User Guide J-Electre-v1.

0
An ELECTRE I, I_s, I_v, II, III, IV, TRI
and TRI ME software.

Valdecy Pereira, Helder Gomes Costa and Livia Dias de Oliveira Nepomuceno
Contents
1- J-Electre-v1.0 - Installation Notes .........................................................................................2
2- J-Electre-v1.0 - First Use .......................................................................................................3
2.1 Electre I ..............................................................................................................................3
2.2 Electre I_s ...........................................................................................................................5
2.3 Electre I_v ...........................................................................................................................8
2.4 Electre II ...........................................................................................................................10
2.5 Electre III ..........................................................................................................................12
2.6 Electre IV ..........................................................................................................................14
2.7 Electre TRI ........................................................................................................................15
2.8 Electre TRI ME ..................................................................................................................17
3- References..........................................................................................................................20
 1- J-Electre-v1.0 - Installation Notes

The J-Electre-v1.0 (jar) is a runnable .jar file that does not need to be installed and it run in any
OS. The unique requisite is the need to have the latest Java SE program installed. Check if your
computer has the latest release, if not please download it (preferably Java SE 7 or superior) at:
www.oracle.com
or www.oracle.com/technetwork/pt/java/javase/downloads/index.html?ssSourceSiteId=otnes
The J-Electre-v1.0 (exe) is a runnable .exe file that does not need to be installed and also do not
need the installation of the latest Java SE program.

Download J-Electre-v1.0 at: https://sourceforge.net/projects/j-electre/files/

 2- J-Electre-v1.0 - First Use

In the main screen choose the Electre Method between the following options:

a) Electre I
b) Electre I_s
c) Electre I_v
d) Electre II
e) Electre III
f) Electre IV
g) Electre TRI
h) Electre TRI ME (Multi-Evaluator)

2.1 Electre I

To explain how to use the J-Electre-v1.0 in order to solve Electre I problems, the following
example will be used:

g1 g2 g3 g4
W 2 1 5 3
a1 150 1 20 3000
a2 300 0 10 0
a3 250 0 10 2250
a4 110 1 20 2800
a5 120 1 50 1000

This problem has 5 alternatives (a1, a2, a3, a4, a5) and 4 criteria (g1, g2, g3, g4). The weights
(importance) of each criterion is represented by the W row.
After Choosing Electre I method, two parameters need to be set: d (discordance index varying
from 0 to 1) and c (concordance index varying from 0 to 1). For the given example d = 0.6 and
c = 0.7.

Insert the number of Alternatives (varying from 2 to 1,000), Criteria (varying from 2 to 1,000)
and press the Matrix button to build the performance matrix.
Insert the values in the performance matrix and then press the Solve button to solve the
problem. To save (export) the results to a spreadsheet press the Save button.

The output contains: Concordance Matrix, Discordance Matrix, Credibility Matrix, Kernel (set
of alternatives that are not dominated) and Dominated (set of alternatives that are dominated
by the alternatives in the Kernel set).

2.2 Electre I_s

To explain how to use the J-Electre-v1.0 in order to solve Electre I_s problems, the following
example (ROY & SKALKA, 1985) will be used:

g1 g2 g3 g4 g5 g6 g7
Q 2000 2 1 1 1 50 0.1
P 3000 5 2 3 2 82 0.2
V 3500 7 3 5 6 90 0.5
W 0.3 0.1 0.3 0.1 0.2 0.2 0.1
a1 16000 201 8 40 5 378 31.3
a2 18000 199 8 35 5 474 33.0
a3 16000 195 8 36 1 480 33.9
a4 18000 199 8 35 5 430 33.1
a5 17000 191 8 34 1 430 34.4
a6 17000 199 8 35 4 494 32.0
a7 15000 194 8 37 3 452 33.8
a8 18000 200 8 36 6 475 33.8
a9 17000 209 7 37 3 440 30.9

This problem has 9 alternatives (a1, a2, a3, a4, a5, a6, a7, a8, a9) and 7 criteria (g1, g2, g3, g4,
g5, g6, g7). The Q row represents the weak preference, the P row represents the strong
preference, the V row represents the Veto (respecting: ) and finally, the weights
(importance) of each criterion is represented by the W row.

After Choosing Electre I_s method, two parameters need to be set: L (lambda index varying
from 0.5 to 1) and the maximum number of Cycles (varying from 0 to 9,000) that will be
removed.

