Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
11
,
No.
1
,
Febr
uar
y
2021
, pp.
575
~
588
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v11
i
1
.
pp575
-
588
575
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om
Target
-
b
ased test
path pri
or
iti
zati
on for
UM
L
acti
vity dia
gram
usin
g weight a
ssi
gnment
methods
Wala
ip
orn S
or
nkli
ang
1
,
Thi
maporn
Phe
tkaew
2
1
Mana
gement
of
Inform
at
ion
Tec
hnolog
y
Progra
m
,
School
of
In
f
orm
at
ic
s
,
W
alailak
Univer
si
t
y
,
T
hai
l
and
2
School
of
In
for
m
at
ic
s
,
W
alailak
Univer
sit
y
,
Th
a
il
and
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
pr
9
, 2
0
20
Re
vised
Jun
11
,
20
20
Accepte
d
J
un
28
, 20
20
The
ben
efit
of
e
xplora
tor
y
te
stin
g
and
ad
ho
c
t
est
ing
b
y
te
st
er’
s
e
xper
ie
n
ce
is
tha
t
cru
c
ia
l
bu
gs
are
found
q
uic
kl
y
.
Regr
ession
te
sting
an
d
te
st
ca
se
priori
tiza
ti
on
ar
e
important
pro
ce
ss
es
of
software
t
esti
ng
whe
n
software
func
ti
ons
h
ave
b
ee
n
cha
ng
ed.
W
e
propose
a
t
est
pat
h
pr
iori
t
iz
a
ti
o
n
m
et
hod
t
o
gene
ra
te
a
sequ
e
nce
of
te
st
p
at
hs
tha
t
would
m
atch
the
t
este
rs’
i
n
te
rests
and
foc
uses
on
th
e
target
area
of
inter
est
or
on
th
e
changed
ar
ea.
W
e
g
ene
ra
te
te
st
pat
hs
form
the
ac
ti
vi
t
y
d
ia
gr
ams
and
surve
y
the
te
st
pat
h
prior
it
i
za
t
ion
from
te
sters.
W
e
def
i
ne
node
and
edg
e
weigh
t
to
the
sy
m
bo
ls
of
ac
t
ivit
y
dia
g
rams
b
y
apply
i
ng
Ti
m
e
m
ana
gement
,
Pare
to,
Buffe
t
t,
Bina
r
y
,
and
Bip
ola
r
m
et
hod
.
The
n
we
propos
e
a
te
st
p
at
h
sco
re
equ
at
ion
to
p
riori
tize
te
st
p
aths
.
W
e
al
so
propose
eva
luati
on
m
et
hods
i.
e.
,
the
diffe
r
ence
an
d
the
sim
il
ari
t
y
of
te
st
pat
h
priori
tiza
ti
on
to
te
sters’
in
te
r
es
ts.
Our
proposed
m
et
hod
had
the
l
ea
st
av
era
g
e
of
the
diff
ere
n
c
e
and
th
e
m
ost
ave
rag
e
of
th
e
sim
il
arit
y
co
m
par
e
with
the
t
este
r’s
prio
rit
izat
ion
of
t
est
pat
hs.
The
Bip
ola
r
m
et
hod
wa
s
the
m
ost
suita
ble
for
assigning
weigh
ts
to
m
at
ch
te
st
pat
h
ran
k
b
y
the
t
e
ster
.
Our
proposed
m
et
hod
al
so
ha
s
give
n
the
af
fe
ct
ed
p
at
h
b
y
ch
angi
ng
ar
ea
highe
r
pr
iori
t
y
t
han
th
e
o
the
r tes
t
pa
th.
Ke
yw
or
d
s
:
Ed
ge wei
ght as
sign
m
ent
Node wei
ght a
ssign
m
ent
Re
gr
essi
on test
ing
Targ
et
-
base
d p
rior
it
iz
at
ion
Test
p
at
h p
rio
r
it
iz
at
ion
UML aci
t
ivit
y diag
ram
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
B
Y
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
Thim
apo
rn P
he
tkaew,
School
of In
for
m
at
ic
s,
Walai
la
k Un
i
ve
rsity
,
222 Thai
buri,
Thasala
Di
stric
t, Na
khonSiT
ham
m
arat, 8
0160, T
haila
nd.
Em
a
il
: pthi
m
a
po
@
wu.ac.th
,
t
him
apo
rn.p@g
m
ai
l.co
m
1.
INTROD
U
CTION
T
est
case
pr
i
ori
ti
zat
ion
by
im
po
rtance
is
i
nten
ded
to
rea
rr
a
ng
e
the
te
st
cases
t
o
s
uit
the
ne
eds
of
so
ft
war
e
te
sti
ng
.
Tho
se
te
st
c
ases
that
are
th
e
m
os
t
i
m
po
rta
nt
will
be
te
ste
d
pr
i
or
to
the
le
ss
im
po
rtant
cases.
Test
case
pr
i
ori
ti
zat
ion
is
very
us
ef
ul
wh
e
n
faced
with
li
m
it
ed
reso
urce
s
or
li
m
it
ed
testing
tim
e.
Test
case
pr
i
or
it
iz
at
ion
c
ou
l
d
ra
nk
te
st
cases
base
d
on
i
m
po
rtant
obje
ct
ives,
i.e.
fa
ci
li
ta
te
ta
rg
et
area
or
m
od
ific
at
ion
-
awar
e
te
sti
ng.
Sever
al
studi
es
br
oa
dly
cat
egorized
the
te
st
case
pri
or
it
iz
at
ion
into
co
ver
a
ge
-
ba
sed,
requirem
ents
-
ba
sed,
risk
-
bas
ed,
sea
rc
h
-
ba
sed,
fa
ult
-
bas
ed,
histo
ry
-
ba
sed,
a
nd
othe
rs
e.
g.,
c
os
t
-
awa
re
base
d
[1
-
3]
.
Cov
e
ra
ge
-
base
d
te
st
case
pr
i
ori
ti
zat
ion
is
a
te
sti
ng
m
et
ho
d
that
e
xam
ines
the
co
de
dir
ect
ly
.
The
pr
im
ary
m
easur
es
i.e.
,
co
de,
sta
te
m
ent,
br
a
nch,
a
nd
f
un
ct
io
n
c
ov
e
ra
ge
we
re
the
m
os
t
wide
ly
us
ed
crit
eria
[1,
3
-
5]
.
Re
qu
irem
ents
-
base
d
te
st
case
pr
i
or
it
iz
at
ion
us
es
t
he
re
qu
i
rem
ents
info
rm
at
ion
to
cl
assif
y
serio
us
te
st
cas
es
[
1].
It
is
to
c
heck
w
hethe
r
t
he
delivere
d
s
oft
war
e
m
eet
s
t
he
needs of
t
he
cust
om
er
or
not
[
6].
Ri
sk
-
based
te
s
t
case
pr
io
riti
zat
ion
pri
ori
ti
zes
the
te
st
cases
that
co
nc
ern
t
he
m
ajo
r
risks
that
a
ffec
te
d
the
so
ft
war
e
[
1].
Searc
h
-
bas
ed
te
st
case
pr
i
or
it
iz
at
ion
f
ind
s
the
opti
m
al
ran
king
of
the
te
st
cases
by
searchi
ng
f
rom
the
gl
ob
al
sp
ace
to
fit
the
ob
j
ect
ives,
su
c
h
as
gr
ee
dy,
gen
et
ic
al
gorithm
,
ant
colo
ny
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
11
, No
.
1,
Febr
uar
y
2021
:
57
5
-
588
576
op
ti
m
iz
ation
[
1,
2,
7].
S
om
e
stud
ie
s
us
ed
a
c
on
ce
pt
of
hy
bri
dizin
g
A
*
al
gorithm
and
ant
colo
ny
op
ti
m
i
zat
ion
to
ge
ner
at
e
a
nd
opti
m
iz
e
the
te
st
paths
[
8].
Fault
-
base
d
te
st
case
pr
io
ri
ti
zat
ion
seq
ue
nces
the
te
st
c
ase
to
detect
ta
rg
et
e
d
fau
lt
s
[
1,
9
]
. Hist
or
y
-
based
te
st
case
pr
i
or
it
iz
at
ion
u
ses
the h
ist
o
ry d
at
a,
f
or
exam
ple,
the co
unt
of exec
utio
ns
wh
ic
h reveale
d a fa
ult t
o pr
i
ori
ti
ze the test
case [
1].
Object
-
Or
ie
nte
d
te
ch
no
l
og
y
is
cur
re
ntly
w
idely
us
ed
in
dev
el
op
i
ng
s
of
t
war
e.
T
he
con
ce
pt
of
Object
-
Or
ie
nte
d
program
m
ing
s
uch
as
data
de
pende
nce,
c
on
t
ro
l
d
e
pende
nce,
a
nd
dep
e
ndency
due
to
obj
ect
relat
ion
s
are
use
d
f
or
te
st
case
sel
ect
ion
and
te
st
case
pri
or
it
iz
at
ion
[
10]
.
The
unifie
d
m
od
el
ing
la
ngua
ge
(U
ML
)
is
a
stan
da
r
d
m
od
el
us
ed
to
desi
gn
obj
ect
-
or
ie
nted
program
m
ing
[4
]
.
Se
ver
al
stud
ie
s
hav
e
em
plo
y
e
d
a
U
ML
m
od
el
to
gen
e
rate
an
d
pri
or
it
iz
e
te
st
paths
by
co
nverti
ng
act
ivit
y
diag
ram
in
to
a
tree
structure
[11],
or
a
gra
ph
str
uc
ture
[5,
12
-
17
]
and
t
hen
ge
ne
rati
ng
the
te
st
paths
.
