I
AE
S
I
n
t
e
r
n
at
ion
al
Jou
r
n
al
of
Ar
t
if
icial
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
Vol.
14
,
No
.
2
,
Apr
il
2025
,
pp.
97
5
~
98
3
I
S
S
N:
2252
-
8938
,
DO
I
:
10
.
11591/i
jai
.
v
14
.i
2
.
pp
97
5
-
98
3
975
Jou
r
n
al
h
omepage
:
ht
tp:
//
ij
ai
.
iaes
c
or
e
.
c
om
Op
t
imiz
in
g q
u
e
u
e
e
f
f
ic
i
e
n
c
y:
A
r
t
ific
ia
l
in
t
e
ll
ig
e
n
c
e
-
d
r
iv
e
n
t
an
d
e
m
q
u
e
u
e
s with
r
e
n
e
gi
n
g
Kee
r
t
h
ik
a
Radh
ak
r
is
h
n
an
1
,
S
u
b
r
am
an
i
P
alan
i
Nir
an
j
an
1
,
B
alak
r
is
h
n
an
Kom
ala
Durga
2
,
S
u
r
e
s
h
a
Ram
ar
e
d
d
y
3
1
D
e
pa
r
tm
e
nt
of
M
a
th
e
ma
ti
c
s
, V
e
l
T
e
c
h
R
a
nga
r
a
ja
n D
r
. S
a
gunt
ha
la
R
&D I
ns
ti
tu
te
of
S
c
ie
nc
e
a
nd
T
e
c
hnol
ogy, C
h
e
nna
i,
I
ndi
a
2
D
e
pa
r
tm
e
nt
of
M
a
th
e
ma
ti
c
s
, V
ir
udhuna
ga
r
H
in
du N
a
da
r
s
’
S
e
n
th
ik
uma
r
a
N
a
da
r
C
ol
le
ge
, V
ir
udhuna
ga
r
, I
ndi
a
3
D
e
pa
r
tm
e
nt
of
M
a
th
e
ma
ti
c
s
, F
a
c
ul
ty
of
E
ngi
ne
e
r
in
g a
nd T
e
c
h
nol
ogy
,
J
A
I
N
(
D
e
e
me
d
-
to
-
be
U
ni
ve
r
s
it
y)
,
B
a
nga
lo
r
e
, I
ndi
a
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
M
a
r
12
,
2024
R
e
vis
e
d
Oc
t
24
,
2024
Ac
c
e
pted
Nov
14
,
2024
T
h
i
s
p
a
p
er
d
e
l
v
e
s
i
n
t
o
t
h
e
t
h
eo
re
t
i
ca
l
i
n
t
e
g
rat
i
o
n
o
f
q
u
e
u
ei
n
g
t
h
eo
r
y
an
d
a
rt
i
f
i
ci
a
l
i
n
t
e
l
l
i
g
e
n
ce
(
AI
)
,
ex
am
i
n
i
n
g
t
h
e
b
e
n
ef
i
t
s
a
n
d
i
mp
l
i
ca
t
i
o
n
s
o
f
t
h
e
i
r
co
n
v
erg
e
n
ce.
Q
u
e
u
ei
n
g
s
y
s
t
em
s
s
erv
e
a
s
fu
n
d
ame
n
t
a
l
mo
d
el
s
fo
r
v
ar
i
o
u
s
real
-
w
o
rl
d
ap
p
l
i
ca
t
i
o
n
s
,
fro
m
t
e
l
eco
mm
u
n
i
cat
i
o
n
s
n
et
w
o
r
k
s
t
o
h
ea
l
t
h
care
faci
l
i
t
i
es
.
T
h
i
s
res
earc
h
p
res
e
n
t
s
a
t
ran
s
fo
rma
t
i
v
e
framew
o
r
k
fo
r
el
ev
at
i
n
g
t
h
e
effi
c
i
en
c
y
an
d
p
erf
o
rman
c
e
o
f
q
u
e
u
ei
n
g
s
y
s
t
ems
b
y
i
n
f
u
s
i
n
g
A
I
-
d
r
i
v
e
n
t
an
d
em
q
u
eu
e
an
a
l
y
s
i
s
.
T
h
e
i
mp
l
i
ca
t
i
o
n
s
o
f
t
h
i
s
ap
p
r
o
ach
t
ran
s
cen
d
i
n
d
u
s
t
ri
es
,
p
r
o
mi
s
i
n
g
s
t
ream
l
i
n
ed
o
p
era
t
i
o
n
s
,
red
u
ce
d
w
ai
t
i
n
g
t
i
mes
,
a
n
d
res
o
u
rce
o
p
t
i
mi
za
t
i
o
n
.
T
h
i
s
w
o
r
k
i
n
v
i
t
e
s
f
u
rt
h
er
ex
p
l
o
ra
t
i
o
n
a
n
d
a
p
p
l
i
ca
t
i
o
n
,
o
fferi
n
g
a
p
at
h
t
o
mo
re
effect
i
v
e
a
n
d
re
s
p
o
n
s
i
v
e
q
u
e
u
ei
n
g
s
y
s
t
em
s
g
l
o
b
al
l
y
.
O
v
er
t
h
e
y
ears
,
res
earc
h
ers
a
n
d
p
rac
t
i
t
i
o
n
er
s
h
a
v
e
ex
p
l
o
red
n
u
mero
u
s
t
ech
n
i
q
u
e
s
t
o
en
h
a
n
ce
t
h
e
effi
c
i
en
c
y
an
d
p
erf
o
rman
ce
o
f
q
u
eu
e
i
n
g
s
y
s
t
em
s
.
In
recen
t
t
i
mes
,
i
n
t
e
g
rat
i
n
g
A
I
i
n
t
o
t
h
e
rea
l
m
o
f
q
u
eu
ei
n
g
an
a
l
y
s
i
s
h
as
o
p
e
n
ed
u
p
n
e
w
av
e
n
u
e
s
fo
r
o
p
t
i
mi
za
t
i
o
n
a
n
d
i
n
n
o
v
at
i
o
n
.
T
h
i
s
p
a
p
er
s
t
u
d
i
e
s
a
t
w
o
-
s
erv
er
t
a
n
d
em
q
u
eu
e
i
n
g
mo
d
e
l
w
i
t
h
ren
e
g
i
n
g
cu
s
t
o
mer
s
u
s
i
n
g
AI
t
ech
n
i
q
u
e
s
.
A
s
s
u
mi
n
g
t
h
at
t
h
e
arri
v
al
rat
e
fo
l
l
o
w
s
t
h
e
Po
i
s
s
o
n
p
ro
ce
s
s
an
d
t
h
e
s
er
v
i
ce
rat
e
fo
l
l
o
w
s
an
ex
p
o
n
e
n
t
i
al
d
i
s
t
r
i
b
u
t
i
o
n
,
u
s
i
n
g
t
h
e
b
i
r
t
h
-
d
eat
h
p
ro
ce
s
s
,
p
r
o
b
a
b
i
l
i
t
y
g
en
era
t
i
n
g
f
u
n
c
t
i
o
n
an
d
AI
m
o
d
u
l
e,
w
e
d
eri
v
e
s
t
ead
y
-
s
t
a
t
e
d
i
ffere
n
ce
eq
u
at
i
o
n
,
ex
p
ec
t
ed
n
u
mb
er
o
f
p
eo
p
l
e
i
n
cu
s
t
o
mer
s
,
an
d
mean
w
ai
t
i
n
g
t
i
me.
K
e
y
w
o
r
d
s
:
Ar
ti
f
icia
l
int
e
ll
igenc
e
B
ir
th
de
a
t
h
pr
oc
e
s
s
E
xpe
dit
e
wa
it
ing
t
im
e
R
e
ne
ging
T
a
nde
m
que
ue
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e.
C
or
r
e
s
pon
din
g
A
u
th
or
:
Ke
e
r
thi
ka
R
a
dha
kr
is
hna
n
De
pa
r
t
ment
o
f
M
a
the
mati
c
s
,
Ve
l
T
e
c
h
R
a
nga
r
a
jan
Dr
.
S
a
gun
tha
la
R
&D
I
ns
t
it
u
te
of
S
c
ienc
e
a
nd
T
e
c
h
nolog
y
Ava
di,
C
he
nna
i,
T
a
mi
lnadu,
I
ndia
E
mail:
ke
e
r
thi
27789
@gmail
.
c
om
1.
I
NT
RODU
C
T
I
ON
Que
ue
ing
s
ys
tems
s
e
r
ve
a
s
the
ba
c
kbone
of
numer
ous
indus
tr
ies
a
nd
ope
r
a
ti
ona
l
s
c
e
na
r
ios
,
inf
luenc
ing
e
ve
r
ythi
ng
f
r
om
c
us
tom
e
r
s
a
ti
s
f
a
c
ti
on
to
r
e
s
our
c
e
uti
li
z
a
ti
on
a
nd
c
os
t
-
e
f
f
e
c
ti
ve
ne
s
s
.
As
or
ga
niza
ti
ons
s
tr
ive
to
de
li
ve
r
mor
e
e
f
f
icie
nt
a
nd
r
e
s
pons
ive
s
e
r
vice
s
,
the
int
e
gr
a
ti
on
of
a
r
ti
f
icia
l
int
e
ll
igenc
e
(
AI
)
ha
s
e
mer
ge
d
a
s
a
t
r
a
ns
f
or
mative
f
or
c
e
.
Que
ue
ing
s
ys
tems
,
c
ha
r
a
c
ter
ize
d
by
their
inher
e
nt
c
ompl
e
xit
y
a
nd
int
e
r
de
pe
nde
nc
e
of
tas
ks
,
p
r
e
s
e
nt
a
unique
s
e
t
of
c
ha
ll
e
nge
s
f
or
opti
mi
z
a
ti
on
.
T
r
a
dit
ionally
,
que
u
e
theor
y
ha
s
pr
o
vided
va
luable
ins
ight
s
a
nd
methodologi
e
s
f
or
tac
kli
ng
thes
e
c
ha
ll
e
nge
s
.
How
e
ve
r
,
the
a
dve
nt
of
A
I
ha
s
int
r
oduc
e
d
a
dyna
mi
c
a
nd
da
ta
-
dr
iven
d
im
e
ns
ion
to
que
ue
ing
s
ys
tem
opti
mi
z
a
ti
on,
p
r
omi
s
ing
unpr
e
c
e
de
nted
a
dva
nc
e
ments
.
E
f
f
icie
nt
que
ue
mana
ge
ment
is
im
pe
r
a
ti
ve
in
d
iver
s
e
f
ields
s
uc
h
a
s
tele
c
omm
unica
ti
ons
,
he
a
lt
hc
a
r
e
,
a
nd
tr
a
ns
por
tation.
T
a
nde
m
que
ue
s
,
c
ompr
is
ing
int
e
r
c
onne
c
ted
que
ue
s
,
r
e
pr
e
s
e
nt
a
c
omm
on
s
tr
uc
tur
e
in
s
ys
tems
whe
r
e
e
nti
ti
e
s
unde
r
go
mul
ti
ple
s
tage
s
of
p
r
oc
e
s
s
ing
or
s
e
r
vice
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
2
,
Apr
il
2
025
:
97
5
-
98
3
976
How
e
ve
r
,
the
dyna
m
ic
na
tur
e
of
tande
m
que
ue
s
,
c
oupled
with
the
phe
nomenon
of
r
e
ne
ging
(
whe
r
e
c
us
tom
e
r
s
lea
ve
the
que
ue
be
f
or
e
s
e
r
vice
)
,
pos
e
s
s
igni
f
ica
nt
c
ha
ll
e
nge
s
to
tr
a
dit
ional
opti
mi
z
a
ti
on
methods
.
T
his
s
tudy
f
oc
us
e
s
on
ha
r
ne
s
s
ing
the
c
a
pa
bil
it
ies
of
AI
to
a
na
lyze
a
nd
o
pti
mi
z
e
tande
m
que
ue
s
,
pa
r
ti
c
ular
l
y
in
the
c
ontext
of
r
e
ne
ging
be
ha
vior
.
