑘
𝑛 𝑛 . 𝑛
𝑘=1 𝑊𝑘 𝑤𝑛

𝑐𝑜𝑛𝑠𝑖𝑠𝑡𝑒𝑛𝑐𝑦 𝑣𝑒𝑐𝑡𝑜𝑟
e. Calculate 𝜆 𝑛

𝑛𝑛
𝜆= = 𝑛 , So 𝜆 = 𝑛 , Where n is eigen value from 𝑊.∎
𝑛

12

You might also like