S
om
e
stud
ie
s
ha
ve
c
onve
rted
a
se
quence
diag
ram
into
a
gr
a
ph
str
uctu
r
e
[18,
19]
an
d
then
generate
d
te
st
paths.
Som
e
hav
e
c
onve
rted
a
sta
te
m
a
chine
diag
ram
[2
0,
21
]
into
a
gr
a
ph
st
ru
ct
ur
e
a
nd
t
hen
gen
e
r
at
ed
te
st
paths
.
Furthe
r,
UM
L
m
od
el
s
with
sever
al
diag
ram
s
hav
e
been
c
onve
rted
an
d
integ
rat
ed
into
th
e
sa
m
e
con
trol
flo
w
grap
h
to
c
re
at
e
and
pri
ori
ti
ze
te
st
paths
.
For
exa
m
ple,
by
us
ing
a
sequ
e
nce
di
agr
am
and
inte
racti
on
ov
e
r
view
diag
ram
s
[2
2],
us
in
g
a
seq
uen
c
e
diag
ram
and
a
sta
te
cha
rt
diagr
am
[23],
a
nd
us
i
ng
a
us
e
-
c
ase
dia
gr
am
,
a
seq
ue
nce
diag
ram
,
and
an
ac
ti
vity
diag
ram
[2
4].
In
te
st
cas
e
pri
or
it
iz
at
ion,
s
ever
al
st
ud
ie
s
ha
ve
assig
ne
d
weig
hts
f
or
act
ivit
y
diagra
m
s
or
sy
m
bo
ls o
f
c
ontrol f
l
ow grap
h t
o
cal
culat
e th
e sco
res fo
r
te
s
t paths, an
d
t
he
scores
wer
e t
he
n used t
o p
rio
riti
ze
the
te
st
path
s.
Ther
e
are
t
wo
gro
up
s
of
a
ppr
oach
e
s
to
as
sign
wei
gh
ts
t
o
t
he
sym
bo
ls
of
act
ivit
y
diagr
a
m
s
or
con
t
ro
l
flo
w
gr
a
phs.
In
the
first
gro
up,
t
he
weig
hts
ar
e
assig
ned
to
nodes
or
sy
m
bo
ls
of
t
he
gr
a
ph.
Each
sym
bo
l
has
a
diff
ere
nt
weig
ht.
All
the
sy
m
bo
l
weigh
ts
aff
ect
the
te
st
path
score
[15,
16]
.
The
weig
ht
c
al
culat
ion
of
a
co
ntr
ol
fl
ow
gr
a
ph
node
de
r
ives
f
r
om
the
nu
m
ber
s
of
in
c
om
ing
ed
ges
to
th
e
node
m
ulti
plied
by
the
c
ount
of
outg
oing
e
dge
s
from
the
no
de
[
13
]
. In
the
s
econd grou
p,
t
he
weig
hts
ar
e assign
e
d
t
o
th
e
ed
ges
of
the
gra
ph.
So
m
e
s
tud
ie
s
hav
e
assig
ne
d
the
weig
hts
to
the
ou
tg
oi
ng
e
dg
es
of
decisi
on
node
s
[12]
.
All
the
weigh
t
s
are
cal
culat
ed
to
fin
d
the
te
st
path
score
and
the
sc
or
e
s
are
us
e
d
to
pr
i
ori
ti
ze
the
te
st
paths
by
it
s i
m
po
rtance
.
None
of
the
a
bove
st
ud
ie
s
pri
or
it
iz
ed
by
t
he
im
po
rtance
of
the
t
ar
get
area
or
of
the
change
d
a
rea
.
Be
sides,
they
did
not
put
an
e
m
ph
asi
s
on
t
he
seq
uen
ce
acc
ordin
g
to
the
t
est
ers
f
or
their
te
st
paths
of
in
te
rest.
Althou
gh
s
pec
ia
l
value
te
sti
ng
or
ad
ho
c
te
sti
ng
[
25]
can
be
us
e
d
to
te
st
the
ta
rg
et
ar
ea
or
c
hange
d
area,
these
ty
pes
of
te
sti
ng
are
on
ly
do
ne
by
pr
of
essi
onal
s
in
so
ft
war
e
te
sti
ng.
T
hese
pr
of
e
ssion
al
s
pu
t
a
lot
of
effor
t
into
deliveri
ng
a
ppr
opr
ia
te
te
st
resu
lts
that
su
it
the
custom
er’
s
re
quirem
ents.
So
,
this
stud
y
pro
po
s
es
a
te
st
path
pri
or
it
iz
at
ion
m
eth
od
t
o
gen
e
rat
e
a
se
qu
e
nce
of
te
st
pat
hs
t
ha
t
would
m
at
ch
the
te
ste
rs
’
i
nt
erests
and f
oc
us
es
on
the tar
get a
rea
of intere
st o
r o
n
the
ch
a
nged
area.
The
rest
of
thi
s
pa
per
is
di
vid
ed
as
f
ollo
ws
.
Sect
io
n
2
ex
plains
t
he
rese
arch
m
et
hod
use
d
in
th
i
s
stud
y.
Ex
pe
rim
en
ta
l
resu
lt
s
and
disc
us
si
on
are
presente
d
in
s
ect
i
on
3.
En
han
cem
ent
of
t
he
al
gorith
m
s
are
pr
ese
nted
in s
e
ct
ion
4.
Finall
y, s
ect
io
n 5 c
oncl
ud
es
the
pa
pe
r.
Re
la
te
d
w
ork
.
Test
case
pri
or
it
iz
at
ion
is
a
re
orde
rin
g
of
t
he
te
st
cas
es
in
a
te
st
s
uite
f
or
te
s
t
execu
ti
ons
.
Te
st
case
pr
i
or
it
iz
at
ion
ai
m
s
to
cov
e
r
a
ny
giv
e
n
par
t
of
t
he
pro
gr
am
by:
det
ect
ing
fau
lt
s
e
arli
er,
fin
ding
high
pri
or
it
y
er
r
or
s
,
f
ocusi
ng
on
te
s
ti
ng
of
the
la
st
m
od
ifie
d
par
t,
an
d
reducin
g
the
ti
m
e
and
c
os
t
of
te
sti
ng
.
Se
veral
stu
dies
have
exam
ined
m
et
ho
ds
to
pr
ioriti
ze
te
st
pa
ths
t
o
in
crea
se
the
ef
fici
ency
a
nd
eff
ect
ive
ness o
f
te
sti
ng [1,
26
]
.
UML
act
ivit
y
diag
ram
is
essenti
al
ly
a
flow
char
t
that
ch
ron
ologica
ll
y
or
ga
nizes
a
set
of
a
ct
ivit
ie
s
that
ta
ke
place
ov
e
r
tim
e.
The
dia
gr
am
sy
m
bo
ls
consi
st
of
i
niti
al
,
act
ivit
y,
transiti
on,
decisi
on,
m
erg
e,
fork,
j
oi
n,
swim
la
ne,
and
final.
A
UML
act
ivit
y
diag
ram
is
transform
ed
into
a
graph
str
uctu
re
or
a
tree
str
uc
ture
t
o
gen
e
rate
an
d
pr
i
or
it
iz
e
te
st
paths
.
I
n
te
s
t
case
pr
io
riti
zat
ion
,
w
ei
ght
s
are
num
eric
values
as
sig
ned
t
o
the
sym
bo
ls
of
a
UML
act
ivit
y
diag
ram
or
t
o
t
he
sym
bo
ls
of
a
c
ontrol
flo
w
gr
a
ph.
Test
case
pri
ori
ti
zat
ion
is
then
done
acco
rd
i
ng
to
th
e
w
ei
gh
ts
to
deter
m
ine
wh
ic
h
pa
ths
shou
l
d
be
t
est
ed
first
[
16]
.
The
wei
ghts
are
set
accor
ding
to
t
he
sym
bo
ls
in
the
UML
act
ivit
y
diagr
am
.
The
UML
act
i
vity
diagr
am
will
be
c
onve
r
te
d
into
a
directed
gr
a
ph
:
G
={
A,
E,
in,
F}
,
w
her
e
A
represe
nts
the
no
des
co
ns
i
sti
ng
of
act
ion
nodes,
obj
ect
nodes
,
and
c
ontr
ol
nodes
i.e.,
decisi
on,
m
erg
e,
f
ork,
a
nd
join
node
s,
E
re
pr
es
e
nt
s
the
ed
ges
,
E
={
(
x,
y
)
|
x,
y
∈
A
}
,
in
is
an
init
ia
l
node,
and
F
is
a
fi
na
l
node
[
4]
.
T
he
re
are
t
hr
ee
a
ppr
oach
es
t
o
a
ssign
i
ng
weig
ht
s
to
the
sym
bo
ls
of
act
ivit
y
diagr
a
m
s
or
to
sym
b
ols
of
co
ntr
ol
flo
w
gr
a
ph.
Al
l
weig
hts
a
re
cal
culat
ed
to
f
ind
out
the
te
s
t
pat
h
scor
e
s.
T
he
sc
or
es
we
re
us
e
d
to
pri
ori
ti
ze
the
te
st
pat
hs
by
im
po
rtance.
First,
t
he
wei
gh
ts
are
assig
ned
t
o
nodes
of
a
grap
h
[
5,
13
-
18,
23,
24
]
.
S
econd,
the
w
ei
gh
ts
are
ass
ign
e
d
to
ed
ge
s
of
a
gr
a
ph
[1
2].