T
he
main
o
bjec
ti
ve
s
a
r
e
:
−
De
ve
lop
AI
m
ode
ls
:
c
r
e
a
te
AI
a
lgor
it
hms
c
a
pa
ble
of
pr
e
dicting
a
nd
mana
ging
c
us
tom
e
r
be
ha
vior
,
pa
r
ti
c
ular
ly
r
e
ne
ging
,
in
tande
m
que
ue
s
ys
tems
.
−
Optim
ize
que
ue
ope
r
a
ti
ons
:
d
e
s
ign
a
nd
im
pleme
nt
s
tr
a
tegie
s
to
mi
nim
ize
the
ne
ga
ti
ve
im
pa
c
ts
of
r
e
ne
ging
on
s
e
r
vice
e
f
f
icie
nc
y
a
nd
c
us
tom
e
r
s
a
ti
s
f
a
c
ti
on.
−
S
im
ulation
a
nd
v
a
li
da
ti
on:
c
onduc
t
s
im
ulations
to
c
ompar
e
the
pe
r
f
o
r
manc
e
of
A
I
-
dr
iven
q
ue
ue
mana
ge
ment
a
ga
ins
t
tr
a
dit
ional
methods
.
−
P
r
a
c
ti
c
a
l
i
mpl
e
menta
ti
on:
p
r
opos
e
a
c
ti
ona
ble
ins
i
ghts
a
nd
f
r
a
mew
or
ks
f
or
int
e
gr
a
ti
ng
AI
-
dr
iven
q
ue
ue
mana
ge
ment
s
olut
ions
in
r
e
a
l
-
wor
ld
a
ppli
c
a
ti
ons
.
T
a
nde
m
que
ue
s
ys
tems
a
r
e
c
omm
only
us
e
d
i
n
va
r
ious
s
e
r
vice
e
nvir
onments
,
s
uc
h
a
s
ba
nks
,
hos
pit
a
ls
,
a
nd
c
a
ll
c
e
nter
s
.
How
e
ve
r
,
c
us
tom
e
r
r
e
n
e
ging
pos
e
s
s
igni
f
ica
nt
c
ha
ll
e
nge
s
,
lea
ding
to
inef
f
icie
nc
ies
a
nd
r
e
duc
e
d
c
us
tom
e
r
s
a
ti
s
f
a
c
ti
on.
T
r
a
dit
ional
que
ue
mana
ge
ment
s
tr
a
tegie
s
of
ten
f
a
il
to
a
ddr
e
s
s
the
dyna
mi
c
a
nd
unpr
e
dicta
ble
na
tur
e
of
c
us
tom
e
r
be
ha
vior
e
f
f
e
c
ti
ve
ly.
T
h
is
r
e
s
e
a
r
c
h
s
e
e
k
s
to
a
ddr
e
s
s
the
pr
oblem
of
opti
mi
z
ing
tande
m
que
ue
s
ys
tems
by
leve
r
a
ging
AI
to
p
r
e
dict
a
nd
mana
ge
r
e
ne
ging,
thus
im
pr
ov
in
g
ove
r
a
ll
e
f
f
icie
nc
y
a
nd
s
e
r
vice
qua
li
ty.
T
he
im
po
r
tanc
e
of
opti
mi
z
ing
que
ue
e
f
f
icie
nc
y
with
AI
-
dr
iven
tande
m
que
ue
s
a
nd
a
ddr
e
s
s
ing
r
e
ne
ging
:
−
C
us
tom
e
r
s
a
ti
s
f
a
c
ti
on:
e
f
f
icie
nt
que
ue
mana
ge
men
t
r
e
duc
e
s
wa
it
ti
mes
a
nd
unc
e
r
tainty,
e
nha
nc
ing
ov
e
r
a
ll
c
us
tom
e
r
e
xpe
r
ienc
e
a
nd
s
a
ti
s
f
a
c
ti
on.
−
Ope
r
a
ti
ona
l
e
f
f
ic
ienc
y:
AI
op
ti
mi
z
e
s
r
e
s
our
c
e
a
ll
oc
a
ti
on
a
nd
s
taf
f
ing,
m
ini
mi
z
ing
c
os
ts
a
nd
maxi
mi
z
ing
thr
oughput.
−
C
os
t
s
a
vings
:
e
f
f
e
c
ti
ve
mana
ge
ment
r
e
duc
e
s
o
pe
r
a
ti
ona
l
e
xpe
ns
e
s
by
a
li
gning
s
taf
f
ing
leve
ls
with
de
mand,
a
voidi
ng
ove
r
s
taf
f
ing
o
r
unde
r
uti
li
z
a
ti
on.
−
B
us
ines
s
p
e
r
f
or
manc
e
:
i
mp
r
ove
d
e
f
f
icie
nc
y
boos
t
s
r
e
ve
nue
ge
ne
r
a
ti
on
a
nd
c
us
tom
e
r
r
e
tention,
pr
ov
idi
ng
a
c
ompetit
ive
e
dge
.
−
Ada
ptabili
ty:
AI
a
da
pts
to
va
r
ying
de
mands
a
nd
dis
r
upti
ons
,
maintaining
pe
r
f
or
manc
e
unde
r
dyn
a
mi
c
c
ondit
ions
.
Ade
poju
[
1]
inves
ti
ga
tes
the
be
ha
vior
of
tande
m
que
ue
s
withi
n
the
c
ontext
of
por
ous
mediums
.
W
a
ng
e
t
al
.
[
2
]
de
ve
lop
models
to
pr
e
dict
que
u
e
be
ha
vior
a
nd
a
na
lyze
va
r
ious
s
tr
a
tegie
s
to
mi
ti
ga
te
the
ne
ga
ti
ve
e
f
f
e
c
ts
of
im
pa
ti
e
nc
e
.
Alf
r
e
d
[
3]
pa
pe
r
i
nve
s
ti
ga
tes
s
c
e
na
r
ios
whe
r
e
q
ue
ue
s
withi
n
a
ne
t
wor
k
c
a
n
be
c
ome
blocke
d
a
nd
it
e
ms
may
withdr
a
w
f
r
o
m
the
que
ue
.
Ayya
ppa
n
a
nd
T
ha
mi
z
hs
e
lvi
[
4
]
f
oc
u
s
on
the
tr
a
ns
ient
be
ha
vior
o
f
que
ue
ing
s
ys
tems
that
inco
r
por
a
te
p
r
ior
it
y
s
e
r
vice
s
.
Yua
n
e
t
al
.
[
5]
inves
ti
g
a
ted
the
potential
of
AI
in
opt
im
izing
ne
twor
k
a
s
s
oc
iation
f
or
ne
xt
-
ge
ne
r
a
ti
on
mobi
le
ne
twor
ks
,
a
c
hieving
e
nha
nc
e
d
QoS
a
nd
ne
twor
k
pe
r
f
or
manc
e
.
F
e
r
e
n
c
z
a
nd
Z
öld
y
[
6
]
e
xplor
e
d
the
a
ppli
c
a
ti
on
of
AI
methods
in
e
s
ti
mating
r
oa
d
tr
a
f
f
ic
que
ue
lengths
,
s
howc
a
s
ing
pr
omi
s
ing
r
e
s
ult
s
f
or
int
e
ll
igent
tr
a
ns
por
tation
s
ys
tems
.
T
he
s
tudy
by
Anus
s
or
nnit
is
a
r
n
a
nd
L
im
law
a
n
[
7]
de
mons
tr
a
te
d
the
po
tential
o
f
a
r
ti
f
icia
l
n
e
ur
a
l
ne
two
r
k
(
AN
N)
-
ba
s
e
d
wa
it
ing
ti
me
pr
e
diction
in
opti
mi
z
ing
que
ue
man
a
ge
ment,
pr
ovidi
ng
va
luable
ins
ight
s
f
or
s
e
r
vice
i
ndus
tr
ies
s
e
e
king
to
mi
nim
ize
c
us
tom
e
r
wa
it
t
im
e
s
.
Ka
mou
n's
[
8]
f
indi
ngs
a
r
e
s
igni
f
ica
nt
f
or
de
s
igni
ng
a
nd
o
pti
mi
z
ing
que
ue
ing
s
ys
tems
in
e
nvir
onments
whe
r
e
p
r
ior
it
i
z
a
ti
on
is
c
r
uc
ial,
s
uc
h
a
s
in
the
he
a
lt
hc
a
r
e
a
nd
c
us
tom
e
r
s
e
r
vice
s
e
c
tor
s
.
L
iu
a
nd
Qua
n
[
9]
dis
c
us
s
the
s
yn
e
r
gis
ti
c
r
e
lations
hip
be
twe
e
n
que
ue
ing
theor
y
a
nd
AI
in
their
pa
pe
r
in
[
10
]
e
xplo
r
e
the
s
yne
r
gis
ti
c
r
e
lations
hip
be
twe
e
n
que
ue
ing
theor
y
a
nd
AI
.
T
his
s
tudy
highl
ight
s
the
pr
a
c
ti
c
a
l
be
ne
f
it
s
of
incor
por
a
ti
ng
mac
hine
lea
r
ning
int
o
que
ue
mana
ge
ment,
pa
r
t
icula
r
ly
i
n
digi
tal
c
omm
unica
ti
on
a
nd
da
ta
pr
oc
e
s
s
ing
e
nvir
onments
.
Ayya
ppa
n
a
nd
T
hil
a
ga
va
thy
[
11]
e
xa
mi
ne
the
a
ppl
ica
ti
on
of
mac
hine
lea
r
ning
tec
hniques
to
a
na
lyze
c
ompl
e
x
que
ue
ing
s
ys
tem
s
.
M
a
s
e
t
al
.
[
12]
highl
ight
va
r
ious
AI
tec
hniques
,
s
uc
h
a
s
mac
hine
lea
r
ning
a
nd
ne
ur
a
l
ne
twor
ks
,
to
pr
e
dict
wa
it
ti
mes
,
im
p
r
ove
s
e
r
vice
e
f
f
icie
nc
y,
a
nd
e
nha
nc
e
ove
r
a
ll
s
ys
tem
pe
r
f
o
r
manc
e
.
L
i
e
t
al
.
[
13
]
r
e
view
the
int
e
r
s
e
c
ti
on
o
f
que
ue
in
g
theor
y
a
nd
AI
in
the
c
ontext
of
s
e
r
vice
s
ys
tems
.
T
hi
s
c
ompr
e
he
ns
ive
s
ur
ve
y
c
ove
r
s
the
theor
e
ti
c
a
l
f
oun
da
ti
ons
a
nd
pr
a
c
ti
c
a
l
a
ppli
c
a
ti
ons
of
AI
-
e
nha
n
c
e
d
que
ue
ing
models
,
de
mons
tr
a
ti
ng
how
A
I
c
a
n
be
us
e
d
to
a
ddr
e
s
s
c
omm
on
c
ha
ll
e
nge
s
in
s
e
r
vice
s
ys
tems
,
s
uc
h
a
s
va
r
iabili
ty
in
s
e
r
vice
de
mand
a
nd
r
e
s
our
c
e
a
ll
oc
a
ti
on.
L
i
e
t
al
.
[
13]
o
f
f
e
r
va
luable
ins
ight
s
f
or
mana
ging
que
ue
ing
s
ys
tems
in
e
nvi
r
onments
whe
r
e
s
pa
c
e
a
n
d
pa
ti
e
nc
e
a
r
e
l
im
it
e
d,
s
uc
h
a
s
in
tele
c
omm
unica
ti
ons
a
nd
tr
a
ns
por
tation.
S
he
kha
r
e
t
al
.