Cycles invalidate the solution obtained by the Electre I_s algorithm, and in order to deal with
cycles we used the Johnson Algorithm (JOHNSON, 1975), implemented by Meyer (2012) and
modified by us, that can find all cycles in a directed graph. Then we remove each found cycle in
order to have a valid solution.

However, if a problem has to many cycles, as a rule of thumb above 30, consider first increasing
the value of L and if there still too many cycles, consider removing an alternative (or alternatives
one at a time) that appears frequently in most cycles (all cycles are indicated in the output
table and detecting them should not be difficult).

For the given example L = 0.7 and Cycles = 30.

Insert the number of Alternatives (varying from 2 to 1,000), Criteria (varying from 2 to 1,000)
and press the Matrix button to build the performance matrix.

Insert the values in the performance matrix and then press the Solve button to solve the
problem. To save (export) the results to a spreadsheet press the Save button.

The output contains: Concordance Matrix, Discordance Matrix, Credibility Matrix, Cycles (first
cycle: 2 a4 2; second cycle: 3 7 3 and the third cycle: 3 8 7 3),
Kernel (set of alternatives that are not dominated after the cycles are removed) and Dominated
(set of alternatives that are dominated by the alternatives in the Kernel set after the cycles are
removed).
2.3 Electre I_v

To explain how to use the J-Electre-v1.0 in order to solve Electre I_v problems, the following
example will be used:

g1 g2 g3 g4
V 2 2 2 2
W 7 3 5 6
a1 15 9 6 10
a2 10 5 7 8
a3 22 12 1 14
a4 31 10 6 18
a5 8 9 0 9

This problem has 5 alternatives (a1, a2, a3, a4, a5) and 4 criteria (g1, g2, g3, g4). The V row
represents the Veto and the weights (importance) of each criterion is represented by the W row.

After Choosing Electre I_v method, two parameters need to be set: d (discordance index with
only two values 0 or 1) and c (concordance index varying from 0 to 1). For the given example
d = 1 and c = 0.65.
Insert the number of Alternatives (varying from 2 to 1,000), Criteria (varying from 2 to 1,000)
and press the Matrix button to build the performance matrix.

Insert the values in the performance matrix and then press the Solve button to solve the
problem. To save (export) the results to a spreadsheet press the Save button.

The output contains: Concordance Matrix, Discordance Matrix, Credibility Matrix, Kernel (set
of alternatives that are not dominated) and Dominated (set of alternatives that are dominated
by the alternatives in the Kernel set).
2.4 Electre II

To explain how to use the J-Electre-v1.0 in order to solve Electre II problems, the following
example (WANG & TRIANTAPHYLLOU, 2006) will be used:

g1 g2 g3 g4 g5 g6 g7
W 0.078 0.118 0.157 0.314 0.235 0.039 0.059
a1 1 2 1 5 2 2 4
a2 3 5 3 5 3 3 3
a3 3 5 3 5 3 2 2
a4 1 2 2 5 1 1 1
a5 1 1 3 5 4 1 5

This problem has 5 alternatives (a1, a2, a3, a4, a5) and 7 criteria (g1, g2, g3, g4, g5, g6, g7). The
weights (importance) of each criterion is represented by the W row.

After Choosing Electre II method, six parameters need to be set: three levels of concordance
index c-, c and c+ (where 0.5 + 1), two levels of concordance d- and d+ (where
0 + 1), and the maximum number of Cycles (varying from 0 to 9,000) that will be
removed.

Cycles invalidate the solution obtained by the Electre II algorithm, and in order to deal with
cycles we used the Johnson Algorithm (JOHNSON, 1975), implemented by Meyer (2012) and
modified by us, that can find all cycles in a directed graph. Then we remove each found cycle in
order to have a valid solution.

However, if a problem has to many cycles, as a rule of thumb above 30, consider first increasing
the parameters values and if there still too many cycles, consider removing an alternative (or
alternatives one at a time) that appears frequently in most cycles (all cycles are indicated in
the output table and detecting them should not be difficult).

For the given example c- = 0.65, c = 0.75, c+ = 0.85, d- = 0.25, d+ = 0.5 and Cycles = 30.
Insert the number of Alternatives (varying from 2 to 1,000), Criteria (varying from 2 to 1,000)
and press the Matrix button to build the performance matrix.
Insert the values in the performance matrix and then press the Solve button to solve the
problem. To save (export) the results to a spreadsheet press the Save button.