Thir
d,
t
he
wei
gh
ts
a
re
assi
gned
t
o
bo
t
h
node
s
an
d
e
dge
s
of
a
gra
ph
[
11,
22]
.
Accor
ding
to
s
om
e
repor
ts
,
the
sym
bo
ls
in
UML
act
ivi
ty
diagr
am
or
in
c
ontrol
fl
ow
gra
ph
we
r
e
assig
ne
d
we
igh
ts
as
in
Ta
ble
1,
and the calc
ula
ti
on
m
et
ho
ds o
f
te
st pat
h
sc
ores are
sho
wn in Ta
ble
2.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ta
r
get
-
based
test
path prio
rit
izati
on for
U
M
L
activ
it
y
...
(
Walai
po
r
n S
ornkl
iang)
577
Table
1.
C
om
par
iso
n of wei
ght assi
gnm
ent m
et
ho
ds
Na
m
e
Sy
m
b
o
l
Ass
ig
n
ed
weigh
ts
Initial n
o
d
e
1
)
weig
h
t
=0
[
1
5
]
Activ
ity
no
d
e
1
)
weig
h
t
=1
[
5
,
1
1
,
1
5
,
1
6
]
2
)
weig
h
t
=4
,
if
the n
o
d
e was
in
ter
-
d
e
p
en
d
ed
activity
[
1
6
]
3
)
weig
h
t
=6
,
if
the n
o
d
e was c
o
n
trol
n
o
d
e
[
1
6
]
No
r
m
al
edg
e
1
)
weig
h
t=1
[
1
2
]
2
)
weig
h
t=2
[
2
2
]
Edg
e pas
sed
the
d
ecisio
n
node
1
)
weig
h
t=to
tal we
ig
h
t of
all
o
u
tg
o
in
g
edg
es
m
u
st
b
e 1
[
1
2
]
2
)
b
ased
on
80
/2
0
rule [
2
2
]
ed
g
e weigh
t=4
f
o
r
th
e tr
u
e edg
e of
th
e decisi
o
n
no
d
e
e
d
g
e weigh
t=1
f
o
r
th
e f
alse edg
e of
the d
ecisio
n
no
d
e
Decisio
n
no
d
e
1
)
weig
h
t=2
[
1
1
,
1
6
]
2
)
weig
h
t=4
[
5
,
1
5
]
Mer
g
e no
d
e
1
)
weig
h
t=2
[
1
1
]
2
)
weig
h
t=3
[5
,
1
5
],
Fo
rk
-
jo
in
no
d
e
1
)
f
o
rk weig
h
t=2
,
jo
in
weigh
t=2
[
5
,
1
5
]
2
)
f
o
rk weig
h
t=3
,
jo
in
weigh
t=3
[
1
1
]
3
)
f
o
rk weig
h
t=5
,
jo
in
weigh
t=3
[
1
6
]
Fin
al no
d
e
1
)
weig
h
t=0
[
1
5
]
Table
2
.
C
om
par
iso
n of cal
cu
la
ti
on
m
et
ho
ds f
or
test
path
sc
or
es
W
e
i
g
h
t
s
c
o
r
e
Calcu
latio
n
m
eth
o
d
s f
o
r
test p
ath
sco
re
N
o
d
e
1
)
su
m
o
f
all
n
o
d
e
weig
h
ts,
wh
ere
th
e
weig
h
ts
o
f
th
e
n
o
d
es
were
ass
ig
n
ed
accord
in
g
to
th
e
ty
p
e
o
f
sym
b
o
ls
(see
Table
1
)
[
5
,
1
5
,
16]
2
)
s
u
m
o
f
all
n
o
d
e
weig
h
ts,
wh
er
e
th
e
weig
h
t
o
f
th
e
f
irst
n
o
d
e
eq
u
als
1
an
d
th
e
weig
h
t
o
f
th
e
n
ex
t
n
o
d
e
will
h
av
e
1
add
ed
un
til the last n
o
d
e
[
1
8
,
2
3
,
2
4
]
3
)
s
u
m
o
f
all
n
o
d
e
weig
h
ts,
w
h
e
r
e
t
h
e
w
e
i
g
h
t
s
o
f
t
h
e
n
o
d
e
s
w
e
r
e
a
s
s
i
g
n
e
d
u
s
i
n
g
S
t
a
c
k
-
b
a
s
e
d
w
e
i
g
h
t
a
p
p
r
o
a
c
h
a
n
d
t
h
e
B
a
s
i
c
I
F
M
o
d
e
l
[
1
4
,
1
7
]
4)
th
e
n
u
m
b
er
o
f
n
o
d
es+s
u
m
o
f
n
o
d
e
weig
h
ts+th
e
n
u
m
b
er
o
f
p
redicate
n
o
d
es+th
e
n
u
m
b
er
o
f
lo
g
ical
co
n
d
itio
n
n
o
d
es
[
1
3
]
E
d
g
e
1
)
su
m
o
f
all
ed
g
e
weig
h
ts,
w
h
e
r
e
t
h
e
w
e
i
g
h
t
s
o
f
t
h
e
e
d
g
e
s
w
e
r
e
a
s
s
i
g
n
e
d
accord
in
g
to
th
e
ty
p
e
o
f
ed
g
es
(see
Table
1
)
[1
2]
N
o
d
e
a
n
d
e
d
g
e
1
)
su
m
o
f
all
n
o
d
e
weig
h
ts+s
u
m
o
f
all
ed
g
e
weig
h
ts,
wh
ere
th
e
weig
h
ts
o
f
th
e
n
o
d
es
were
ass
ig
n
ed
accord
in
g
to
th
e
ty
p
e
o
f
sym
b
o
ls
(see
Table
1
)
an
d
th
e
weig
h
ts
o
f
th
e
ed
g
es
were
ass
ig
n
ed
b
y
th
e
n
u
m
b
er
o
f
in
co
m
in
g
d
ep
en
d
en
cies
o
f
p
redeces
so
r
n
o
d
e
m
u
ltip
lied
b
y
th
e
n
u
m
b
er
o
f
o
u
tg
o
in
g
d
ep
en
d
en
cies
o
f
th
e
su
ccesso
r
n
o
d
e [
1
1
]
2
)
s
u
m
o
f
a
l
l
n
o
d
e
w
e
i
g
h
t
s
+
su
m
o
f
all
ed
g
e
weig
h
ts
,
w
h
e
r
e
t
h
e
weig
h
ts
o
f
th
e
n
o
d
es
were
ass
ig
n
ed
b
y
u
sin
g
b
ackward
slices
ap
p
roach
an
d
t
h
e
w
e
i
g
h
t
s
o
f
t
h
e
e
d
g
e
s
w
e
r
e
a
s
s
i
g
n
e
d
accord
in
g
to
th
e
ty
p
e
o
f
ed
g
es
(see
Table
1
)
[
2
2
]
Kau
r
et
al
.
[1
3]
pr
op
os
ed
the
pr
ioriti
zat
ion
of
te
st
paths
descen
ded
fr
om
UML
act
ivit
y
diagr
am
by
us
ing
com
plexity
o
f t
he path
as
fo
ll
ow
s.
First,
the
act
ivit
y diagr
am
w
as
transf
or
m
ed
into
a
con
t
ro
l
flow
gr
ap
h.
Secon
d,
the
te
st
paths
wer
e
gen
erated
fr
om
the
con
trol
flow
g
rap
h
us
ing
the
dep
th
first
search.
W
hen
it
fo
un
d
fo
rk
no
des,
it
wo
uld
sel
ect
on
e
rep
resen
ta
ti
ve
path
by
us
ing
br
eadth
first
search
to
get
un
iqu
e
ind
epen
den
t
paths.
Third
,
assignin
g
a
weigh
t
to
each
no
de
IF(N
)
was
by
the
nu
m
ber
of
incom
ing
edg
es
to
the
no
de
FANIN(
N)
m
ulti
plied b
y t
he
nu
m
ber
o
f
ou
tgo
ing
ed
ges
fr
om
that no
de
FANOUT
(
N)
, u
sing
(
1)
.
Each test
path
scor
e
W
p
was
fo
un
d
by
su
m
m
ing
no
de
weigh
ts
W
i
in
the
set
T
p
and
n
was
the
nu
m
ber
of
no
des,
us
ing
(2
).
Fo
ur
th,
the co
m
plexity
C
of
test
p
at
h
was
cal
culat
ed
us
ing
(
3)
by g
at
her
ing
the
values
of
the
nu
m
ber
o
f
no
des
(N
p
)
,
the
scor
e
of
te
st
path
(W
p
)
,
the
nu
m
ber
of
pr
edicat
e
no
des
(P
p
)
,
and
the
nu
m
ber
of
log
ic
al
con
diti
on
s
(C
p
)
.
And
finall
y,
the
te
st
paths
wo
uld
be
pr
ioriti
zed
fr
om
the
hig
hest
path
com
plexity
to
the
lowest
path
com
plexity
.
)
(
)
(
)
(
N
x
F
A
N
O
U
T
N
F
A
N
I
N
N
IF
(1)
p
n
i
i
p
T
P
W
W
1
(2)
p
p
p
p
C
P
W
N
C
(3)
Wang
et
al
.
[
15]
propose
d
te
st
path
pr
i
or
it
iz
at
ion
de
rive
d
f
ro
m
UML
act
ivit
y
diagr
am
as
fo
ll
ows
.
First,
the
act
ivit
y
diagr
am
was
transfor
m
ed
into
a
co
ntro
l
fl
ow
gr
a
ph.
Sec
ond,
the
te
st
pa
ths
wer
e
ge
ne
rated
by
the
de
pth
fi
rst
search
.