[
14]
f
indi
ngs
of
f
e
r
va
luable
i
ns
ight
s
f
or
mana
ging
que
ue
ing
s
y
s
tems
in
e
nvir
onments
whe
r
e
s
pa
c
e
a
nd
pa
ti
e
nc
e
a
r
e
li
m
it
e
d,
s
uc
h
a
s
in
tele
c
omm
unica
ti
ons
a
nd
t
r
a
ns
por
tation
.
B
a
ghe
l
a
nd
J
a
in
[
15]
wor
k
is
pa
r
ti
c
ula
r
ly
r
e
leva
nt
f
or
de
s
igni
ng
r
obus
t
s
ys
tems
in
c
r
it
ica
l
a
ppli
c
a
ti
ons
s
uc
h
a
s
a
e
r
os
pa
c
e
,
he
a
lt
hc
a
r
e
,
a
nd
inf
o
r
mation
tec
hnology
(
I
T
)
inf
r
a
s
tr
uc
tur
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
Optimiz
ing
que
ue
e
ff
icie
nc
y
:
A
r
ti
fi
c
ial
int
e
ll
igenc
e
-
dr
ive
n
tandem
que
ue
s
…
(
K
e
e
r
thi
k
a
R
adhak
r
is
hn
an)
977
T
he
r
e
s
e
a
r
c
h
pr
e
s
e
nted
in
[
16
]
–
[
18]
s
e
r
ve
s
a
s
a
f
ounda
ti
on
f
o
r
g
r
a
s
ping
the
f
unda
menta
l
r
ole
o
f
que
ue
ing
theor
y
in
modeling
a
nd
a
na
lyzing
va
r
i
ous
e
ve
r
yda
y
s
ys
tems
a
nd
pr
oc
e
s
s
e
s
.
Ndia
ye
e
t
al
.
[
17
]
de
ve
lopm
e
nt
of
AI
-
dr
iven
que
ue
mana
ge
ment
s
ys
tems
s
howc
a
s
e
s
the
potential
f
o
r
a
dva
nc
e
d
tec
hno
logi
e
s
to
e
nha
nc
e
e
f
f
icie
nc
y
a
nd
c
us
tom
e
r
e
xpe
r
ienc
e
.
M
e
a
nwhile,
S
ti
ntzing
a
nd
Nor
r
man's
[
18
]
e
xplor
a
ti
on
o
f
pr
e
dictive
mo
de
li
ng
with
AN
Ns
of
f
e
r
s
a
gl
im
ps
e
int
o
the
f
utu
r
e
of
dyna
mi
c
a
nd
int
e
ll
igen
t
que
ue
mana
ge
ment.
Yuva
r
a
ni
a
nd
Vijaya
laks
hmi
[
19]
e
x
plor
e
the
us
e
of
s
tocha
s
ti
c
pr
oc
e
s
s
e
s
in
de
s
igni
ng
que
ue
ing
models
with
AN
Ns
.
T
he
a
uthor
s
in
[
20]
,
[
21]
pr
opos
e
a
que
ue
ne
twor
k
model
us
ing
Ke
nda
ll
's
no
tation
to
mi
nim
ize
int
e
r
s
e
c
ti
on
wa
it
ing
ti
me
in
their
pa
pe
r
.
Anur
uddhika
e
t
al
.
[
22]
pr
ovide
a
c
ompr
e
he
ns
ive
ove
r
view
of
the
dif
f
e
r
e
nt
methodologi
e
s
a
nd
tec
hniques
us
e
d
in
que
ue
ing
mod
e
li
ng,
e
mphas
izing
the
e
volu
ti
on
a
nd
a
dva
nc
e
ments
in
the
f
ield
ove
r
ti
me
.
T
he
a
uthor
s
i
n
[
23]
,
[
24]
e
xplor
e
s
how
que
ue
ing
models
c
a
n
be
a
ppli
e
d
to
opti
mi
z
e
r
e
s
our
c
e
a
ll
oc
a
ti
on,
de
mons
tr
a
ti
ng
th
e
r
e
leva
nc
e
of
que
ue
ing
theor
y
in
e
mer
ging
tec
h
nologi
e
s
a
nd
a
na
lyze
d
mul
ti
-
s
e
r
ve
r
tande
m
que
ue
with
M
a
r
kovian
a
r
r
ival
pr
oc
e
s
s
,
pha
s
e
-
type
s
e
r
vice
ti
mes
,
a
nd
f
ini
te
buf
f
e
r
s
.
Ni
r
a
njan
e
t
al
.
[
25
]
us
e
d
a
n
opti
mu
m
c
os
t
a
na
lys
is
to
mi
nim
ize
the
tot
a
l
a
ve
r
a
ge
c
os
t
wi
th
da
ta
tr
a
ns
mi
s
s
ion
a
nd
da
ta
pr
oc
e
s
s
ing
in
long
ter
m
e
volut
ion
-
a
dva
nc
e
d
(
L
T
E
-
A)
ne
twor
ks
us
ing
the
dis
c
o
nti
nuous
r
e
c
e
pti
on
(
DR
X)
mec
ha
nis
m.
T
he
r
e
is
li
mi
ted
r
e
s
e
a
r
c
h
on
how
AI
-
dr
iven
que
ue
mana
ge
ment
im
pa
c
ts
ove
r
a
ll
c
us
tom
e
r
e
xpe
r
ienc
e
,
pa
r
ti
c
ular
ly
in
s
c
e
na
r
ios
invol
ving
r
e
ne
ging.
S
tudi
e
s
that
e
va
luate
c
us
tom
e
r
s
a
ti
s
f
a
c
ti
on
a
nd
pe
r
c
e
ived
f
a
ir
ne
s
s
in
AI
-
opti
mi
z
e
d
que
ue
ing
s
ys
tems
a
r
e
ne
e
de
d
to
e
ns
ur
e
that
tec
hnologi
c
a
l
a
dva
nc
e
ments
a
li
gn
with
c
us
tom
e
r
e
xpe
c
tations
.
Add
r
e
s
s
ing
thes
e
ga
ps
thr
ough
f
oc
us
e
d
r
e
s
e
a
r
c
h
c
a
n
lea
d
to
mor
e
e
f
f
icie
nt,
s
c
a
lable
,
a
nd
c
us
tom
e
r
-
f
r
iendly
que
ue
mana
ge
ment
s
ys
tems
,
leve
r
a
ging
the
f
u
ll
po
tential
o
f
A
I
in
tande
m
que
ue
s
c
e
na
r
ios
with
r
e
ne
ging.
T
h
is
s
tudy
inves
ti
ga
ted
the
e
f
f
e
c
ts
of
AI
-
dr
iven
opti
mi
z
a
ti
on
on
tande
m
que
ue
s
with
r
e
n
e
ging
be
ha
vior
.
W
hil
e
e
a
r
li
e
r
s
tudi
e
s
ha
ve
e
xplor
e
d
the
im
pa
c
t
of
t
r
a
dit
iona
l
que
ue
mana
ge
ment
tec
hniques
a
nd
s
tatic
opt
im
iza
ti
on
on
que
ue
pe
r
f
or
manc
e
,
they
ha
ve
not
e
xpli
c
it
ly
a
ddr
e
s
s
e
d
the
inf
luenc
e
of
r
e
a
l
-
ti
me,
AI
-
dr
iven
dyna
mi
c
a
da
ptati
on
on
the
e
f
f
ic
ienc
y
a
nd
s
tabili
ty
of
tande
m
que
ue
s
ys
tems
,
e
s
pe
c
ially
in
the
c
ontext
of
c
us
tom
e
r
r
e
ne
ging
.
T
he
pa
pe
r
is
s
tr
uc
tur
e
d
a
s
f
oll
ows
:
s
e
c
ti
on
2
dis
c
us
s
e
s
the
int
e
r
a
c
ti
on
be
twe
e
n
tande
m
que
ue
ing
theor
y
a
nd
the
f
ield
of
AI
.
S
e
c
ti
on
3
outl
ines
the
methods
us
e
d
a
n
d
pr
ovides
a
na
ly
s
is
.
S
e
c
ti
on
4
pr
e
s
e
nts
a
numer
ica
l
il
lus
tr
a
ti
on.
F
inally
,
s
e
c
ti
on
5
of
f
e
r
s
c
onc
lus
ions
.
2.
I
NT
E
RA
CT
I
ON
B
E
T
WE
E
N
T
HE
T
AN
DE
M
QUE
UE
I
NG
T
HE
ORY
AN
D
T
HE
AI
F
I
E
L
D
T
he
int
e
gr
a
ti
on
o
f
tande
m
que
ue
s
a
nd
AI
c
a
n
r
e
volut
ioni
z
e
ope
r
a
ti
ons
r
e
s
e
a
r
c
h
by
opti
mi
z
ing
que
ue
ing
s
ys
tems
a
nd
e
nha
nc
ing
de
c
is
ion
-
making
pr
oc
e
s
s
e
s
.
AI
-
powe
r
e
d
a
lgor
it
hms
c
a
n
a
na
lyze
a
nd
s
im
ulate
tande
m
que
ue
ope
r
a
ti
ons
,
lea
ding
to
im
pr
ove
d
e
f
f
icie
nc
y
a
nd
r
e
duc
e
d
wa
it
t
im
e
s
.
B
y
le
ve
r
a
ging
mac
hine
lea
r
ning
a
nd
que
ue
ing
theo
r
y
,
r
e
s
e
a
r
c
he
r
s
a
nd
pr
a
c
ti
t
ioner
s
c
a
n
de
ve
lop
innovative
s
olut
ions
f
or
c
ompl
e
x
que
ue
ing
s
ys
tems
.
2.
1
.
Qu
e
u
e
m
an
age
m
e
n
t
AI
c
a
n
be
us
e
d
to
opti
mi
z
e
the
mana
ge
ment
o
f
tan
de
m
que
ue
s
in
va
r
ious
s
ys
tems
,
s
uc
h
a
s
c
us
tom
e
r
s
e
r
vice
c
e
nter
s
,
manuf
a
c
tur
ing
pr
oc
e
s
s
e
s
,
a
nd
tr
a
ns
por
tation
s
ys
tems
.
AI
a
lgo
r
it
hms
c
a
n
p
r
e
dict
que
ue
lengths
a
nd
a
r
r
ival
r
a
tes
,
he
lpi
ng
to
a
ll
oc
a
te
r
e
s
our
c
e
s
e
f
f
icie
ntl
y
a
nd
mi
nim
ize
wa
it
ti
mes
.
P
r
e
dictive
m
a
int
e
na
nc
e
:
i
n
manuf
a
c
tur
ing
s
e
tt
ings
whe
r
e
tan
de
m
que
ue
s
a
r
e
c
omm
on
,
AI
c
a
n
be
uti
li
z
e
d
f
or
p
r
e
dictive
maintena
nc
e
.
B
y
a
na
lyzing
da
ta
f
r
om
s
e
ns
or
s
a
nd
pr
oduc
ti
on
p
r
oc
e
s
s
e
s
,
AI
a
lgor
it
h
ms
c
a
n
a
nti
c
ipate
e
quipm
e
nt
f
a
il
ur
e
s
or
bott
lene
c
ks
in
the
que
ue
a
nd
s
c
he
dule
maintena
nc
e
pr
oa
c
ti
ve
ly,
r
e
duc
ing
downt
im
e
a
nd
im
pr
oving
ove
r
a
ll
e
f
f
icie
nc
y
.
2.
2.
Dynam
ic
r
e
s
ou
r
c
e
all
oc
at
ion
AI
a
lgor
it
hms
c
a
n
dyna
mi
c
a
ll
y
a
ll
oc
a
te
r
e
s
our
c
e
s
withi
n
tande
m
que
ue
s
ba
s
e
d
on
r
e
a
l
-
ti
me
da
ta.
F
or
e
xa
mpl
e
,
in
a
c
a
ll
c
e
nter
,
AI
c
a
n
a
na
lyze
c
us
tom
e
r
inqui
r
ies
a
nd
a
s
s
ign
them
to
the
mos
t
a
ppr
op
r
i
a
te
a
ge
nt
ba
s
e
d
on
f
a
c
tor
s
s
uc
h
a
s
s
kil
l
leve
l,
a
va
il
a
bil
it
y,
a
nd
pa
s
t
pe
r
f
or
manc
e
tec
hniques
.