The output contains: Concordance Matrix, Discordance Matrix, Credibility Matrix, Cycles Ss
(cycles from the strong graph), Cycles Ws (cycles from the weak graph), Ranking Ascending
(from the worst alternative to the best), Ranking Descending (from the best alternative to the
worst), Ranking Average (In order to have a complete ranking, we choose to build the final
ranking as an average between the ascending and descending ranking) and Dominance Matrix
(To build the classical pre-ordination or final ranking, the dominance matrix is provided)

2.5 Electre III

To explain how to use the J-Electre-v1.0 in order to solve Electre III problems, the following
example will be used:

g1 g2 g3 g4
Q 0.3 0.3 0.3 0.3
P 0.5 0.5 0.5 0.5
V 0.7 0.7 0.7 0.7
W 0.2754741 0.2735455 0.1758277 0.2221151
a1 8.84 8.79 6.43 6.95
a2 8.57 8.51 5.47 6.91
a3 7.76 7.75 5.34 8.76
a4 7.97 9.12 5.93 8.09
a5 9.03 8.97 8.19 8.1
a6 7.41 7.87 6.77 7.23

This problem has 6 alternatives (a1, a2, a3, a4, a5, a6) and 4 criteria (g1, g2, g3, g4). The Q row
represents the weak preference (as constants), the P row represents the strong preference (as
constants), the V row represents the Veto (respecting: ) and finally, the weights
(importance) of each criterion is represented by the W row.

Insert the number of Alternatives (varying from 2 to 1,000), Criteria (varying from 2 to 1,000)
and press the Matrix button to build the performance matrix.

Insert the values in the performance matrix and then press the Solve button to solve the
problem. To save (export) the results to a spreadsheet press the Save button.

The output contains: Concordance Matrix, Discordance Matrix (one for each criterion),
Credibility Matrix, Ranking Ascending (from the worst alternative to the best), Ranking
Descending (from the best alternative to the worst), Ranking Average (In order to have a
complete ranking, we choose to build the final ranking as an average between the ascending
and descending ranking) and Dominance Matrix (To build the classical pre-ordination or final
ranking, the dominance matrix is provided)

2.6 Electre IV

To explain how to use the J-Electre-v1.0 in order to solve Electre IV problems, the following
example will be used:

g1 g2 g3 g4 g5 g6 g7 g8
Q 10 10 10 10 10 10 10 10
P 20 20 20 20 20 20 20 20
V 100 100 100 100 100 100 100 100
a1 15 80 60 30 60 50 60 70
a2 25 0 40 30 40 40 50 140
a3 25 0 50 30 40 40 50 140
a4 25 0 50 30 50 40 70 140
a5 25 0 50 30 50 40 50 140
a6 15 20 50 30 50 60 60 100
a7 15 80 50 50 40 90 60 100

This problem has 7 alternatives (a1, a2, a3, a4, a5, a6, a7) and 8 criteria (g1, g2, g3, g4, g5, g6,
g7, g8). The Q row represents the weak preference, the P row represents the strong preference,
the V row represents the Veto (respecting: ).

Insert the number of Alternatives (varying from 2 to 1,000), Criteria (varying from 2 to 1,000)
and press the Matrix button to build the performance matrix.
Insert the values in the performance matrix and then press the Solve button to solve the
problem. To save (export) the results to a spreadsheet press the Save button.

The output contains: Credibility Matrix, Ranking Ascending (from the worst alternative to the
best), Ranking Descending (from the best alternative to the worst), Ranking Average (In order
to have a complete ranking, we choose to build the final ranking as an average between the
ascending and descending ranking) and Dominance Matrix (To build the classical pre-ordination
or final ranking, the dominance matrix is provided)

2.7 Electre TRI

To explain how to use the J-Electre-v1.0 in order to solve Electre TRI problems, the following
example (MOUSSEAU, 1999.) will be used:

g1 g2 g3 g4 g5
b2 70 75 80 75 85
b1 50 48 55 55 60
Q 5 5 5 5 5
P 10 10 10 10 10
V 30 30 30 30 30
W 1 1 1 1 1
a1 75 67 85 82 90
a2 28 35 70 90 95
a3 45 60 55 68 60

This problem has 3 alternatives (a1, a2, a3, a4, a5) and 4 criteria (g1, g2, g3, g4). The bn rows
represents the profiles (respecting: ). The Q row represents the weak preference, the
P row represents the strong preference, the V row represents the Veto (respecting: )
and finally, the weights (importance) of each criterion is represented by the W row.
After Choosing Electre TRI method, two parameters need to be set: Classes (the total number
of classes varying from 2 to 100) and Lambda (cut-off level varying from 0.5 to 1). For the
given example Classes = 3 (hence we have two profiles: b2 and b1) and Lambda = 0.75.