Thi
r
d,
the
weig
ht
of
each
no
de
wa
s
assigne
d
bas
ed
on
the
ty
pe
of
sym
bo
ls
in
UML
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
11
, No
.
1,
Febr
uar
y
2021
:
57
5
-
588
578
act
ivit
y
diagr
a
m
.
If
the
node
was
a
n
i
niti
al
no
de
or
a
n
e
nd
node,
it
was
as
sign
e
d
0.
I
f
the
node
wa
s
a
n
a
ct
ivit
y
node,
it
w
as as
sign
e
d
4. If
the
n
ode w
as a fork nod
e
or
a
j
oi
n
node
, it was assign
e
d
2. I
f
the nod
e
was
a
m
erg
e
node,
it
was
as
sign
e
d
3.
N
e
xt,
eac
h
te
st path
scor
e
f(
x
)
was
the
s
um
of
no
de
weig
hts
W
i
on
that
te
st path
and
n
was
t
he
nu
m
ber
of
node
s
i
n
the
c
hrom
os
om
e
C
in
th
e
popula
ti
on
P
,
us
in
g
(4)
.
Finall
y,
this
r
esearch
pr
i
or
it
iz
ed
the
te
st
paths
fro
m
the
highest
scor
e
.
I
f
tw
o
t
est
paths
ha
d
t
he
sa
m
e
scor
e,
it
would
rand
om
l
y
sel
ect
the test
pat
h
pri
ori
ty
o
r
de
r betwee
n
the
se.
P
C
W
x
f
n
i
i
1
)
(
(4)
Ma
hali
and
Ach
arya
[1
6]
pr
op
os
ed
the
pr
ioriti
zat
ion
of
te
st
paths
us
ing
UML
act
ivit
y
diagr
am
and
evo
luti
on
ary
al
go
rithm
as
fo
ll
ow
s.
First,
the
ac
ti
vity
diagr
am
was
transf
or
m
ed
into
an
act
ivit
y
gr
aph
and
al
l
te
st
paths
wer
e
create
d.
Secon
d,
each
no
de
was
assigned
a
weigh
t
based
on
the
sym
bo
ls
of
UML
act
ivit
y
diagr
am
.
If
the
no
de
was
a
no
rm
al
act
ivit
y
no
de,
it
was
assigned
1.
If
the
no
de
was
a
decisi
on
no
de,
it
was
assigned
2.
If
the
no
de
was
an
act
ivit
y
of
on
e
thread
dep
end
s
on
each
oth
er,
it
was
assigned
3.
If
the
al
l
act
ivit
y
no
de
was
within
the
decisi
on
sym
bo
l
or
the
fo
rk
-
j
oin
sym
bo
l,
it
was
assigned
4.
If
the
no
de
was
a
fo
rk
no
de,
it
was
assigne
d
5.
If
the
no
de
was
a
con
trol
no
de,
it
was
assigned
6.
Third
,
the
te
st
path
scor
e
F(
x)
was
the
su
m
of
no
de
weigh
ts
C
ri
in
that
te
st
path,
us
ing
(5
)
.
This
research
pr
ioriti
zed
the
im
po
rtance
of
te
st paths
fr
om
the h
igh
est
to
the low
est
p
at
h
scor
e.
n
i
ri
C
x
F
1
)
(
(5)
Jena
et
al
.
[2
2]
pr
op
os
ed
te
st
path
gen
erati
on
and
pr
ioriti
zat
ion
as
fo
ll
ow
s.
First,
the
interact
ion
ov
erv
ie
w
diagr
am
was
con
ver
te
d
into
an
interact
ion
gr
aph
.
The
sequ
ence
diagr
am
was
transf
or
m
ed
into
m
essage
sequ
ence
dep
end
ency
gr
ap
h
.
An
d
then,
the
two
gr
aph
s
wer
e
com
bin
ed
into
a
sequ
ence
interact
ion
gr
aph
.
Secon
d,
al
l
po
ssible
te
st
paths
wer
e
create
d
fr
om
the
gr
aph
by
us
ing
dep
th
first
search
.
Third
,
the
weigh
ts
wer
e
set
to
each
no
de
of
the
gr
aph
accord
ing
to
the
nu
m
ber
of
no
des
af
fected
by
the
cur
ren
t
no
de,
determ
ined
with
back
war
d
sli
ci
ng
.
The
weigh
t
of
no
rm
al
edg
es
was
2.
The
edg
e
fr
om
a
decisi
on
no
de
was
assigned
by
us
ing
80
/20
ru
le
and
weigh
t
4
was
assigned
to
the
true
valued
edg
e
(8
0%)
wh
il
e
weigh
t
1
was
assigned
to
the
fals
e
valued
edg
e
(2
0%)
.
Fo
ur
th,
the
scor
e
of
a
path
was
the
su
m
of
no
de
weigh
ts
plu
s
the
su
m
of
edg
e
weigh
ts
al
on
g
the
te
st
path,
us
ing
(6
).
Finall
y,
this
stud
y
pr
ioriti
zed
the
te
st
paths
fr
om
the h
igh
est
sco
re to
the low
est
sco
re.
n
i
n
j
i
i
e
d
g
e
w
e
i
g
h
t
n
o
d
e
w
e
i
g
h
t
p
a
t
h
S
c
o
r
e
1
1
)
(
)
(
)
(
(6
)
2.
RESEA
R
CH MET
HO
D
This
stu
dy
ai
m
ed
to p
ri
or
it
iz
e
the
im
po
rtanc
e
of
t
he
te
st
pat
hs
acc
ordin
g
to
the
ta
rg
et
a
rea
of
i
nterest,
the ch
a
nged
ar
ea, and t
he
te
st
p
at
hs
pr
io
riti
zed
by the test
e
r
s.
T
he
ste
ps
a
r
e as foll
ows.
2
.
1.
De
fine
the
criteri
a t
o
se
le
ct t
he
UML
act
i
vity dia
gr
am
The
UML
act
ivit
y
diagr
am
s
wer
e
sel
ect
ed
accor
ding
to
the
f
ollow
i
ng
c
rite
ria:
the
UML
act
ivit
y
diag
ram
s
that
wer
e
fr
e
quentl
y
us
ed
i
n
rese
arch
st
ud
ie
s
on
s
of
twa
re
te
st
ing
,
t
he
UM
L
act
ivit
y
diagr
a
m
s
that
wer
e
us
e
d
in
daily
li
ves,
the
U
ML
act
ivit
y
diag
ram
s
that
wer
e
ea
sy
to
unde
rstan
d,
a
nd
the
UML
ac
ti
vity
diag
ram
s
that
had
va
rio
us
c
on
t
ro
l
fl
ows
a
ff
ect
in
g
the
te
st
path
sel
ect
ion.
T
he
sel
ect
ion
of
UML
a
ct
ivit
y
diag
ram
s
dep
end
e
d
on
the
co
ntr
ol
structu
re.
Each
UML
ac
ti
vity
diagr
am
sel
ect
ed
for
this
stud
y
ha
d
dif
fer
e
nt
con
t
ro
l
str
uctu
res.
As
a
res
ul
t,
the
sel
ect
ed
four
UML
act
ivit
y
diagr
am
s
for
the
e
xp
e
ri
m
ent
wer
e
as
f
ollows
i.e., A
TM W
it
hdra
w
act
ivit
y
diag
ram
[2
7], S
hippin
g
Orde
r
Syst
em
act
ivi
ty
diagr
am
[13
]
,
Buy
Be
ve
rage
from
Vendin
g
Ma
c
hi
n
e act
ivit
y dia
gr
am
[
11]
, a
nd Mi
le
age Purc
ha
se
W
e
b Po
rtal
acti
vity
d
ia
gr
a
m
[
28
]
.
The
A
TM
W
it
hdra
w
act
ivit
y
diag
ram
,
app
li
ed
f
ro
m
Bog
hdady
et
al
.
[27
]
,
com
pr
ise
s
m
any
ty
pes
of
con
t
ro
l
struct
ures
s
uc
h
as
sel
ect
ion
c
ontrol
structu
res
co
nsi
sti
ng
of
one
-
w
ay
sel
e
ct
ion
,
two
-
way
sel
ect
ion,
a
nd
nested
sel
ect
ion
;
an
d
a
one
-
w
ay
sel
ect
ion
co
ntr
ol
struct
ur
e
within
for
k
-
j
oi
n
str
uctu
res
as
show
n
in
Fig
ur
e
1.
The
S
hippin
g
Order
Syst
e
m
act
ivit
y
diagr
a
m
,
app
li
ed
fro
m
Kau
r
et
al
.
[13]
,
com
pr
ise
s
a
fork
-
j
oi
n
struct
ur
e
with
no
c
on
t
rol
struct
ur
e
insi
de,
sel
ect
ion
c
on
t
ro
l
struct
ure
co
ns
ist
in
g
of
two
-
way
sel
e
ct
ion
,
but
t
here
is
no
it
erati
on
co
ntr
ol
struct
ur
e
as
sh
ow
n
in
Fig
ur
e
2
.
The
Ve
nd
i
ng
Ma
c
hin
e
ac
ti
vity
diagr
am
,
app
li
ed
from
Sapna
and
M
ohanty
[
11
]
,
w
hich
c
om
pr
ise
s
two
it
erati
o
n
c
ontr
ol
s
tructu
res
with
pr
e
-
te
st
loop.
This
dia
gr
am
does
no
t
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ta
r
get
-
based
test
path prio
rit
izati
on for
U
M
L
activ
it
y
...