AI
tec
hniques
s
uc
h
a
s
r
e
inf
or
c
e
ment
lea
r
ning
c
a
n
be
a
ppli
e
d
to
opti
mi
z
e
the
pe
r
f
o
r
manc
e
of
tande
m
que
ue
s
.
B
y
c
onti
nuous
ly
lea
r
ning
f
r
om
pa
s
t
e
xpe
r
ienc
e
s
a
nd
f
e
e
dba
c
k,
A
I
a
lgor
it
hms
c
a
n
a
d
a
pti
ve
ly
a
djus
t
que
ue
ing
s
tr
a
tegie
s
to
im
pr
ove
thr
oughput
,
r
e
duc
e
c
onge
s
ti
on,
a
nd
e
nha
n
c
e
ove
r
a
ll
s
ys
tem
e
f
f
icie
nc
y.
2.
3.
S
i
m
u
lat
io
n
an
d
m
o
d
e
li
n
g
AI
c
a
n
e
nha
nc
e
the
a
c
c
ur
a
c
y
of
tande
m
que
u
e
s
im
ulations
by
leve
r
a
ging
mac
hine
lea
r
ning
a
lgor
it
hms
to
a
na
lyze
his
tor
ica
l
da
ta.
T
he
s
e
a
l
gor
it
hms
c
a
n
identif
y
pa
tt
e
r
ns
a
nd
tr
e
nds
,
e
na
bli
ng
the
de
ve
lopm
e
nt
of
mor
e
r
e
a
li
s
ti
c
s
im
ulation
models
.
B
y
incor
por
a
ti
ng
thes
e
ins
ight
s
,
r
e
s
e
a
r
c
he
r
s
c
a
n
opti
mi
z
e
que
ue
ing
pa
r
a
mete
r
s
a
nd
be
tt
e
r
p
r
e
dict
s
ys
tem
be
ha
vior
,
lea
d
ing
to
im
p
r
ove
d
pe
r
f
or
manc
e
a
nd
e
f
f
icie
nc
y.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
2
,
Apr
il
2
025
:
97
5
-
98
3
978
2.
4.
Dynam
ic
p
r
icin
g
an
d
q
u
e
u
e
m
a
n
age
m
e
n
t
I
n
s
e
r
vice
indus
tr
ies
l
ike
a
i
r
li
ne
s
o
r
theme
pa
r
ks
,
AI
c
a
n
he
lp
opti
mi
z
e
pr
icing
s
tr
a
tegie
s
ba
s
e
d
on
de
mand
a
nd
que
ue
length.
B
y
a
na
lyzing
his
tor
i
c
a
l
da
ta
a
nd
c
ur
r
e
nt
d
e
mand
t
r
e
nds
,
AI
a
lgor
it
hms
c
a
n
dyna
mi
c
a
ll
y
a
djus
t
pr
ice
s
to
ba
lanc
e
s
upply
a
nd
de
mand
a
nd
mi
nim
ize
que
ue
c
onge
s
ti
on.
T
he
no
ve
lt
y
of
int
e
gr
a
ti
ng
tande
m
que
ue
s
with
AI
li
e
s
in
the
a
bil
i
ty
to
ha
r
ne
s
s
a
dva
nc
e
d
c
omput
a
ti
ona
l
tec
hniques
to
tac
kle
c
ompl
e
x
que
ue
i
ng
pr
oblems
in
innovative
wa
ys
.
H
e
r
e
a
r
e
s
ome
a
s
pe
c
ts
c
ontr
ibut
ing
to
the
nove
lt
y
.
2.
5.
P
e
r
s
on
ali
z
e
d
q
u
e
u
e
m
a
n
age
m
e
n
t
AI
-
powe
r
e
d
que
ue
mana
ge
ment
c
a
n
be
tailor
e
d
to
indi
vidual
pr
e
f
e
r
e
nc
e
s
,
taking
in
to
a
c
c
ount
his
tor
ica
l
be
ha
vior
a
nd
other
r
e
leva
nt
f
a
c
tor
s
.
T
his
pe
r
s
ona
li
z
e
d
a
ppr
oa
c
h
e
na
bles
a
mo
r
e
e
f
f
ici
e
nt
a
nd
s
a
ti
s
f
ying
c
us
tom
e
r
e
xpe
r
ienc
e
.
B
y
de
viating
f
r
om
tr
a
dit
ional
one
-
s
ize
-
f
it
s
-
a
ll
methods
,
bus
ines
s
e
s
c
a
n
s
igni
f
ica
ntl
y
e
nha
nc
e
c
us
tom
e
r
s
a
ti
s
f
a
c
ti
on
a
nd
loy
a
lt
y.
2.
6.
I
n
t
e
gr
at
ion
wit
h
in
t
e
r
n
e
t
of
t
h
in
gs
an
d
s
e
n
s
or
d
at
a
AI
-
powe
r
e
d
que
ue
mana
ge
ment
s
ys
tem
s
c
a
n
ha
r
ne
s
s
r
e
a
l
-
ti
me
da
ta
f
r
om
int
e
r
ne
t
of
thi
ng
s
(
I
o
T
)
de
vice
s
a
nd
s
e
ns
or
s
to
ga
in
a
de
e
pe
r
unde
r
s
tanding
of
que
ue
dyna
mi
c
s
.
T
his
int
e
gr
a
ti
on
e
na
bles
mor
e
pr
e
c
is
e
moni
tor
ing
a
nd
c
ontr
o
l
o
f
que
ue
i
ng
pr
oc
e
s
s
e
s
,
f
a
c
il
it
a
ti
ng
da
ta
-
dr
iven
de
c
is
ion
-
making.
B
y
leve
r
a
ging
I
o
T
da
ta,
bus
ines
s
e
s
c
a
n
r
e
s
pond
mor
e
e
f
f
e
c
ti
ve
ly
to
c
ha
nging
que
ue
c
ondit
ions
,
opti
mi
z
ing
the
c
us
tom
e
r
e
xpe
r
ienc
e
.
2.
7.
S
i
m
u
lat
e
d
e
n
vironm
e
n
t
s
f
or
e
xp
e
r
im
e
n
t
at
i
on
AI
e
na
bles
the
c
r
e
a
ti
on
of
s
im
ulate
d
e
nvir
onme
nts
f
or
tes
ti
ng
a
nd
opti
mi
z
ing
que
ue
ing
s
ys
tems
,
a
ll
owing
f
o
r
e
xpe
r
i
menta
ti
on
with
va
r
ious
s
tr
a
tegi
e
s
.
T
his
vir
tual
tes
ti
ng
g
r
ound
f
a
c
il
it
a
tes
r
a
pid
it
e
r
a
ti
on
a
nd
im
pr
ove
ment
without
dis
r
upti
ng
r
e
a
l
-
wor
ld
ope
r
a
ti
ons
.
B
y
leve
r
a
ging
A
I
-
dr
iven
s
im
ulation,
bus
ines
s
e
s
c
a
n
identif
y
opti
mal
que
ue
ing
s
olut
ions
a
nd
im
pleme
nt
them
mor
e
e
f
f
icie
ntl
y
.
2.
8.
Nove
l
ap
p
r
oac
h
e
s
of
ar
t
i
f
icial
in
t
e
ll
igence
t
e
c
h
n
iq
u
e
s
Nove
l
a
ppr
oa
c
he
s
to
que
ue
mana
ge
ment
of
ten
in
volve
c
ombi
ning
mul
ti
ple
AI
tec
hn
iques
,
s
uc
h
a
s
r
e
inf
or
c
e
ment
lea
r
ning
a
nd
de
e
p
lea
r
ning.
B
y
leve
r
a
ging
the
s
tr
e
ngths
o
f
e
a
c
h
tec
hnique,
r
e
s
e
a
r
c
he
r
s
c
a
n
de
ve
lop
mor
e
r
obus
t
a
nd
e
f
f
e
c
ti
ve
s
olut
ions
that
a
ddr
e
s
s
va
r
ious
a
s
p
e
c
ts
of
que
ue
mana
ge
me
nt.
T
his
mul
ti
-
f
a
c
e
ted
a
ppr
oa
c
h
e
na
bles
the
c
r
e
a
ti
on
of
s
o
phis
ti
c
a
ted
que
ue
mana
ge
ment
s
ys
tems
that
c
a
n
a
da
pt
to
c
ha
nging
c
ondit
ions
.
2.
9.
T
h
e
s
ign
if
icance
of
t
h
e
f
ield
T
he
s
e
r
vice
indus
tr
y
ha
s
unde
r
gone
a
tr
a
ns
f
or
mative
s
hif
t
s
ince
the
digi
tal
r
e
volut
ion,
with
a
utom
a
ti
on
r
e
plac
ing
tr
a
dit
ional
c
us
to
mer
s
uppor
t
methods
.
L
e
ve
r
a
ging
AI
a
nd
c
ons
umer
da
ta,
bu
s
ines
s
e
s
c
a
n
now
opti
mi
z
e
ope
r
a
ti
ons
,
r
e
duc
e
c
os
ts
,
a
nd
e
nha
nc
e
c
us
tom
e
r
s
a
ti
s
f
a
c
ti
on.
I
ntegr
a
ti
ng
tande
m
que
ue
s
with
AI
int
r
oduc
e
s
innovative
que
ue
mana
ge
ment
tec
hniques
,
e
leva
ti
ng
e
f
f
icie
nc
y
a
nd
op
e
r
a
ti
ona
l
pe
r
f
or
manc
e
a
c
r
os
s
indus
tr
ies
s
uc
h
a
s
mar
ke
ti
ng,
e
-
c
omm
e
r
c
e
,
s
a
les
,
a
nd
da
ta
s
c
ien
c
e
.
B
y
ha
r
ne
s
s
in
g
r
e
f
ined
c
ons
umer
da
ta,
c
ompanie
s
c
a
n
unlock
AI
's
f
ull
p
otential,
r
e
volut
ion
izing
c
us
tom
e
r
int
e
r
a
c
ti
ons
a
n
d
dr
ivi
ng
gr
owth
thr
ough
s
tr
e
a
ml
ined,
s
oph
is
ti
c
a
ted
ope
r
a
ti
ons
.
3.
M
ODE
L
DE
S
CR
I
P
T
I
ON
I
n
thi
s
s
e
c
ti
on
,
we
c
ons
ider
a
two
-
ter
mi
na
l
s
ys
tem
whe
r
e
c
us
tom
e
r
s
r
e
quir
e
s
e
r
vice
s
a
t
both
e
nds
.
C
us
tom
e
r
s
e
nter
the
s
ys
tem
a
t
1
,
que
ue
f
o
r
one
s
e
r
v
ice
a
s
they
e
xit
,
a
nd
then
go
to
the
s
e
c
ond
s
e
r
vice
to
us
e
it
.
T
he
AI
modul
e
s
it
e
s
be
twe
e
n
the
que
ue
s
a
nd
s
e
r
ve
s
a
s
the
int
e
ll
igent
pr
oc
e
s
s
ing
lay
e
r
,
F
igur
e
1
r
e
pr
e
s
e
nt
the
AI
-
dr
iven
tande
m
que
ue
ing
model
.
T
his
diagr
a
m
il
lus
tr
a
tes
a
tan
de
m
que
ue
s
ys
tem
opti
mi
z
e
d
by
AI
.
Ar
r
ival
c
us
tom
e
r
s
f
ir
s
t
e
nter
s
e
r
ve
r
1
,
whe
r
e
a
por
ti
on
of
them
may
r
e
ne
ge
a
nd
lea
ve
the
que
ue
,
r
e
p
r
e
s
e
nted
by
(
1−
θ)
λ
.