Insert the number of Alternatives (varying from 2 to 1,000), Criteria (varying from 2 to 1,000)
and press the Matrix button to build the performance matrix.
Insert the values in the performance matrix and then press the Solve button to solve the
problem. To save (export) the results to a spreadsheet press the Save button.

The output contains: Concordance c(ai;bh) (global concordance between alternative i and
profile h), Concordance c(bh;ai) (global concordance between profile h and alternative i),
Discordance d(ai;bh) (global discordance between alternative i and profile h), Discordance
d(bh;ai) (global discordance between profile h and alternative i), Credibility Matrix,
Classifcation Pessimist (from the upper profile bn to b1, Class A > B > C), Classification
Optimist (from the lower profile b1 to bn, Class A > B > C).

2.8 Electre TRI ME

To explain how to use the J-Electre-v1.0 in order to solve Electre TRI ME problems (developed
by Livia Dias Nepomuceno and Helder Gomes Costa), the following example will be used:

EV1(g1) EV1(g2) EV1(g3) EV2(g1) EV2(g2) EV2(g3)

b2 2 3 3 3 3 2
b1 1 2 2 1 2 1
Q 0 0 0 0 0 0
P 0 0 0 0 0 0
V 5 5 5 5 5 5
W 0.1 0.2 0.2 0.3 0.3 0.1
a1 2 2 1 3 1 3
a2 0 2 1 0 0 0
a3 1 3 3 2 1 1

This problem has 3 alternatives (a1, a2, a3), 3 criteria (g1, g2, g3, g4) and two Evaluators (EV1,
EV2). Each evaluator judges the same set of criteria and have their own set of weights (which
may be the same). The bn rows represents the profiles (respecting: ). The Q row
represents the weak preference, the P row represents the strong preference, the V row
represents the Veto (respecting: ) and finally, the weights (importance) of each
criterion is represented by the W row.

After Choosing Electre TRI ME method, three parameters need to be set: Classes (the total
number of classes varying from 2 to 100), Evaluators (the total number of evaluators, judges
or decision makers varying from 2 to 100) and Lambda (cut-off level varying from 0.5 to 1).
For the given example Classes = 3 (hence we have two profiles: b2 and b1), Evaluators = 2 and
Lambda = 0.75.

Insert the number of Alternatives (varying from 2 to 1,000), Criteria (varying from 2 to 1,000)
and press the Matrix button to build the performance matrix.
Insert the values in the performance matrix and then press the Solve button to solve the
problem. To save (export) the results to a spreadsheet press the Save button.

The output contains: Concordance c(ai;bh) (global concordance between alternative i and
profile h), Concordance c(bh;ai) (global concordance between profile h and alternative i),
Discordance d(ai;bh) (global discordance between alternative i and profile h), Discordance
d(bh;ai) (global discordance between profile h and alternative i), Credibility Matrix,
Classifcation Pessimist (from the upper profile bn to b1, Class A > B > C), Classification
Optimist (from the lower profile b1 to bn, Class A > B > C).
 3- References

JOHNSON, D. B. (1975). Finding All the Elementary Circuits of a Directed Graph. SIAM Journal
on Computing 4, no. 1, 77-84. http://dx.doi.org/10.1137/0204007

MEYER, F. (2012). Johnson Algorithm Implementation: <http://www.normalisiert.de> or

<https://github.com/josch/cycles_johnson_meyer>

MOUSSEAU, V.; SLOWINSKI, R.; ZIELNIEWICZ, P. (1999). ELECTRE TRI 2.0a: Methodological
Guide and User's Manual. Documents du LAMSADE, n 111. University Paris - Dauphine.

ROY, B. M.; SKALKA, J. (1985). ELECTRE IS: Aspcts Methodologiques et Guide d'utilization.
Cahier du LAMSADE. Universit de ParisDauphine.

WANG, X.; TRIANTAPHYLLOU, E. (2006). Ranking Irregularities When Evaluating Alternatives by

Using Some ELECTRE Methods, Omega, Vol. 36, No. 1, pp. 45-63.