(
Walai
po
r
n S
ornkl
iang)
579
hav
e
the
f
ork
-
j
oi
n
struct
ur
e
as
sh
ow
n
in
Figure
3
.
T
he
Mi
le
age
Pu
rc
ha
se
act
ivit
y
diagr
am
,
app
li
ed
fr
om
Paiboo
nk
a
sem
su
t
an
d
Lim
piya
ko
r
n
[
28]
,
com
pr
ise
s
sel
ect
ion
co
ntr
ol
structu
res
c
onsist
ing
of
two
-
way
sel
ect
ion
a
nd
nested
sel
ect
io
n,
a
nd
the
re
is
a
fo
r
k
-
join
st
ru
ct
ur
e.
T
his
diag
ram
do
es
no
t
ha
ve
a
n
it
erati
on
con
t
ro
l st
ru
ct
ure
as s
how
n
in
Figure
4
.
Figure
1
.
A
TM
w
it
hdra
w
act
i
vity
d
ia
gram
Figure
2
.
S
hippin
g
o
r
der
s
yst
e
m
acti
vity
d
ia
gr
am
Figure
3
.
V
e
nd
ing
m
achine ac
ti
vity
d
ia
gr
am
Figure
4
.
Mi
le
age
p
urc
hase a
ct
ivit
y diagr
am
2
.
2.
Tr
ansf
or
m t
he
UML
ac
tivit
y
di
ag
r
am
to
an
ac
tivit
y flow
graph
The
four
act
ivit
y
diagr
am
s
wer
e
tra
ns
f
or
m
ed
to
a
n
act
ivit
y
flow
grap
h.
Fo
r
the
c
oncu
r
ren
ce
re
gion,
a
qu
e
sti
onnair
e
-
base
d
s
urve
y
fo
r
te
st
path
by
te
ste
rs'
interest
,
65
prof
e
ssion
al
s
of
t
wa
re
te
ste
rs
s
how
that
the
path
m
us
t
be
trave
rse
d
f
r
om
le
ft
-
to
-
rig
ht
.
Each
th
read
is
li
ste
d
fr
om
t
he
act
ivit
y
of
top
-
le
vel
to
act
ivit
y
of
low
le
vel acc
ordin
g
to
the
de
pt
h
fir
st searc
h m
et
ho
d.
The
ATM
Wi
thdraw
act
ivit
y
diag
ram
was
tran
sf
or
m
ed
into
an
act
iv
it
y
flow
grap
h,
w
hich
is
a
directed
grap
h.
Eac
h
node
r
epr
ese
nts
the
s
ym
bo
ls
of
act
ivit
y
diagr
am
a
nd
eac
h
ed
ge
r
epr
ese
nts
the
f
low
in
the
act
ivit
y
di
agr
am
.
The
ac
ti
vity
flow
gra
ph
of
ATM
W
it
hd
ra
w
act
ivit
y
diag
ram
is
dep
ic
te
d
i
n
Fi
gure
5
,
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
11
, No
.
1,
Febr
uar
y
2021
:
57
5
-
588
580
wh
e
re
I
is
t
he
init
ia
l
no
de
,
F
is
the
final
node,
di
is
a
deci
sion
node
,
fi
is
a
fork
node
,
a
nd
ji
is
a
join
node.
The
S
hippin
g
Order
Syst
e
m
act
ivity
dia
gr
am
,
the
Ven
di
ng
Ma
c
hi
ne
act
ivit
y
diagr
am
,
and
t
he
Mi
le
age
Pu
r
chase
acti
vi
ty
d
ia
gr
am
were
trans
f
or
m
ed
into a
n
act
i
vity
f
lo
w gr
a
ph, de
picte
d
in
Fi
gur
e
s
6
-
8 res
pecti
vely
.
Figure
5
.
A
TM
w
it
hdra
w
act
i
vity
f
lo
w gr
a
ph
Figure
6
.
S
hippin
g
o
r
der
s
yst
e
m
acti
vity
f
low gr
a
ph
Figure
7
.
V
e
nd
ing
m
achine ac
ti
vity
f
low gra
ph
Figure
8
.
Mi
le
age
p
urc
hase a
ct
ivit
y flow
gr
aph
2
.
3
.
Que
stion
na
ir
e
-
b
as
e
d
s
urve
y fo
r
t
es
t
p
ath
p
ri
oritiz
at
i
on
b
y
t
es
ter
s'
i
nt
erest
The
four
act
ivit
y
diagr
am
s
in
sect
ion
2.1
w
ere
use
d
t
o
ge
ner
at
e
te
st
pat
hs
in
e
ach
act
i
vity
diagr
am
.
The
n,
th
e
te
st
paths
we
re
gi
ve
n
to
65
prof
e
s
sion
al
s
of
t
war
e
te
ste
rs
to
pr
i
ori
ti
ze
the
m
by
im
po
rtance
acc
ordin
g
to
their
i
nteres
ts
to
create
a
baseli
ne
for
c
om
par
isons.
T
he
data
obta
in
ed
wer
e
a
naly
zed
for
pr
io
riti
zat
ion
fr
e
qu
e
nc
ie
s
of
the
te
st
paths.
The
survey
r
esults
on
the
pr
i
or
it
iz
at
ion
of
the
te
st
paths
in
ATM
withdra
w
,
Sh
ip
ping
orde
r
syst
e
m
,
Vend
ing
m
achine
,
a
nd
Mi
le
age
purch
ase
act
ivit
y
diag
ram
accord
i
ng
to
t
he
te
ste
rs’
interest
s
are
s
umm
arized
in
Table
s
3
-
6
res
pecti
vely
.
I
n
Table
3,
t
he
ATM
with
dr
a
w
act
ivit
y
diagr
am
has
gen
e
rated
14
t
est
paths
(
P1
to
P
14)
.
All
65
te
ste
rs
will
ass
ign
a
pat
h
(ro
w)
to
rank
t
he
i
m
po
rtance
of
the
te
st
accor
ding
t
o
their
inte
rest
by
arr
a
ng
i
ng
t
hem
in
order
1
-
14
(c
olu
m
ns
)
.
T
he
high
est
value
de
pic
te
d
i
n
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ta
r
get
-
based
test
path prio
rit
izati
on for
U
M
L
activ
it
y
...
(
Walai
po
r
n S
ornkl
iang)
581
bo
l
d
-
faced
i
n
each
col
um
n
sh
ows
t
hat
th
e
te
ste
rs
ha
ve
the
m
os
t
agr
ee
m
ent
in
that
order.
F
or
e
xam
ple,
m
os
t t
est
ers
agr
ee that
P5 sh
ould test
i
n
the
thir
d
ra
nk
of th
e test
p
at
hs.
Table
3
.
T
he
r
e
su
lt
s of t
he
s
urvey o
n
p
rior
it
i
zat
ion
of the
t
e
st
p
at
hs i
n AT
M
w
it
hdra
w
a
c
ti
vity
d
ia
gr
am
a
ccordin
g
t
o
th
e
t
est
ers’
i
ntere
sts
Test
p
ath
no
.
Test path
Ran
k
of
test p
ath
The f
requ
en
cy
dis
t
ribu
tio
n
of
r
an
k
in
g
1
2
3
4
5
6
7
8
9
10
11
12
13
14
P
1
I
-
1
-
2
-
d1
-
3
-
4
-
d2
-
5
-
f1
-
6
-
d3
-
7
-
j1
-
11
-
F
1
32
9
4
1
6
1
1
3
0
3
0
1
1
3
P
2
I
-
1
-
2
-
d1
-
2
-
d1
-
3
-
4
-
d2
-
5
-
f1
-
6
-
d3
-
7
-
j1
-
11
-
F
7
6
4
7
8
5
2
18
8
2
2
2
1
0
0
P
3
I
-
1
-
2
-
d1
-
3
-
4
-
d2
-
5
-
f1
-
6
-
d3
-
j1
-
11
-
F
2
7
19
7
3
2
7
4
4
1
2
6
0
3
0
P
4
I
-
1
-
2
-
d1
-
2
-
d1
-
3
-
4
-
d2
-
5
-
f1
-
6
-
d3
-
j1
-
11
-
F
8
0
3
5
8
2
5
9
16
5
4
2
3
2
1
P
5
I
-
1
-
2
-
d1
-
3
-
4
-
d2
-
8
-
d4
-
9
-
d5
-
5
-
f
1
-
6
-
d3
-
7
-
j1
-
11
-
F
3
4
3
18
9
8
3
5
2
6
3
1
2
1
0
P
6
I
-
1
-
2
-
d1
-
2
-
d1
-
3
-
4
-
d2
-
8
-
d4
-
9
-
d5
-
5
-
f1
-
6
-
d3
-
7
-
j1
-
11
-
F
9
6
0
5
3
2
7
3
5
16
11
2
3
1
1
P
7
I
-
1
-
2
-
d1
-
3
-
4
-
d2
-
8
-
d4
-
9
-
d5
-
5
-
f1
-
6
-
d3
-
j1
-
11
-
F
4
0
2
6
18
3
5
9
8
3
2
4
1
3
1
P
8
I
-
1
-
2
-
d1
-
2
-
d1
-
3
-
4
-
d2
-
8
-
d4
-
9
-
d5
-
5
-
f1
-
6
-
d3
-
j1
-
11
-
F
10
1
6
1
1
4
5
4
9
6
16
6
2
0
4
P
9
I
-
1
-
2
-
d1
-
3
-
4
-
d2
-
8
-
d4
-
9
-
d5
-
10
-
11
-
F
5
0
1
2
0
19
7
6
3
9
5
5
6
1
1
P
10
I
-
1
-
2
-
d1
-
2
-
d1
-
3
-
4
-
d2
-
8
-
d4
-
9
-
d5
-
10
-
11
-
F
11
1
1
3
3
3
1
2
3
7
8
21
6
6
0
P
11
I
-
1
-
2
-
d1
-
3
-
4
-
d2
-
8
-
d4
-
10
-
11
-
F
6
0
0
1
7
3
16
1
4
6
5
7
10
4
1
P
12
I
-
1
-
2
-
d1
-
2
-
d1
-
3
-
4
-
d2
-
8
-
d4
-
10
-
11
-
F
12
1
0
6
2
4
0
3
0
1
2
8
26
5
7
P
13
I
-
1
-
2
-
d1
-
10
-
11
-
F
13
8
6
2
0
3
1
3
0
0
1
1
1
30
9
P
14
I
-
1
-
2
-
d1
-
2
-
d1
-
10
-
11
-
F
14
4
6
3
1
3
2
0
0
1
0
3
0
6
36
Table
4
.