T
he
AI
s
ys
tem
dyna
mi
c
a
ll
y
mana
ge
s
t
he
f
low
be
twe
e
n
s
e
r
ve
r
1
a
nd
s
e
r
ve
r
2
to
opti
mi
z
e
ove
r
a
ll
que
ue
e
f
f
icie
nc
y.
C
us
tom
e
r
s
s
uc
c
e
s
s
f
ull
y
pr
oc
e
s
s
e
d
by
s
e
r
ve
r
1,
a
t
a
r
a
te
1
,
a
r
e
then
di
r
e
c
ted
to
s
e
r
ve
r
2
,
whic
h
pr
oc
e
s
s
e
s
them
a
t
a
r
a
te
2
,
lea
ding
to
their
e
ve
ntual
de
pa
r
tur
e
f
r
om
the
s
ys
tem.
3.
1
.
Not
a
t
ion
s
AI
modul
e
:
-
a
r
r
ival
r
a
te
a
t
s
e
r
ve
r
1
;
1
-
s
e
r
vice
r
a
te
a
t
s
e
r
ve
r
1
;
2
-
s
e
r
vice
r
a
te
a
t
s
e
r
ve
r
2
;
1
-
Numbe
r
of
c
li
e
nts
on
s
e
r
ve
r
1
.
2
-
Numbe
r
of
c
li
e
nts
on
s
e
r
ve
r
2
.
-
R
e
ne
ging
c
li
e
nt
f
r
om
s
e
r
ve
r
1
.
1
=
1
,
2
=
2
=
tr
a
f
f
ic
de
ns
it
y.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
Optimiz
ing
que
ue
e
ff
icie
nc
y
:
A
r
ti
fi
c
ial
int
e
ll
igenc
e
-
dr
ive
n
tandem
que
ue
s
…
(
K
e
e
r
thi
k
a
R
adhak
r
is
hn
an)
979
F
igur
e
1.
A
I
-
dr
iven
tande
m
que
ue
ing
model
I
nc
or
por
a
ti
ng
AI
to
de
ter
mi
ne
or
op
ti
mi
z
e
the
ti
me
t+h
in
a
bir
th
-
de
a
th
pr
oc
e
s
s
is
a
n
int
e
r
e
s
ti
ng
a
ppli
c
a
ti
on
that
c
a
n
he
lp
a
da
pt
the
model’
s
ti
me
r
e
s
olut
ion.
‒
1
,
2
u
nit
s
in
the
s
ys
tem
a
t
ti
me
t
a
nd
no
a
r
r
ival
a
nd
no
s
e
r
vice
dur
ing
t+h
in
both
s
e
r
vice
s
tations
‒
1
−
1
,
2
u
nit
s
in
the
s
ys
tem
a
t
ti
me
t
,
one
a
r
r
ival
a
nd
no
s
e
r
vice
a
t
s
e
r
ve
r
1,
no
a
r
r
ival
a
nd
no
s
e
r
vice
a
t
s
e
r
ve
r
2
du
r
ing
t+h
‒
1
+
1
,
2
−
1
u
nit
s
in
the
s
ys
tem
a
t
ti
me
t
,
no
a
r
r
ival
a
nd
one
s
e
r
vice
a
t
s
tation
1,
one
a
r
r
ival
a
nd
no
s
e
r
vice
a
t
s
tation
2
dur
ing
t+h
‒
1
,
2
+
1
u
nit
s
in
the
s
ys
tem
a
t
ti
me
t
,
no
a
r
r
ival
a
nd
no
s
e
r
vice
a
t
s
e
r
ve
r
1,
no
a
r
r
ival
a
nd
one
s
e
r
vice
a
t
s
e
r
ve
r
2
du
r
ing
t+h
‒
1
+
1
,
2
u
nit
s
in
the
s
ys
tem
a
t
ti
me
t
,
no
a
r
r
ival
a
nd
one
s
e
r
vice
a
t
s
e
r
ve
r
1,
no
a
r
r
ival
a
nd
no
s
e
r
vice
a
t
s
e
r
ve
r
2
du
r
ing
t+h
1
,
2
(
+
ℎ
)
=
1
,
2
(
)
(
1
−
ℎ
)
(
1
−
1
ℎ
)
(
1
−
1
(
1
−
)
)
(
1
−
2
ℎ
)
+
1
−
1
,
2
(
)
ℎ
(
1
−
1
ℎ
)
(
1
−
1
(
1
−
)
)
(
1
−
2
ℎ
)
1
ℎ
1
+
1
,
2
−
1
(
)
(
1
−
ℎ
)
(
1
−
2
ℎ
)
+
(
1
−
)
1
ℎ
1
+
1
,
2
(
)
(
1
−
ℎ
)
(
1
−
2
ℎ
)
+
1
,
2
+
1
(
)
(
1
−
ℎ
)
(
1
−
1
ℎ
)
(
1
−
1
(
1
−
)
)
2
ℎ
(
1)
S
im
pli
f
ying
(
1)
gives
:
1
,
2
(
+
ℎ
)
=
1
,
2
(
1
−
1
−
2
−
)
+
ℎ
1
−
1
,
2
+
1
ℎ
1
+
1
,
2
−
1
+
(
1
−
)
1
ℎ
1
+
1
,
2
+
2
ℎ
1
,
2
+
1
(
2)
l
im
ℎ
→
0
1
2
1
2
,,
(
)
(
)
K
K
K
K
Q
t
h
P
t
h
+−
=
−
1
,
2
(
1
+
2
+
)
+
1
−
1
,
2
+
1
1
+
1
,
2
−
1
+
(
1
−
)
1
1
+
1
,
2
+
2
1
,
2
+
1
(
3)
At
s
tea
dy
s
tate
(
i.
e
.
→
∞
)
1
,
2
(
)
=
0
.
He
r
e
(
)
is
no
longer
a
f
unc
ti
on
of
t.
I
n
(
3
)
now
be
c
omes
:
1
,
2
(
1
+
2
+
)
=
1
−
1
,
2
+
1
1
+
1
,
2
−
1
+
(
1
−
)
1
1
+
1
,
2
+
2
1
,
2
+
1
(
4)
T
his
e
qua
ti
on
is
known
a
s
the
s
tea
dy
-
s
tate
di
f
f
e
r
e
nc
e
e
qua
ti
on.
Now
,
the
bounda
r
y
c
ondit
ion
f
or
th
is
model
is
a
s
f
oll
ows
:
W
he
n,
1
=
0
,
(
4)
be
c
omes
(
+
2
)
0
,
2
=
1
1
,
2
−
1
+
(
1
−
)
1
1
,
2
+
2
0
,
2
+
1
(
5)
W
he
n,
2
=
0,
(
4)
be
c
omes
(
+
1
)
1
,
0
=
1
−
1
,
0
+
(
1
−
)
1
1
+
1
,
0
+
2
1
,
1
(
6)
W
he
n,
1
=
0
,
2
=
0
,
(
4)
be
c
omes
0
,
0
=
(
1
−
)
1
1
,
0
+
2
0
,
1
(
7)
3.
2.
P
r
ob
a
b
il
it
y
ge
n
e
r
at
in
g
f
u
n
c
t
ion
T
he
pr
oba
bil
it
y
-
ge
ne
r
a
ti
ng
f
unc
ti
on
s
e
r
ve
s
a
s
a
c
r
uc
ial
tool
f
o
r
a
na
lyzing
a
nd
unde
r
s
tanding
the
pr
ope
r
ti
e
s
of
r
a
ndom
va
r
iable
s
.
T
h
is
f
unc
ti
on
e
n
a
bles
r
e
s
e
a
r
c
he
r
s
to
c
a
ptur
e
the
unde
r
lyi
ng
pr
oba
bil
it
y
dis
tr
ibut
ion
of
a
s
ys
tem.
W
e
de
f
ine
the
p
r
oba
bil
it
y
-
ge
ne
r
a
ti
ng
f
unc
ti
on
a
s
f
oll
ows
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
2
,
Apr
il
2
025
:
97
5
-
98
3
980
1
,
2
(
1
,
2
)
=
∑
∑
1
1
2
2
1
,
2
∞
2
=
0
∞
1
=
0
;
1
,
0
(
1
)
=
∑
1
1
1
,
0
∞
1
=
0
0
,
2
(
2
)
=
∑
2
2
0
,
2
∞
2
=
0
;
1
,
1
(
1
)
=
∑
1
1
1
,
1
∞
1
=
0
;
1
,
2
(
2
)
=
∑
2
2
1
,
2
∞
2
=
0
(
8)
3.
3.
S
olu
t
ion
of
t
h
e
s
t
e
ad
y
-
s
t
at
e
p
r
ob
a
b
il
it
y
Now
s
im
pli
f
ying
the
lef
t
-
ha
nd
s
ide
of
(
4)
we
ge
t
:
(
1
+
2
+
)
∑
∑
1
1
2
2
1
,
2
=
∞
2
=
1
∞
1
=
1
(
1
+
2
+
)
∑
1
1
{
∑
2
2
1
,
2
−
1
,
0
}
∞
2
=
0
∞
1
=
1
(
9)
=
(
1
+
2
+
)
.
He
nc
e
:
∑
∑
1
1
2
2
1
,
2
=
∞
2
=
1
∞
1
=
1
[
∑
1
1
∑
2
2
1
,
2
−
∑
1
1
1
,
0
∞
1
=
1
∞
2
=
0
∞
1
=
1
]
=
[
1
,
2
(
1
,
2
)
−
0
,
2
(
2
)
−
1
,
0
(
1
)
+
0
,
0
]
(
10)
T
he
f
i
r
s
t
it
e
m
on
the
R
HS
of
(
4)
is
:
∑
∑
1
1
2
2
1
−
1
,
2
=
∞
2
=
1
∞
1
=
1
1
[
∑
∑
1
1
−
1
2
2
1
−
1
,
2
∞
2
=
1`
∞
1
=
1
]
=
1
[
∑
1
1
∑
2
2
1
,
2
∞
2
=
0
∞
1
=
0
]
−
1
∑
1
1
∞
1
=
0
1
,
0
=
1
1
,
2
(
1
,
2
)
−
1
1
,
0
(
1
)
(
11)
T
he
s
e
c
ond
ter
m
on
the
R
HS
o
f
(
4)
is
:
1
1
+
1
,
2
−
1
=
1
∑
∑
1
1
2
2
∞
2
=
1
∞
1
=
1
1
+
1
,
2
−
1
=
1
1
−
1
2
(
∑
∑
1
1
2
2
1
,
2
−
∑
2
2
0
,
2
−
1
∑
2
2
1
,
2
∞
2
=
0
∞
2
=
0
∞
2
=
0
∞
1
=
0
)
=
1
1
−
1
2
1
,
2
(
1
,
2
)
−
1
1
−
1
2
0
,
2
(
2
)
−
1
2
1
,
2
(
2
)
(
12)
T
he
thi
r
d
ter
m
on
the
R
HS
of
(
4)
is
:
(
1
−
)
1
1
+
1
,
2
=
(
1
−
)
1
∑
∑
1
1
∞
2
=
1
∞
1
=
1
2
2
1
+
1
,
2
=
(
1
−
)
1
1
−
1
∑
∑
1
1
+
1
∞
2
=
1
∞
1
=
1
2
2
1
+
1
,
2
(
13)
L
e
t
=
(
1
−
)
1
1
−
1
then
(
13)
be
c
ome,
(
1
−
)
1
1
+
1
,
2
=
∑
1
1
+
1
[
∑
2
2
1
+
1
,
2
−
1
+
1
,
0
∞
2
=
0
]
∞
1
=
1
=
[
1
,
2
(
1
,
2
)
−
0
,
2
(
2
)
−
1
1
,
2
(
2
)
−
1
,
0
(
1
)
+
0
,
0
+
1
1
.
