2. Consider the system S in Figure 9.7, which has three input parameters X, Y, and Z.

Assume that set D, a set of

input test data values, has been selected for each of the input variables such that D(X)={True, False}, D(Y)={0,5},
and D(Z)={P,Q,R}. Using the orthogonal array method discussed in this chapter, generate pairwise test cases for
this system. Compare the results with the test suite generated using the IPO algorithm.
Orthogonal array method / Taguchi method
1. The total number of possible test cases is 2x2x3 = 12 test cases. The Orthogonal array method generates
six test cases
2. Maximum number of independent input variable = 3, this will map to the factors of the array
3. Maximum number of values that each independent variable will take = 3, this will map to the levels of the
array
4. A suitable orthogonal array with the smallest number of runs L6 (33). The array has six rows, three levels for
the value, and three factors for the variables
Runs
1
2
3
4
5
6

1
1
1
2
2
3
3

Factors
2
1
2
3
1
2
3

3
1
2
3
2
3
1

5. Map the variables to the factors and values to the levels of the array
Factors
1

Variables Levels of the array Values

X
1
True
2
(left-over)
3
False
2
Y
1
0
2
(left-over)
3
5
3
Z
1
P
2
Q
3
R
6. There are left-over levels in the array that are not being mapped. Mapping after filling in the remaining
levels using the cycling technique
Test Case
ID
TC1
TC2
TC3
TC4
TC5
TC6

Orthogonal Array after Mapping

Factor
X
Y
Z
True
0
P
True
2
Q
2
5
R
2
0
Q
False
2
R
False
5
P

Test Case
ID

Generates Test Cases after

Mapping Left-Over Levels
X
Y
Z
TC1
True
0
P
TC2
True
0
Q
TC3
True
5
R
TC4
False
0
Q
TC5
False
5
R
TC6
False
5
P
7. Generate nine test cases taking the test case values from each run. The result:
Each X is tested with every Y and with every Z
Each Y is tested with every X and with every Z
Each Z is tested with every X and with every Y
In Parameter Order (IPO algorithm)
1. Initialization phase, test cases are generated to cover two input variables
2. Horizontal growth phase, the existing test cases are extended with the values of the other input variables
3. Vertical growth phase, additional test cases are created such that the test suite satisfied pairwise coverage
for the values of the new variables
(
(
[
(
(

(
(

)
)
]
)
)

)
)

(
[ (

(
(

)
)

)
)
(
(

(
(

)
)

)
)
(
(

)
) ]

Z has three values, q = 3 and |T| = 4 q = 3. So:

(
(
[
(
(

)
)
]
)
)

(
(

)
(

[ (

)
(

) ]

Select one of P, Q, and R for (False, 5). Choose that the extended test can cover most missing pairs:
(
(
[
(
(

)
)
]
)
)

(
(
[

(
(

)
)

Make the test in T to covered the four test in

(
(
(
(
(
[(

)
)
)
)
)
)]

)
)
]

. The union T T generate the six pairwise test cases as follows:

Result Orthogonal array algorithm

(
(
(
(
(
[(

)
)
)
)
)
)]

Result IPO algorithm

(
(
(
(
(
[(

)
)
)
)
)
)]

Dalam test suite tersebut terdapat sedikit perbedaan dalam hal test case, hal tersebut disebabkan kedua algoritma
mempunyai cara yang berbeda dalam mengenerate test case.
Algoritma Orthogonal array menghasilkan tes suite yang memiliki pemerataan semua kombinasi berpasangan.
Dengan hanya menguji kombinasi semua pasangan dari input variabel yang dipilih, sehingga menghasilkan test case
lebih sedikit (tidak seperti pendekatan semua kombinasi)
Algoritma IPO menghasilkan test suite yang memenuhi pairwise coverage untuk nilai-nilai dari dua parameter
pertama, lalu test suite diperpanjang oleh algoritma untuk memenuhi pairwise coverage untuk nilai-nilai parameter
ketiga, dan begitu seterusnya untuk nilai-nilai dari masing-masing parameter tambahan sampai semua parameter
termasuk dalam test suite.