T
he
r
e
su
lt
s of t
he
s
urvey o
n p
rior
it
i
zat
ion
of the te
st paths
in
s
hi
ppin
g
o
r
der
s
yst
e
m
acti
vity
d
ia
gr
am
accor
ding t
o
th
e test
ers’
i
ntere
sts
Test path
no.
Test path
Ran
k
of
test p
ath
The f
requ
en
cy
dis
t
ribu
tio
n
o
f
r
an
k
in
g
1
2
3
4
5
6
7
P
1
I
-
1
-
2
-
d1
-
3
-
d2
-
15
-
F
6
8
10
7
3
1
23
13
P
2
I
-
1
-
2
-
d1
-
3
-
d2
-
f1
-
4
-
5
-
6
-
j1
-
d3
-
f2
-
7
-
8
-
j2
-
15
-
F
1
30
7
12
6
4
3
3
P
3
I
-
1
-
2
-
d1
-
3
-
d2
-
f1
-
4
-
5
-
6
-
j1
-
d3
-
14
-
15
-
F
3
1
10
27
10
10
5
2
P
4
I
-
1
-
2
-
d1
-
9
-
d4
-
10
-
f3
-
11
-
12
-
j3
-
13
-
3
-
d2
-
15
-
F
5
5
3
11
7
28
10
1
P
5
I
-
1
-
2
-
d1
-
9
-
d4
-
10
-
f3
-
11
-
12
-
j3
-
13
-
3
-
d2
-
f1
-
4
-
5
-
6
-
j1
-
d3
-
f2
-
7
-
8
-
j2
-
15
-
F
2
11
24
2
9
12
4
3
P
6
I
-
1
-
2
-
d1
-
9
-
d4
-
10
-
f3
-
11
-
12
-
j3
-
13
-
3
-
d2
-
f1
-
4
-
5
-
6
-
j1
-
d3
-
14
-
15
-
F
4
1
7
11
23
7
7
9
P
7
I
-
1
-
2
-
d1
-
9
-
d4
-
14
-
15
-
F
7
12
4
2
1
2
11
33
Table
5
.
T
he
r
e
su
lt
s of t
he
s
urvey o
n p
rior
it
i
zat
ion
of the te
st paths
in
v
e
ndin
g
m
achine a
ct
ivit
y diagr
am
accor
ding t
o
th
e test
ers’
i
nter
e
sts
Test path
no
.
Test path
Ran
k
of
test p
ath
The f
requ
en
cy
dis
t
ribu
tio
n
of
r
an
k
in
g
1
2
3
4
5
6
7
8
9
10
P
1
I
-
1
-
d1
-
8
-
F
10
10
1
3
1
2
1
2
0
2
43
P
2
I
-
1
-
d1
-
2
-
d2
-
8
-
F
9
1
12
1
6
3
4
2
4
29
3
P
3
I
-
1
-
d1
-
2
-
d2
-
3
-
4
-
5
-
8
-
F
1
34
5
6
3
4
5
3
2
1
2
P
4
I
-
1
-
d1
-
6
-
1
-
d1
-
8
-
F
8
1
5
5
2
8
4
1
29
9
1
P
5
I
-
1
-
d1
-
6
-
1
-
d1
-
2
-
d2
-
8
-
F
7
0
1
2
6
3
10
28
10
4
1
P
6
I
-
1
-
d1
-
2
-
d2
-
7
-
2
-
d2
-
8
-
F
5
0
5
10
7
20
7
7
3
6
0
P
7
I
-
1
-
d1
-
6
-
1
-
d1
-
2
-
d2
-
7
-
2
-
d2
-
8
-
F
6
0
2
1
9
9
21
10
6
4
3
P
8
I
-
1
-
d1
-
6
-
1
-
d1
-
2
-
d2
-
3
-
4
-
5
-
8
-
F
3
2
5
24
7
7
5
7
7
1
0
P
9
I
-
1
-
d1
-
2
-
d2
-
7
-
2
-
d2
-
3
-
4
-
5
-
8
-
F
2
9
25
10
3
5
1
1
1
6
4
P
10
I
-
1
-
d1
-
6
-
1
-
d1
-
2
-
d2
-
7
-
2
-
d2
-
3
-
4
-
5
-
8
-
F
4
9
6
2
20
4
9
3
2
2
8
Table
6
.
T
he
r
e
su
lt
s of t
he
s
urvey o
n p
rior
it
i
zat
ion
of the te
st paths
in
m
il
e
age
p
urc
hase a
ct
ivit
y diagr
am
accor
ding t
o
th
e test
ers’
i
nter
e
sts
Test path
no.
Test path
Ran
k
of
test p
ath
The f
requ
en
cy
dis
t
ribu
tio
n
of
r
an
k
in
g
1
2
3
4
5
6
7
8
9
10
P
1
I
-
1
-
2
-
d1
-
16
-
F
10
13
2
1
1
1
1
0
1
1
44
P
2
I
-
1
-
2
-
d1
-
3
-
d2
-
10
-
f1
-
11
-
12
-
13
-
14
-
j1
-
15
-
d6
-
4
-
5
-
d3
-
6
-
d4
-
7
-
d5
-
8
-
9
-
F
5
12
4
5
3
28
3
1
0
6
3
P
3
I
-
1
-
2
-
d
1
-
3
-
d2
-
10
-
f1
-
11
-
12
-
13
-
14
-
j1
-
15
-
d6
-
4
-
5
-
d3
-
6
-
d4
-
7
-
d5
-
16
-
F
6
3
8
4
5
6
28
4
2
2
3
P
4
I
-
1
-
2
-
d1
-
3
-
d2
-
10
-
f1
-
11
-
12
-
13
-
14
-
j1
-
15
-
d6
-
4
-
5
-
d3
-
6
-
d4
-
16
-
F
7
0
2
9
8
1
8
28
6
3
0
P
5
I
-
1
-
2
-
d1
-
3
-
d2
-
10
-
f1
-
11
-
12
-
13
-
14
-
j1
-
15
-
d6
-
4
-
5
-
d3
-
16
-
F
8
0
3
7
8
4
3
6
32
2
0
P
6
I
-
1
-
2
-
d1
-
3
-
d2
-
10
-
f1
-
11
-
12
-
13
-
14
-
j1
-
15
-
d6
-
16
-
F
9
1
6
2
2
4
5
4
2
35
4
P
7
I
-
1
-
2
-
d1
-
3
-
d2
-
4
-
5
-
d3
-
6
-
d4
-
7
-
d5
-
8
-
9
-
F
1
39
4
1
1
5
3
2
3
1
6
P
8
I
-
1
-
2
-
d1
-
3
-
d2
-
4
-
5
-
d3
-
6
-
d4
-
7
-
d5
-
16
-
F
2
1
28
7
6
3
9
4
3
3
1
P
9
I
-
1
-
2
-
d1
-
3
-
d2
-
4
-
5
-
d3
-
6
-
d4
-
16
-
F
3
1
0
28
5
8
4
9
7
3
0
P
10
I
-
1
-
2
-
d1
-
3
-
d2
-
4
-
5
-
d3
-
16
-
F
4
1
7
1
25
3
4
5
10
7
2
2
.
4
.
Pr
opose
ap
pr
oaches
to
t
es
t
p
ath pri
or
itiz
at
ion
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
11
, No
.
1,
Febr
uar
y
2021
:
57
5
-
588
582
2
.
4
.
P
r
opos
e
ap
pr
oaches
to
tes
t
p
ath
pri
or
i
ti
z
at
ion
The
weig
ht
ass
ign
m
ents
in
the
act
ivit
y
flow
gr
ap
h
tra
ns
f
orm
ed
fr
om
the
UML
act
ivit
y
diag
ram
are
the
top
ic
of
t
his
sect
ion
a
nd
t
hen
t
he
sc
or
es
of
t
he
te
st
pat
hs
are
cal
culat
e
d.
T
he
act
ivit
y
flo
w
grap
h
c
onsist
s
of no
des
a
nd e
dg
e
s a
s s
how
n i
n
Fig
ures
5
,
6
,
7
,
and
8. T
he nodes
and e
dg
es h
a
ve
as
sig
ne
d weig
hts as f
ollows.
2.4.1. De
fine
w
ei
gh
ts of t
he
no
des
in
th
e
ac
tivit
y flow
gr
ap
h
In
it
ia
l
node
an
d
final
node
pres
ente
d
the
fl
ow
sta
rt
a
nd
t
he
final
ste
p
a
re
assig
ned
0.
Me
rg
e,
for
k,
and
join
node
wh
ic
h
are
t
he
con
t
ro
l
node
t
hat
de
pu
te
d
fl
ow
of
the
pro
gra
m
are
assigne
d
0.