0
]
(
14)
T
he
f
our
th
it
e
m
on
the
R
HS
of
(
14)
is
,
2
1
,
2
+
1
=
2
2
−
1
∑
∑
1
1
2
2
+
1
1
,
2
+
1
∞
2
=
1
∞
1
=
1
=
2
2
−
1
[
∑
∑
1
1
2
2
1
,
2
−
∑
2
2
0
,
2
−
∑
1
1
1
,
0
+
0
,
0
−
2
0
,
1
∞
1
=
0
∞
2
=
0
∞
2
=
0
∞
1
=
0
]
He
nc
e
2
1
,
2
+
1
e
qua
ls
:
=
2
2
−
1
[
1
,
2
(
1
2
)
]
−
0
,
2
(
2
)
−
1
,
0
(
1
)
+
0
,
0
−
2
1
,
2
(
2
)
+
1
0
,
1
(
15)
3.
4.
E
xp
e
c
t
e
d
n
u
m
b
e
r
o
f
c
u
s
t
om
e
r
s
in
t
h
e
s
ys
t
e
m
T
he
e
xpe
c
ted
number
of
pe
ople
in
the
s
ys
tem
a
t
s
tea
dy
s
tate
f
or
a
model
with
two
s
tations
c
a
n
be
obtaine
d
a
s
f
oll
ow
s:
(
1
,
2
)
=
∑
∑
1
,
2
1
,
2
∞
2
=
0
∞
1
=
0
(
16)
T
he
e
qua
ti
on
(
16
)
c
a
n
be
wr
it
ten
a
s
:
(
1
,
2
)
=
∑
∑
1
,
2
1
0
,
2
(
1
)
2
1
,
0
∞
2
=
0
∞
1
=
0
(
1
,
2
)
=
(
1
−
1
)
(
1
−
2
)
∑
∑
1
,
2
1
2
∞
2
=
0
∞
1
=
0
W
he
r
e
0
,
2
(
1
)
=
1
−
1
&
1
,
0
(
1
)
=
1
−
2
,
(
1
,
2
)
=
(
1
−
1
)
(
1
−
2
)
∑
∑
1
,
2
1
2
∞
2
=
0
∞
1
=
0
=
1
2
(
1
−
1
)
(
1
−
2
)
(
17)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
Optimiz
ing
que
ue
e
ff
icie
nc
y
:
A
r
ti
fi
c
ial
int
e
ll
igenc
e
-
dr
ive
n
tandem
que
ue
s
…
(
K
e
e
r
thi
k
a
R
adhak
r
is
hn
an)
981
3.
5.
M
e
an
wai
t
in
g
t
im
e
T
he
mea
n
wa
it
ing
ti
me
in
a
two
-
s
tation
s
y
s
tem
c
a
n
be
c
a
lcula
ted
u
s
ing
a
s
tea
dy
-
s
tate
a
na
lys
i
s
,
pr
ovidi
ng
va
luable
ins
ight
s
int
o
s
ys
tem
pe
r
f
or
manc
e
.
T
his
c
a
lcula
ti
on
take
s
int
o
a
c
c
ount
the
a
r
r
iv
a
l
r
a
tes
,
s
e
r
vice
r
a
tes
,
a
nd
other
ke
y
f
a
c
tor
s
that
inf
luenc
e
wa
it
ing
ti
mes
.
B
y
de
ter
mi
ning
the
mea
n
wa
it
i
ng
ti
me,
r
e
s
e
a
r
c
he
r
s
a
nd
pr
a
c
ti
ti
one
r
s
c
a
n
opti
mi
z
e
s
ys
tem
de
s
ign
a
nd
im
pr
ove
ove
r
a
ll
e
f
f
icie
nc
y.
E
(
W
)
=
(
1
1
−
1
)
0
1
(
2
+
−
2
−
2
2
−
1
)
=
(
1
1
−
1
)
0
(
2
2
+
2
−
2
2
−
2
(
18)
4.
NU
M
E
RI
C
AL
I
L
L
UST
RA
T
I
ON
I
n
t
h
i
s
s
e
c
t
i
on
,
t
h
e
o
r
e
t
ic
a
l
r
e
s
u
l
ts
a
r
e
j
us
t
i
f
i
e
d
w
i
t
h
s
u
i
t
a
b
le
n
u
m
e
r
ic
a
l
r
e
s
u
l
ts
.
I
n
o
r
d
e
r
t
o
s
t
u
d
y
t
h
e
e
f
f
e
c
t
o
f
a
r
r
i
v
a
l
r
a
te
,
s
e
r
v
i
c
e
r
a
t
e
a
t
n
od
e
1
a
nd
s
e
r
v
ic
e
r
a
te
a
t
n
o
de
2
.
F
r
o
m
F
i
g
u
r
e
2
a
n
d
T
a
b
l
e
1
,
w
e
c
a
n
o
bs
e
r
v
e
th
a
t
t
h
e
a
r
r
i
va
l
r
a
t
e
in
c
r
e
a
s
e
s
w
he
n
AI
i
s
d
r
iv
e
n
in
t
h
is
m
o
d
e
l
t
he
n
m
e
a
n
wa
i
t
i
n
g
c
us
t
o
me
r
s
s
p
e
e
d
i
ly
d
e
c
r
e
a
s
e
s
.
T
a
ble
1.
Ar
r
ival
r
a
te
vs
.
mea
n
wa
it
ing
ti
me
1
=
0
.
5
,
2
=
0
.
4
,
=
0
.
2
,
=
0
.
8
Χ
S
e
r
ve
r
1
S
e
r
ve
r
2
S
ys
te
m
1
1.7884
2.4564
4.2448
2
0.65
2.2653
2.9153
3
0.9907
2.006
2.9967
4
0.6505
1.856
2.5065
5
0.4808
1.75
2.2308
6
0.2056
1.6503
1.8559
F
igur
e
2.
AI
-
dr
iven,
a
r
r
ival
r
a
te
vs
.
wa
it
ing
ti
me
5.
CONC
L
USI
ON
T
his
pa
pe
r
int
r
oduc
e
s
a
n
innovative
a
ppr
oa
c
h
t
o
tande
m
que
ue
a
na
lys
is
us
ing
AI
tec
hniques
.
AI
-
dr
iven
tande
m
que
ue
a
na
lys
is
r
e
pr
e
s
e
nts
a
tr
a
ns
f
or
mative
f
o
r
c
e
a
c
r
os
s
indus
tr
ies
,
of
f
e
r
ing
mul
t
if
a
c
e
ted
a
dva
ntage
s
.
I
n
thi
s
a
r
ti
c
le,
we
c
ons
ider
a
tande
m
que
ue
w
it
h
two
s
e
r
vice
s
tations
u
s
ing
AI
tec
hniq
ue
s
.
T
he
s
tea
dy
-
s
tate
dif
f
e
r
e
nc
e
e
qua
ti
on
f
or
thi
s
model
i
s
de
r
ived
a
nd
a
m
e
a
n
number
of
c
us
tom
e
r
s
a
r
e
de
r
ived.
C
ompanie
s
c
a
n
r
e
volut
ioni
z
e
ope
r
a
ti
ons
thr
o
ugh
a
utom
a
ted
que
ue
ing,
s
tr
e
a
ml
ini
ng
da
ta
a
na
lys
is
,
a
c
c
umul
a
ti
on,
a
nd
A
I
-
dr
iven
c
us
tom
e
r
s
olut
ion
s
to
dr
ive
bus
ines
s
s
uc
c
e
s
s
.
E
xis
ti
ng
models
may
lac
k
r
obus
tnes
s
to
s
uc
h
unc
e
r
tainty,
lea
ding
to
s
ubopti
mal
pe
r
f
o
r
manc
e
in
pr
a
c
ti
c
e
.
R
e
c
e
nt
obs
e
r
va
ti
ons
s
ugge
s
t
that
the
e
f
f
icie
nc
y
of
tande
m
que
ue
s
c
a
n
s
igni
f
ica
ntl
y
f
lu
c
tuate
due
to
unpr
e
dicta
ble
c
us
tom
e
r
r
e
ne
ging
be
ha
vior
.
Our
f
indi
ngs
p
r
ovide
c
onc
lus
ive
e
vid
e
nc
e
that
thi
s
phe
nomenon
is
a
s
s
oc
iate
d
with
dyna
mi
c
c
ha
nge
s
in
s
e
r
vice
c
ondit
ions
a
nd
a
r
r
ival
r
a
tes
,
not
due
to
e
va
luate
d
number
s
of
s
e
r
ve
r
s
or
s
tati
c
que
ue
pa
r
a
me
ter
s
.
F
utu
r
e
s
tudi
e
s
may
e
xplor
e
the
int
e
g
r
a
ti
on
of
r
obus
t
mac
hine
lea
r
ning
tec
hniques
with
f
e
a
s
ibl
e
wa
ys
of
p
r
oduc
ing
a
da
pti
ve
,
r
e
a
l
-
ti
me
opti
mi
z
a
ti
ons
to
f
ur
the
r
e
nha
nc
e
que
ue
e
f
f
icie
nc
y
a
nd
c
us
tom
e
r
s
a
ti
s
f
a
c
ti
on.
F
utur
e
r
e
s
e
a
r
c
h
s
hould
e
xplo
r
e
tec
hnique
s
f
or
r
obus
t
que
ue
mana
ge
ment
that
c
a
n
ha
ndle
unc
e
r
tainty
e
f
f
e
c
ti
ve
ly.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
2
,
Apr
il
2
025
:
97
5
-
98
3
982
RE
F
E
RE
NC
E
S
[
1]
G.
D
.
A
de
poj
u,
“
S
ta
ti
s
ti
c
a
l
a
na
ly
s
is
of
ta
nde
m
que
ue
s
w
it
h
ma
r
kovi
a
n
por
ous
me
di
ums
,”
M
.A
.
T
he
s
is
,
D
e
pa
r
tm
e
nt
of
M
a
th
e
ma
ti
c
s
,
M
a
r
s
ha
ll
U
ni
ve
r
s
it
y,
H
unt
in
gt
on
, U
S
A
,
2019.
[
2]
J
.
W
a
ng,
H
.
A
boue
e
-
M
e
hr
iz
i,
O
.
B
a
r
on,
a
nd
O
.
B
e
r
ma
n,
“
T
a
nde
m
que
ue
s
w
it
h
im
pa
ti
e
nt
c
us
to
me
r
s
,”
P
e
r
fo
r
m
anc
e
E
v
al
ua
ti
on
,
vol
. 135, Nov. 2019, do
i:
10.1016/j
.pe
va
.2019.102011.
[
3]
A
.
A
.
A
lf
r
e
d,
“
B
lo
c
ke
d
ne
twor
k
o
f
ta
nde
m
que
ue
s
w
it
h
w
i
th
dr
a
w
a
l,
”
K
r
aguj
e
v
ac
J
o
u
r
nal
of
M
at
he
m
at
ic
s
,
vol
.
23,
pp.
63
–
73,
2001.
[
4]
G
.
A
yya
ppa
n
a
nd
P
.
T
ha
mi
z
h
s
e
lv
,
“
T
r
a
n
s
ie
nt
a
n
a
ly
s
is
of
M
[
X1
]
,
M
[
X2
]
/
G
1
,
G
2
(
a
,b)
/1
que
ue
in
g
s
y
s
te
m
w
it
h
pr
io
r
it
y
s
e
r
vi
c
e
s
,
”
B
ul
le
ti
n of
P
ur
e
and A
ppl
ie
d Sc
ie
nc
e
s
, vol
. 36, no. 2, pp. 112
–
132, 2017.
[
5]
X
.
Y
ua
n,
H
.
Y
a
o,
J
.
W
a
ng,
T
.
M
a
i
a
nd
M
.
G
ui
z
a
ni
,
“
A
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
e
mpowe
r
e
d
Q
oS
-
or
ie
nt
e
d
ne
twor
k
a
s
s
oc
ia
ti
on
f
or
ne
xt
-
ge
ne
r
a
ti
on
mobi
le
ne
twor
ks
,”
in
I
E
E
E
T
r
ans
ac
ti
ons
on
C
ogni
ti
v
e
C
om
m
uni
c
at
io
ns
and
N
e
tw
or
k
in
g
,
vol
.