Decisi
on
node
is
assigne
d
0,
we
us
e
se
ve
ral
m
et
hods
to
co
ns
i
der
t
he
e
dg
e
w
ei
gh
ts
for
the
de
ci
sion
node
i
nst
ead.
Acti
vity
node
con
ta
ine
d
t
he p
rogr
am
stat
e
m
ent is assi
gn
e
d 1.
2.4.2. De
fine
w
ei
gh
ts of t
he
edges
in the
a
ctivit
y flow
gr
ap
h
Norm
al
edg
e
is
assigne
d
1.
E
dg
e
of
decisi
on
node
is
base
d
on
the
w
ork
br
ea
kdow
n
str
uctu
re.
It
can
hav
e
m
or
e
tha
n
2
e
d
ges
but
the
total
weig
ht
of
T
ru
e v
al
ue
and
false
val
ue
is
100%.
T
he
Tru
e
e
dge
is
as
sign
e
d
m
or
e
weigh
t
than
the
false
edg
e
.
I
n
this
res
earch
,
the
wei
gh
t
of
an
e
dg
e
is
based
on
ei
ther
t
im
e
m
ana
gem
ent
m
et
ho
d,
Pa
ret
o
m
e
tho
d,
Buffet
t
m
et
ho
d,
Bi
nar
y
m
et
ho
d,
or
B
ip
olar
m
e
t
hod
as
show
n
in
Table
7.
Th
e
total
weig
ht
of
ed
ge
s
of
a
de
ci
sion
no
de
is
5.
T
he
Tim
e
m
anag
em
ent,
a
70
/
30
ru
le
is
ap
plied
to
assi
gn
wei
ghts
to
edg
e
s
f
r
om
decisi
on
node
[
29
]
.
F
or
70
%
the
Tru
e
ed
ge
wei
gh
t
a
ssig
ned
is
3.5,
an
d
30%
t
he
f
al
se
ed
ge
weig
ht
assigne
d
is
1.5.
The
Pa
reto
,
a
n
80
/
20
r
ul
e
is
ap
plied
to
assi
gn
wei
gh
ts
to
ed
ges
from
the
de
ci
sion
node
[22,
29]
.
For
80%
t
he
Tru
e
ed
ge
weigh
t
as
sig
ned
is
4
a
nd
20%
t
he
false
ed
ge
w
ei
gh
t
assi
gn
e
d
is
1.
The
B
uffett
,
a
90
/
10
ru
le
is
a
pp
li
ed
to
assig
n
we
ig
hts
t
o
e
dg
e
s
from
a
d
e
ci
sion
node
[
29]
.
F
or
90
%
th
e
T
r
ue
edg
e
weig
ht
a
ssign
e
d
is
4.5
,
and
10%
the
false
ed
ge
wei
gh
t
assi
gn
e
d
i
s
0.5.
T
he
Bi
na
ry,
a
100/
0
r
ule
is
app
li
ed
t
o
assi
gn
weig
hts
to
edg
e
s
from
the
decisi
on
node
.
Fo
r
100%
the
Tru
e
e
dge
wei
gh
t
assi
gn
e
d
is
5
an
d
0%
t
he
false
e
dg
e
wei
ght
ass
ign
e
d
is
0.
T
he
Bi
po
la
r,
a
100/
-
10
0
ru
le
is
ap
plied
t
o
as
s
ign
wei
gh
ts
t
o
ed
ges
from
the
decisi
on
no
de.
F
or
100%
the
T
r
ue
edg
e
weig
ht
assigne
d
is
5
and
-
100%
th
e
false
edg
e
weig
ht
assigne
d
is
-
5.
Table
7.
Def
i
ne
w
ei
ghts
of th
e ed
ges fr
om
a d
eci
sio
n node
Edg
e
W
eig
h
t assig
n
m
en
t
m
eth
o
d
s
Ti
m
e
m
an
ag
e
m
en
t
(70
/3
0
r
u
le)
Pareto
(80
/2
0
r
u
le)
B
u
ff
ett
(90
/1
0
r
u
le
)
Bin
ary
(10
0
/0
r
u
le)
Bip
o
lar
(10
0
/
-
1
0
0
r
u
le)
Tr
u
e
3
.5
4
.0
4
.5
5
.0
5
.0
False
1
.5
1
.0
0
.5
0
.0
-
5
.0
2.4.3.
C
alcula
t
ing s
c
ores
and
prioritiz
ing
f
or test
p
at
hs
The
te
st
path
s
from
each
act
ivit
y
diag
ram
wer
e
gi
ven
sc
or
es
Sc
or
e
p
to
pr
i
or
it
iz
e
their
im
po
rtance
.
W
e
is
an
ed
ge
weig
ht
in
t
he
t
est
path
an
d
E
is
the
nu
m
ber
of
ed
ges
i
n
t
he
te
st
path
.
W
n
i
s
the
weig
ht
of
the
te
st
path
(
su
m
of
node
weig
hts
in
te
st
path)
a
nd
N
is
the
nu
m
be
r
of
no
des
in
t
he
te
st
path.
T
he
scor
e
of
a
te
s
t
path
Score
p
is cal
cu
la
te
d
by
(7),
an
d
the
test
p
at
hs are
pr
i
or
it
iz
ed a
ccordin
g
t
o
th
ei
r
sco
res
i
n de
scen
ding
order.
n
N
n
E
e
e
p
W
W
S
c
o
r
e
1
1
(
7)
2.5.
C
om
pa
ri
ng
c
om
put
e
d
resul
ts
on
pri
oritiz
at
ion
of
test
p
at
h
s
w
it
h
t
he
sur
ve
y
r
esults
on
th
e
t
esters
’
prioriti
z
at
io
n
of t
es
t
p
aths
The
ob
j
ect
ive
of
the
pro
po
se
d
ap
proac
h
is
to
rank
the
te
s
t
paths
the
sa
m
e
as
the
te
st
ers’
inter
est
s.
In
thi
s
sect
io
n,
the
dif
fer
e
nce
and
t
he
sim
il
arit
y
between
te
s
t
path
pri
ori
ti
zat
ion
by
the
propose
d
m
et
ho
ds
and
the s
ubj
ect
ive t
est
p
at
h p
rio
riti
zat
ion
acc
ordin
g
to
the test
e
rs’ interest
s
will
be evaluate
d.
2.5.1.
T
he c
alculat
i
on
of the
difference
to s
ubj
ec
tive
prio
ri
tiz
at
ion
of the tes
t
p
aths
The
diff
e
re
nce
Diff
(
T,R)
bet
ween
te
st
path
rank
by
the
propose
d
m
et
ho
ds
(
R)
and
the
s
ubj
ect
ive
te
st
path
pri
or
it
iz
at
ion
acc
ordi
ng
to
the
te
ste
rs
’
i
nterests
(
T)
is
qu
a
ntifie
d
in
t
his
sect
io
n.
Th
e
dif
fer
e
nce
be
twee
n
te
st
pa
th
pr
i
or
i
ti
zat
ion
can
be
cal
culat
ed
by
(8)
w
he
re
P
is
the
num
ber
of
te
st
path
s
in
act
ivit
y
flow
grap
h.
The
value
of
diff
e
ren
ce
Diff
(
T,R)
is
between
0
a
nd
1,
the
id
eal
value
f
or
D
if
f(
T,R
)
is
0.
D
p
is
the
diff
e
re
nce
of
a
T
-
R
pair
of
t
he
te
st
path.
T
p
is
the
te
st
path
ra
nk
acc
ordi
ng
t
o
the
te
ste
r
s’
interest
s,
a
nd
R
p
is
te
st
path
ra
nk
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ta
r
get
-
based
test
path prio
rit
izati
on for
U
M
L
activ
it
y
...
(
Walai
po
r
n S
ornkl
iang)
583
by
a
pro
po
se
d
al
gorithm
ic
m
et
ho
d.
I
f
R
p
is
eq
ual
to
T
p
,
this
m
eans
th
at
the
propose
d
m
et
ho
d
can
m
at
ch
the
te
st
path
ra
nk
t
o
the
te
ste
r
s’
intere
sts,
the
dif
fer
e
nce
of
t
he
te
st
pa
t
h
is
0.
If
R
p
is
m
ore
than
T
p
,
this
m
eans
that
rank
by
the
pr
opos
e
d
m
et
ho
d
is
t
oo
la
te
fo
r
te
sti
ng.
T
he
dif
fer
e
nce
of
th
e
te
s
t
path
is
puni
sh
e
d
as
the
double
of
the
dif
fer
e
nce
of
their
ra
nks.
If
R
p
is
le
ss
th
an
T
p
,
the
dif
f
eren
ce
of
the
te
st
path
cal
cula
te
d
as
the d
i
ff
e
ren
ce
of their
ra
nk
s
.
Max
Diff
is t
he m
axi
m
u
m
su
m
of the
d
if
fer
e
nc
e of T
-
R pai
r of al
l paths
.
M
a
xD
i
f
f
D
R
T
D
i
f
f
P
p
p
/
)
,
(
1
,
(
8)
wh
e
re
p
p
p
p
p
p
p
p
p
p
p
T
R
R
T
T
R
T
R
R
T
D
,
,
0
,
2
,
O
d
d
Nu
m
b
er
is
P
P
E
ve
n
Nu
m
b
e
r
is
P
P
M
a
x
D
i
f
f
,
4
1
3
,
2
3
2
2
.
2.5.2.
T
he c
alculat
i
on
of the
simi
larity to s
ubj
ec
tive pri
or
itiz
at
ion
of the tes
t
p
aths
The
sim
il
arit
y
Sim(T,R)
between
te
st
path
ran
k
by
the
pr
opose
d
m
et
ho
ds
(
R
)
and
the
subj
ect
iv
e
te
st
path
pr
i
or
it
iz
at
ion
acc
ordin
g
to
the
te
ste
rs’
i
nterests
(
T)
is
qu
a
ntifie
d
i
n
this
sect
ion.
Th
e
si
m
il
ari
ty
between
te
st
path
pri
or
i
ti
zat
ion
can
be
cal
culat
ed
by
(9),
w
her
e
P
is
the
nu
m
ber
of
te
st
path
s
in
a
ct
ivit
y
flow
gr
aph.
The
value
of
si
m
il
arity
Sim(T,
R)
is
between
0
a
nd
1,
the
i
de
al
value
for
Si
m(T,R)
is
1.
S
p
is
the
sim
i
la
rity
of
a
T
-
R
pair
of
t
he
te
st
path.
T
p
is
th
e
te
st
path
ra
nk
acc
ordi
ng
t
o
the
te
ste
r
s’
interest
s,
a
nd
R
p
is
te
st
path
ra
nk
by
a
pro
posed
al
gorithm
ic
method.
If
R
p
is
le
ss
than
T
p
,
th
is
m
eans
that
r
ank
by
the
pro
po
s
ed
m
et
ho
d
is
no
t
too
lat
e f
or test
ing
,
the sim
il
ar
it
y of
the test
pa
th is 1. If
R
p
i
s
eq
ual to
T
p
, t
his m
eans th
at
the prop
os
e
d m
et
ho
d
can
m
at
ch
the
t
est
path
ra
nk
t
o
the
te
ste
rs’
i
nterests,
the
si
m
il
arity
of
t
he
te
st
path
is
do
ubly
re
warded.
I
f
R
p
is
m
or
e
than
T
p
,
this
m
eans
tha
t
rank
by
t
he
pro
po
se
d
m
et
h
od
is
t
oo
la
te
f
or
te
sti
ng,
the
si
m
il
arity
of
t
he
te
st
path
is
0.
M
ax
Sim
is
the m
axim
u
m
su
m
o
f
th
e sim
il
arit
y of
T
-
R
pair of
all
p
at
hs.
M
a
xS
i
m
S
R
T
Sim
P
p
p
/
)
,
(
1
,
(9)
wh
e
re
p
p
p
p
p
p
p
R
T
R
T
R
T
S
,
0
,
2
,
1
,
P
M
ax
S
i
m
2
.
3.
RESU
LT
S
A
ND
DI
SCUS
S
ION
S
The
stu
dy
ass
essed
the
pro
pose
d
wei
gh
t
a
ssign
m
ent
m
eth
ods
i.e
.
no
de
weig
hts
an
d
edg
e
weig
hts
app
ly
in
g
with
Ti
m
e
m
anag
e
m
ent
m
et
ho
d,
Paret
o
m
et
ho
d,
Bu
ff
et
t
m
et
hod
,
Bi
na
ry
m
e
tho
d,
an
d
Bi
po
la
r
m
et
ho
d.
I
n
the
exp
e
rim
ent,
the
te
st
paths
fr
om
fo
ur
UML
a
ct
ivit
y
diagr
am
s
(as
descr
i
bed
in
sect
ion
2.3)
wer
e
pr
i
or
it
iz
ed
a
nd
com
par
ed wit
h p
rior resea
rc
he
s
.
3.1.
The res
ults
of
tes
t
p
ath
prioriti
z
at
io
n
The
pr
io
riti
zat
i
on
s
of
te
st
paths
f
ro
m
the
act
ivit
y
flow
gra
ph
s
we
re
obta
ined
by
assig
ni
ng
weig
ht
s
us
in
g
fi
ve
pro
po
s
ed
m
et
ho
ds:
cal
culat
ing
th
e
scor
e
s
f
or
t
he
te
st
paths
by
(7),
cal
c
ulati
ng
the
dif
fer
e
nc
e
an
d
the
si
m
il
arity
t
o
sub
j
ect
ive
pr
ioriti
zat
ion
of
the
te
st
paths
accor
ding
to
t
he
te
ste
rs'
interest
s
by
(
8)
a
nd
(
9)
,
resp
ect
ively
. If
two
test
p
at
hs
h
ad
the sam
e s
cor
e
, it wo
uld
assign
the sam
e ran
k value and
the n
ext num
ber(s
)
will
be
sk
ip
pe
d.
T
he
res
ults
of
te
st
path
pri
or
it
iz
at
ion
f
or
ATM
W
it
hdra
w
act
ivit
y
flow
grap
h
are
s
how
n
in
Table
8
,
f
or
S
hi
pp
in
g
Order
S
yst
e
m
activity
flo
w
grap
h
are
sh
ow
n
in
Ta
bl
e
9,
f
or
Ve
nd
i
ng
Ma
c
hin
e
ac
ti
vity
flo
w
gr
a
ph
are
sh
ow
n
in
Table
10,
a
nd f
or Mi
le
age
Purch
ase
act
ivit
y
flo
w
gr
a
ph
are
s
how
n
i
n
Ta
ble 11
. In
al
l
four
ta
bles,
“E
arly
”
m
eans
the
nu
m
ber
of
te
st
paths
that
r
a
nk
by
the
propose
d
m
et
ho
d
is
earli
er
than
ra
nk
by
the
te
ste
r.
“La
te
”
m
eans
the
nu
m
ber
of
te
st
path
s
that
ra
nk
by
the
pro
pose
d
m
et
ho
d
i
s
la
te
r
tha
n
ra
nk
by
the test
er. “Eq
ual” m
eans th
e
prop
os
ed
m
eth
od ca
n
m
at
ch t
he
te
st pat
h ra
nk to
t
he
te
ste
r
s’
inte
rests.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
11
, No
.
1,
Febr
uar
y
2021
:
57
5
-
588
584
Table
8.
T
he
r
e
su
lt
s of test
path pr
i
or
it
iz
at
ion f
or
ATM
w
it
hdra
w
act
ivit
y fl
ow gra
ph
Test path
no.
Test path
rank
by
p
rop
o
sed
weigh
t assig
n
m
en
t
m
e
th
o
d
s
Test path
r
an
k
b
y
tester
Ti
m
e
m
a
n
ag
e
m
en
t
Pareto
Bu
ff
ett
Bin
ary
Bip
o
lar
P1
5
4
2
2
1
1
P2
6
6
5
4
3
7
P3
7
7
7
6
5
2
P4
8
8
8
8
7
8
P5
1
1
1
1
2
3
P6
3
2
3
3
4
9
P7
2
3
4
4
6
4
P8
4
5
6
7
8
10
P9
10
9
9
9
9
5
P1
0
9
10
10
10
10
11
P1
1
12
11
11
11
11
6
P1
2
11
11
12
12
13
12
P1
3
14
14
14
13
12
13
P1
4
13
13
13
14
14
14
Ear
l
y
8
8
6
5
7
Equ
al
1
1
3
5
2
Late
5
5
5
4
5
Dif
f
erence
0
.43
0
.37
0
.33
0
.29
0
.31
Si
m
ila
rity
0
.36
0
.36
0
.43
0
.54
0
.39
Ta
ble
9.
T
he
r
e
su
lt
s of test
path pr
i
or
it
iz
at
ion f
or
s
hi
pp
i
ng
o
rd
e
r
s
yst
em
act
ivit
y flow g
raph
Test path
no.
Test path
r
an
k
by
p
rop
o
sed
weigh
t assig
n
m
en
t
m
e
th
o
d
s
Test path
r
an
k
b
y
tester
Ti
m
e
m
a
n
ag
e
m
en
t
Pareto
Bu
ff
ett
Bin
ary
Bip
o
lar
P1
2
2
3
4
5
6
P2
1
1
1
1
1
1
P3
3
2
2
2
3
3
P4
6
6
6
6
6
5
P5
4
4
3
3
2
2
P6
5
5
5
5
4
4
P7
7
7
7
7
7
7
Ear
l
y
1
2
2
2
1
Equ
al
3
2
2
2
5
Late
3
3
3
3
1
Dif
f
erence
0
.33
0
.36
0
.28
0
.25
0
.08
Si
m
ila
rity
0
.50
0
.43
0
.43
0
.43
0
.79
Table
10. T
he results
of test
pat
h
pri
ori
ti
zat
i
on for
v
e
ndin
g
m
achine acti
vity
f
lo
w gr
a
ph
Test path
no.
Test path
r
an
k
by
p
rop
o
sed
weigh
t assig
n
m
en
t
m
e
th
o
d
s
Test path
r
an
k
b
y
tester
Ti
m
e
m
a
n
ag
e
m
en
t
Pareto
Bu
ff
ett
Bin
ary
Bip
o
lar
P1
6
9
9
9
9
10
P2
1
1
1
1
4
9
P3
4
2
2
2
1
1
P4
10
1
0
10
10
10
8
P5
2
3
5
5
6
7
P6
2
3
5
5
6
5
P7
5
7
8
8
8
6
P8
7
5
3
3
2
3
P9
7
5
3
3
2
2
P1
0
9
8
7
7
5
4
Ear
l
y
5
4
3
3
4
Equ
al
0
0
2
2
2
Late
5
6
5
5
4
Dif
f
erence
0
.79
0
.55
0
.39
0
.39
0
.27
Si
m
ila
rity
0
.25
0
.20
0
.35
0
.35
0
.40
Evaluation Warning : The document was created with Spire.PDF for Python.