7,
no
.
3,
pp. 856
-
870, S
e
pt
. 2021, doi:
10.1109/T
C
C
N
.2021.3065463.
[
6]
C.
F
e
r
e
nc
z
a
nd
M
.
Z
öl
dy
,
“
R
o
a
d
tr
a
f
f
ic
que
ue
le
ngt
h
e
s
ti
ma
ti
on
w
it
h
a
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
(
A
I
)
me
th
ods
,
”
C
ogni
ti
v
e
s
us
ta
in
abi
li
ty
, vol
2, no 3, 2023
,
do
i:
10.55343/cogs
us
t.
65
.
[
7]
P
.
A
nus
s
or
n
ni
ti
s
a
r
n
a
nd
V
.
L
im
la
w
a
n,
“
D
e
s
ig
n
of
a
dva
nc
e
d
que
ue
s
ys
t
e
m
us
in
g
a
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
k
f
or
w
a
it
in
g
ti
me
pr
e
di
c
ti
on,”
R
e
v
is
ta
e
s
pac
io
s
, vol
. 41, no. 40, pp. 111
–
124, 2020.
[
8]
F
. K
a
moun, “
P
e
r
f
or
ma
nc
e
a
na
ly
s
is
of
t
w
o pr
io
r
it
y
que
ui
ng s
ys
te
ms
i
n t
a
nde
m,”
A
m
e
r
ic
an J
our
nal
of
O
pe
r
at
io
ns
R
e
s
e
a
r
c
h
, vol
. 2,
no. 4, pp. 509
–
518, 2012, doi:
10.4236/ajor
.2012.24060.
[
9]
Y
.
L
iu
a
nd
R
.
Q
ua
n, “
R
e
s
e
a
r
c
h a
nd
a
n
a
ly
s
is
on
th
e
in
te
r
a
c
ti
on be
twe
e
n
que
ui
ng
th
e
or
y
a
nd a
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
,”
in
P
r
oc
e
e
di
ngs
of
th
e
2023
I
nt
e
r
n
at
io
nal
C
onf
e
r
e
nc
e
on
I
m
age
,
A
lg
or
it
hm
s
and
A
r
ti
fi
c
ia
l
I
nt
e
ll
ig
e
nc
e
(
I
C
I
A
A
I
2023
)
,
A
tl
a
nt
is
P
r
e
s
s
,
2023,
pp. 391
–
401
, doi
:
10.2991/978
-
94
-
6463
-
300
-
9_40.
[
10]
M
.
S
.
C
he
n
a
nd
H
.
W
.
Y
e
n,
“
A
ppl
ic
a
ti
on
s
of
ma
c
hi
ne
le
a
r
ni
ng
a
ppr
oa
c
h
on
mul
ti
-
que
ue
me
s
s
a
ge
s
c
he
dul
in
g,”
E
x
p
e
r
t
Sy
s
te
m
s
w
it
h A
ppl
ic
at
io
ns
, vol
. 38, no. 4, pp. 3323
–
3335, 2011, doi:
10.1016/j
.e
s
w
a
.2010.08.117.
[
11]
G
.
A
yya
ppa
n
a
nd
K
.
T
hi
la
ga
va
th
y,
“
M
a
c
hi
n
e
le
a
r
ni
ng
a
na
ly
s
is
of
que
ue
s
w
it
h
m
a
p,
r
e
ne
gi
ng,
pha
s
e
ty
pe
s
e
r
vi
c
e
s
,
va
c
a
ti
ons
a
nd
r
e
pa
ir
s
,”
A
C
M
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
P
r
oc
e
e
di
ng Se
r
ie
s
, pp.
197
–
208, 2021, doi:
10.1145/3484824.3
484872.
[
12]
L
. M
a
s
, J
. V
il
a
pl
a
na
, J
. M
a
te
o, a
nd F
. S
ol
s
ona
, “
A
que
ui
ng t
h
e
or
y mode
l
f
o
r
f
og
c
omput
in
g,”
J
our
nal
of
Supe
r
c
om
put
in
g
, vol
. 78,
no. 8, pp. 11138
–
11155, 2022, doi:
10.1007/s
11227
-
022
-
04328
-
3.
[
13]
Y
.
L
i
,
W
.
G
ua
n,
a
nd
W
.
L
i,
“
Q
ue
ui
ng
th
e
or
y
a
nd
a
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
in
s
e
r
vi
c
e
s
ys
te
ms
:
a
s
ur
ve
y,”
M
at
he
m
at
ic
s
,
vol
.
8
,
no
.
5,
2020.
[
14]
C
.
S
he
kha
r
,
N
.
K
uma
r
,
A
.
G
upt
a
,
a
nd
R
.
K
.
T
iwa
r
i,
“
F
in
it
e
c
a
pa
c
it
y
ta
nde
m
que
ue
in
g
ne
twor
k
w
it
h
r
e
ne
gi
ng,”
in
M
at
he
m
at
ic
al
M
ode
li
ng and C
om
put
at
io
n of
R
e
al
-
T
ime
P
r
obl
e
m
s
, B
oc
a
R
a
to
n:
C
R
C
P
r
e
s
s
, 2021, pp. 33
–
46
, doi
:
10.1201/978100305
503
7
-
4.
[
15]
K
.
P
.
S
.
B
a
ghe
l
a
nd
M
.
J
a
in
,
“
T
r
a
ns
ie
nt
s
ol
ut
io
n
of
ma
r
kov
mode
l
f
or
f
a
ul
t
to
le
r
a
nt
s
ys
te
m
w
it
h
r
e
dunda
nc
y,”
in
M
at
he
m
at
ic
al
M
ode
li
ng and C
om
put
at
io
n of
R
e
al
-
T
ime
P
r
obl
e
m
s
, C
R
C
P
r
e
s
s
, 2021
, doi
:
10.1201/978100305
5037
-
5.
[
16]
S
.
S
ha
nmuga
s
unda
r
a
m
a
nd
P
.
U
ma
r
a
ni
,
“
Q
ue
ue
in
g
th
e
or
y
a
p
pl
ie
d
in
our
da
y
to
da
y
li
f
e
,”
I
nt
e
r
nat
io
nal
J
our
nal
of
Sc
ie
nt
if
ic
&
E
ngi
ne
e
r
in
g R
e
s
e
ar
c
h
, vol
. 6, no. 4, pp. 533
–
541, 2015
.
[
17]
J
.
N
di
a
ye
,
O
.
S
ow
,
Y
.
T
r
a
or
e
,
M
.
A
.
D
io
p,
A
.
S
.
F
a
ye
,
a
nd
A
.
D
io
p,
“
E
le
c
tr
oni
c
s
ys
te
m
us
in
g
a
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
f
or
q
ue
ue
ma
na
ge
me
nt
,”
O
pe
n J
ou
r
nal
of
A
ppl
ie
d Sc
ie
nc
e
s
, vol
. 12, no. 12, pp. 2019
–
2036, 2022, doi:
10.4236/oj
a
pps
.2022.1212141.
[
18]
J
. S
ti
nt
z
in
g a
nd F
.
N
or
r
ma
n,
“
P
r
e
di
c
ti
on of
que
ui
ng be
ha
vi
our
th
r
ough the
us
e
of
a
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
ks
,”
M
.
Sc
.
T
he
s
is
,
S
c
hool
o
f
C
omput
e
r
S
c
ie
nc
e
a
nd C
omm
uni
c
a
ti
on
,
K
th
R
oya
l
I
ns
ti
tu
te
o
f
T
e
c
hnol
ogy
,
S
to
c
khol
m,
S
w
e
de
n,
2017.
[
19]
C
.
Y
uva
r
a
ni
a
nd
C
.
V
ij
a
ya
la
ks
hmi
,
“
A
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
k
a
ppr
oa
c
h
in
de
s
ig
n
of
que
ue
in
g
mode
ls
,”
in
St
oc
has
ti
c
P
r
oc
e
s
s
e
s
and T
he
ir
A
ppl
ic
at
io
ns
i
n A
r
ti
fi
c
ia
l
I
nt
e
ll
ig
e
nc
e
, 2023, pp. 23
–
38
, doi
:
10.4018/978
-
1
-
6684
-
7679
-
6.c
h003.
[
20]
C
.
R
ove
tt
o
e
t
al
.
,
“
M
in
im
iz
in
g
in
te
r
s
e
c
ti
on
w
a
it
in
g
ti
me
:
pr
opos
a
l
of
a
que
ue
ne
twor
k
mode
l
us
in
g
ke
nd
a
ll
’
s
not
a
ti
on
in
pa
na
ma
c
it
y,”
A
ppl
ie
d Sc
ie
nc
e
s
, vol
. 13, no. 18, 2023, doi
:
10.3390/a
pp
131810030.
[
21]
M
.
S
.
S
unda
r
i
a
nd
S
.
P
a
la
ni
a
mm
a
l,
“
S
im
ul
a
ti
on
of
m
/m
/1
que
ui
ng
s
ys
te
m
us
in
g
a
nn
que
ui
ng
mode
ls
,”
M
al
ay
a
J
our
nal
of
M
at
e
m
at
ik
, vol
. 5, no. 1, pp. 279
–
294, 2015.
[
22]
T
.
M
.
V
A
nur
uddhika
,
S
.
P
r
a
s
a
nt
h,
a
nd
R
.
M
.
K
.
T
.
R
a
th
na
y
a
ka
,
“
T
he
a
ppr
oa
c
he
s
ut
il
iz
e
d
in
que
ui
ng
mode
li
ng:
a
s
y
s
te
m
a
ti
c
li
te
r
a
tu
r
e
r
e
vi
e
w
,”
A
s
ia
n
J
our
nal
of
C
onv
e
r
ge
n
c
e
i
n
T
e
c
hnol
ogy
,
vol
.
8,
no.
2,
pp.
24
–
30,
2022,
doi
:
10.33130/ajc
t.
2022v08i02.006.
[
23]
S
.
S
id
di
qui
,
M
.
D
a
r
ba
r
i,
a
nd
D
.
Y
a
gya
s
e
n,
“
M
ode
ll
in
g
a
nd
s
im
ul
a
ti
on
of
que
ui
ng
mode
ls
th
r
ough
th
e
c
onc
e
pt
of
pe
tr
i
ne
ts
,”
A
dc
ai
j
-
A
dv
anc
e
s
i
n D
is
tr
ib
ut
e
d C
om
put
in
g and A
r
ti
fi
c
ia
l
I
nt
e
ll
ig
e
nc
e
J
our
nal
, vol
. 9, no. 3, pp. 17
–
28, 2020.
[
24]
H
.
B
a
uma
nn
a
nd
W
.
S
a
ndma
nn,
“
M
ul
ti
-
s
e
r
ve
r
ta
nde
m
que
ue
w
it
h
ma
r
kovi
a
n
a
r
r
iv
a
l
pr
oc
e
s
s
,
pha
s
e
-
ty
pe
s
e
r
vi
c
e
ti
me
s
,
a
nd
f
in
it
e
buf
f
e
r
s
,”
E
ur
ope
an J
our
nal
of
O
pe
r
at
io
nal
R
e
s
e
ar
c
h
, vol
. 256,
no. 1, pp. 187
–
195, 2017, doi:
10.1016/j
.e
jo
r
.2016.07.035.
[
25]
S
.
P
.
N
ir
a
nj
a
n,
S
.
D
.
L
a
th
a
,
M
.
M
a
hda
l,
a
nd
K
.
K
a
r
th
ik
,
“
M
ul
ti
pl
e
c
ont
r
ol
pol
ic
y
in
unr
e
li
a
bl
e
two
-
pha
s
e
bul
k
que
u
e
in
g
s
y
s
te
m
w
it
h a
c
ti
ve
be
r
noul
li
f
e
e
dba
c
k a
nd va
c
a
ti
on,”
M
at
he
m
at
ic
s
, vol
. 12, no. 1, 2024, doi
:
10.3390/m
a
th
12010075.
B
I
OG
RA
P
HI
E
S
OF
AU
T
HO
R
Keerthi
ka
R
a
dha
kr
i
s
hn
a
n
c
u
rren
t
l
y
p
u
r
s
u
i
n
g
h
e
r
d
o
ct
o
r
o
f
p
h
i
l
o
s
o
p
h
y
(Ph
.
D
.
)
d
eg
ree
at
V
el
T
ec
h
Ran
g
a
raj
a
n
D
r.
Sag
u
n
t
h
al
a
R&D
I
n
s
t
i
t
u
t
e
o
f
Sci
e
n
ce
an
d
T
ech
n
o
l
o
g
y
.
Si
n
ce
J
an
2
0
2
3
.
Sh
e
al
s
o
rece
i
v
e
d
h
er
B.
Sc.
,
M.
Sc.
,
an
d
M.
Ph
i
l
.
(Mat
h
emat
i
cs
)
fr
o
m
Mad
ra
s
U
n
i
v
er
s
i
t
y
,
Ch
e
n
n
a
i
i
n
2
0
0
9
,
2
0
1
2
a
n
d
2
0
1
3
re
s
p
ect
i
v
e
l
y
a
n
d
a
l
s
o
recei
v
e
d
h
er
B.
E
d
.
(Mat
h
ema
t
i
c
s
)
fro
m
T
am
i
l
n
ad
u
T
eac
h
er
E
d
u
ca
t
i
o
n
U
n
i
v
er
s
i
t
y
i
n
2
0
1
0
.
Sh
e
h
a
s
w
o
rk
e
d
as
an
A
s
s
i
s
t
an
t
P
r
o
fes
s
o
r
o
f
Mat
h
emat
i
cs
o
v
er
7
+
y
e
ars
.
H
er
r
es
earch
area
o
f
i
n
t
eres
t
i
n
cl
u
d
e
s
q
u
eu
e
i
n
g
t
h
e
o
ry
,
t
an
d
em
q
u
e
u
ei
n
g
mo
d
el
,
s
t
o
c
h
as
t
i
c
an
d
p
ro
b
ab
i
l
i
t
y
.
Sh
e
h
a
s
p
u
b
l
i
s
h
ed
o
v
er
6
p
ap
er
s
i
n
i
n
t
ern
a
t
i
o
n
a
l
j
o
u
r
n
al
s
,
n
at
i
o
n
a
l
j
o
u
rn
a
l
s
an
d
,
co
n
fere
n
ces
.
Sh
e
can
b
e
co
n
t
ac
t
ed
at
emai
l
:
k
eert
h
i
2
7
7
8
9
@
g
mai
l
.
co
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
Optimiz
ing
que
ue
e
ff
icie
nc
y
:
A
r
ti
fi
c
ial
int
e
ll
igenc
e
-
dr
ive
n
tandem
que
ue
s
…
(
K
e
e
r
thi
k
a
R
adhak
r
is
hn
an)
983
D
r.
Subra
m
a
n
i
P
a
l
a
n
i
N
i
r
a
nja
n
i
s
w
o
r
k
i
n
g
as
A
s
s
i
s
t
a
n
t
Pro
fe
s
s
o
r
o
f
Mat
h
ema
t
i
cs
i
n
V
e
l
T
ech
Ran
g
araj
a
n
D
r.
Sag
u
n
t
h
a
l
a
R&D
In
s
t
i
t
u
t
e
o
f
Sci
e
n
ce
an
d
T
ec
h
n
o
l
o
g
y
,
A
v
ad
i
,
Ch
en
n
ai
.
H
e
h
a
s
ei
g
h
t
y
ears
o
f
t
each
i
n
g
an
d
res
earc
h
ex
p
er
i
en
ce.
H
e
o
b
t
a
i
n
e
d
a
mas
t
e
r
d
eg
ree
i
n
s
ci
e
n
ce
at
PSG
Co
l
l
e
g
e
o
f
T
ec
h
n
o
l
o
g
y
,
Co
i
m
b
a
t
o
re.
H
e
aw
ard
e
d
D
o
c
t
o
r
o
f
Ph
i
l
o
s
o
p
h
y
i
n
Mat
h
ema
t
i
c
s
b
y
t
h
e
V
e
l
l
o
re
I
n
s
t
i
t
u
t
e
o
f
T
ech
n
o
l
o
g
y
,
V
e
l
l
o
re
i
n
t
h
e
y
ear
o
f
2
0
1
8
.
H
e
w
o
rk
e
d
as
a
j
u
n
i
o
r
res
earch
fel
l
o
w
i
n
V
IT
,
V
el
l
o
re,
s
p
o
n
s
o
re
d
b
y
N
BH
M
-
D
A
E
.
H
i
s
res
earch
areas
i
n
cl
u
d
e
q
u
e
u
ei
n
g
t
h
e
o
ry
a
n
d
o
p
era
t
i
o
n
s
res
earc
h
.
H
e
p
u
b
l
i
s
h
ed
2
0
re
s
earch
ar
t
i
c
l
es
,
o
n
e
p
at
e
n
t
,
an
d
t
w
o
b
o
o
k
ch
ap
t
ers
i
n
rep
u
t
ed
i
n
t
er
n
at
i
o
n
al
j
o
u
rn
a
l
s
.
H
e
can
b
e
co
n
t
act
e
d
at
emai
l
:
n
i
ra
n
.
j
ay
a
n
0
9
2
@
g
ma
i
l
.
c
o
m.
D
r.
Ba
l
a
kri
s
hna
n
Ko
m
a
l
a
D
urg
a
i
s
c
u
rren
t
l
y
w
o
rk
i
n
g
a
s
an
A
s
s
i
s
t
a
n
t
Pr
o
fes
s
o
r
i
n
t
h
e
D
ep
ar
t
me
n
t
o
f
Mat
h
ema
t
i
c
s
at
V
i
r
u
d
h
u
n
ag
ar
H
i
n
d
u
N
ad
ar
s
’
Sen
t
h
i
k
u
mara
N
a
d
ar
Co
l
l
eg
e
(A
u
t
o
n
o
m
o
u
s
)
affi
l
i
a
t
ed
w
i
t
h
Ma
d
u
ra
i
K
amara
j
U
n
i
v
ers
i
t
y
.
Sh
e
o
b
t
a
i
n
e
d
h
er
Ph
.
D
.
i
n
Mat
h
ema
t
i
c
s
u
n
d
er
t
h
e
g
u
i
d
a
n
ce
o
f
P
ro
fe
s
s
o
r
D
r.
E
.
Ch
an
d
ra
s
ek
ara
n
at
V
e
l
T
ech
Ra
n
g
ara
j
an
D
r.
Sag
u
n
t
h
al
a
R&D
In
s
t
i
t
u
t
e
o
f
Sci
en
ce
a
n
d
T
ec
h
n
o
l
o
g
y
,
l
o
cat
e
d
i
n
A
v
a
d
i
,
Ch
e
n
n
a
i
,
an
d
T
ami
l
N
ad
u
.
Sh
e
co
mp
l
et
ed
h
er
u
n
d
e
rg
ra
d
u
a
t
e
an
d
p
o
s
t
g
rad
u
at
e
fr
o
m
St
el
l
a
Mari
s
C
o
l
l
eg
e,
Ch
en
n
ai
,
affi
l
i
a
t
ed
w
i
t
h
t
h
e
U
n
i
v
er
s
i
t
y
o
f
Mad
ra
s
.
T
h
e
au
t
h
o
r
h
a
d
t
h
e
d
e
s
i
re
a
n
d
p
as
s
i
o
n
t
o
t
each
m
at
h
ema
t
i
c
s
,
w
h
i
c
h
co
n
s
eq
u
en
t
l
y
mad
e
h
er
s
t
art
h
er
t
each
i
n
g
career
i
n
2
0
1
2
.
Sh
e
h
as
t
eac
h
i
n
g
ex
p
er
i
en
ce
o
f
mo
re
t
h
an
1
0
y
ear
s
i
n
t
h
e
area
o
f
en
g
i
n
eeri
n
g
mat
h
emat
i
cs
.
T
h
e
au
t
h
o
r
h
as
pu
b
l
i
s
h
ed
n
u
mero
u
s
re
s
earch
p
ap
er
s
i
n
Sc
o
p
u
s
-
i
n
d
ex
e
d
j
o
u
r
n
al
s
.
Sh
e
can
b
e
co
n
t
act
ed
at
ema
i
l
:
k
o
ma
l
ad
u
rg
a.
b
@
v
h
n
s
n
c.
e
d
u
.
i
n
.
D
r.
Sures
ha
R
a
m
a
reddy
i
s
an
A
s
s
i
s
t
an
t
Pr
o
fes
s
o
r
i
n
t
h
e
D
e
p
art
me
n
t
o
f
Mat
h
ema
t
i
c
s
at
t
h
e
Facu
l
t
y
o
f
E
n
g
i
n
eer
i
n
g
an
d
T
ec
h
n
o
l
o
g
y
,
J
A
I
N
(
d
eemed
-
to
-
be
-
U
n
i
v
ers
i
t
y
),
Ban
g
a
l
o
re,
K
ar
n
at
ak
a,
In
d
i
a.
H
e
h
o
l
d
s
a
B.
Sc
.
an
d
a
B.
E
d
.
i
n
mat
h
emat
i
cs
fro
m
Ba
n
g
a
l
o
re
U
n
i
v
er
s
i
t
y
,
an
M.
Sc
.
in
mat
h
em
at
i
cs
fro
m
Ba
n
g
a
l
o
re
U
n
i
v
ers
i
t
y
,
an
d
a
Ph
.
D
.
i
n
fl
u
i
d
mec
h
an
i
cs
fro
m
V
i
s
v
e
s
v
ara
y
a
T
ec
h
n
o
l
o
g
i
cal
U
n
i
v
ers
i
t
y
,
Bel
g
a
u
m,
In
d
i
a.
H
e
i
s
an
act
i
v
e
memb
er
o
f
s
ev
era
l
p
ro
fe
s
s
i
o
n
al
o
rg
a
n
i
za
t
i
o
n
s
,
i
n
cl
u
d
i
n
g
t
h
e
Rama
n
u
j
an
Mat
h
ema
t
i
ca
l
So
c
i
et
y
(RMS),
t
h
e
I
n
d
i
an
Mat
h
ema
t
i
ca
l
So
ci
et
y
(IMS),
an
d
t
h
e
In
d
i
a
n
So
ci
e
t
y
o
f
T
h
eo
ret
i
cal
an
d
A
p
p
l
i
e
d
Mech
an
i
cs
(IST
A
M).
H
e
h
as
o
v
er
1
1
+
y
ears
o
f
t
eac
h
i
n
g
e
x
p
er
i
en
ce
,
h
as
p
u
b
l
i
s
h
e
d
n
u
mero
u
s
re
s
earc
h
p
a
p
ers
i
n
b
o
t
h
n
a
t
i
o
n
a
l
an
d
i
n
t
er
n
at
i
o
n
al
j
o
u
r
n
al
s
,
an
d
h
o
l
d
s
s
ev
eral
p
at
en
t
s
.
H
i
s
res
ea
rch
i
n
t
ere
s
t
s
are
fl
u
i
d
mec
h
an
i
cs
,
t
r
i
b
o
l
o
g
y
,
s
c
h
ed
u
l
i
n
g
p
r
o
b
l
ems
,
an
d
q
u
e
u
ei
n
g
t
h
e
o
ry
.
H
e
ca
n
b
e
c
o
n
t
ac
t
ed
a
t
emai
l
:
s
u
re
s
h
aramare
d
d
y
@
g
ma
i
l
.
co
m.
Evaluation Warning : The document was created with Spire.PDF for Python.