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.
3
,
June
2021
,
pp. 2
516~
2524
IS
S
N: 20
88
-
8708
,
D
O
I: 10
.11
591/
ijece
.
v11
i
3
.
pp2516
-
25
24
2516
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om
Solving
th
e order
batchi
ng and
sequ
encing proble
m with
mu
ltip
le pickers:
A
gr
ouped gene
tic algo
rith
m
Jo
s
e A
le
jandr
o Cano
1
,
P
ab
l
o Co
r
tés
2
,
Em
ir
o
A
ntoni
o C
ampo
3
, Ale
xa
nder
Albert
o Co
rre
a
-
Espi
n
al
4
1
Facul
t
y
of Econ
om
ic
s a
nd
Adm
ini
strat
iv
e
Sc
ie
n
c
es,
Unive
rsid
ad de
Mede
ll
ín
,
Co
l
om
bia
2
Escue
l
a
T
éc
n
ica
Superior
de
In
geni
er
ía
,
Sevi
ll
e
Univer
sit
y
,
Spai
n
3
ESACS
-
Escue
la
Superior
en
Adm
ini
strac
ión
de
Cade
na
de
Sum
ini
stro,
Co
lombia
4
Depa
rta
m
ent
o
d
e
Inge
n
ie
r
ía
d
e l
a
Organi
zaci
ón
,
Univer
sidad
Na
c
iona
l
de
Colom
bia
,
Colom
b
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
ug
21, 202
0
Re
vised
Oct
7
,
2020
Accepte
d
Oct
27
, 202
0
Thi
s
pape
r
int
ro
duce
s
a
groupe
d
gene
tic
a
lgori
th
m
(GG
A)
to
solve
the
ord
er
bat
ch
ing
and
se
quenc
ing
prob
lem
with
m
ult
ipl
e
pic
ker
s
(OBS
PM
P)
with
t
he
obje
c
ti
ve
of
m
in
imizi
ng
total
co
m
ple
ti
on
ti
m
e
.
T
o
the
best
of
our
knowledge,
for
the
first
ti
m
e,
an
OBS
PM
P
is
solved
b
y
m
ea
ns
of
GG
A
conside
rin
g
pic
king
d
evice
s
with
he
te
rog
en
eous
loa
d
ca
p
a
ci
t
y
.
For
thi
s
,
a
n
enc
od
ing
sche
m
e
is
proposed
to
rep
rese
nt
in
a
chr
om
osom
e
the
orde
rs
assigned
to
bat
ch
es,
and
ba
t
che
s
assigned
to
pic
k
ing
dev
ices.
L
ike
wise
,
th
e
oper
at
ors
o
f
the
proposed
a
l
gorit
hm
are
ad
apt
ed
to
the
sp
ec
if
ic
req
u
ire
m
ent
s
of
the
OBS
P
MP
.
Comput
at
ion
al
exper
iments
show
that
th
e
GG
A
per
f
orm
s
m
uch
bet
t
er
tha
n
six
orde
r
bat
ch
ing
an
d
seque
nci
ng
he
uristi
cs,
l
ea
ding
to
func
ti
on
obje
c
ti
ve
savin
gs
of
18.
3%
o
n
ave
rag
e
.
As
a
conc
lusion
,
t
he
proposed
al
gorit
hm
provi
des
fea
sible
s
olut
ions
for
the
oper
ations
pla
nning
in
ware
houses
an
d
distri
but
ion
ce
nt
ers,
improv
ing
m
arg
ins
b
y
r
educing
oper
ating
ti
m
e
f
or
orde
r
pi
cke
rs,
and
improving
c
ustom
er
servic
e
b
y
r
educing
pic
king
service t
imes.
Ke
yw
or
d
s
:
Groupe
d ge
netic
algorit
hm
s
Heter
og
e
ne
ou
s
load capa
ci
ty
Mult
iple picke
rs
Order batc
hing
Sequenci
ng
This
is an
open
acc
es
s arti
cl
e
un
der
the
CC
B
Y
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
Jo
se
Aleja
ndr
o C
an
o
Faculty
of Ec
onom
ic
s an
d A
dm
inist
rati
ve
Sc
ie
nces
Un
i
ver
si
dad d
e
Medell
in
Ca
rr
era
87 #
30
-
65, Me
delli
n,
Col
om
bia
Em
a
il
: jacano@udem
.ed
u.
c
o
1.
INTROD
U
CTION
The
order
pick
ing
pro
blem
is
in
cha
rg
e
of
re
trie
ving
the
set
of
it
em
s
fr
om
stora
ge
locat
io
ns
to
fu
l
fill
custom
er
orde
rs
an
d
delive
r
them
on
tim
e,
wh
il
e
pic
ke
rs
walk
or
op
erate
a
picking
de
vice
thr
ou
gh
t
he
war
e
h
ou
se
[
1,
2]
.
Ma
nual
ord
er
picking
syst
e
m
s
are
pre
valent
in
pract
ic
e
involvin
g
hum
an
op
e
rato
rs
at
la
rg
e
scal
e
du
e
to
th
e
flexibili
ty
and
auto
nom
y
offer
e
d
by
them
[3
]
and
the
l
ow
la
bor
costs
in
te
rr
it
or
ie
s
w
her
e
autom
at
ed
syst
e
m
s ar
e not
via
ble
[
4, 5]
.
Con
ce
rn
i
ng
to
achieve
pic
kin
g
ef
fici
ency,
the
orde
r
batchin
g
groups
c
us
tom
er
orde
rs
into
batche
s
with
a
m
axi
m
u
m
fixed
capaci
ty
[6]
,
then
the
batches
a
re
as
sign
e
d
to
a
picking
de
vice
an
d
batc
h
seq
ue
nc
in
g
determ
ines
the
picking
sc
he
duli
ng
an
d
the
c
om
ple
ti
on
tim
e
batches
a
nd
cu
stom
er
or
de
rs
[
3,
7
]
.
The
refo
r
e,
th
e
j
oi
nt
orde
r
bat
chin
g
a
nd
seq
uen
ci
ng
pro
bl
e
m
with
m
ultip
le
pic
ker
s
(
O
BSPM
P)
is
pi
vo
ta
l
to
en
ha
nc
e
the
eff
ic
ie
ncy
a
nd
custom
er
ser
vi
ce
[8
-
10]
.
One
of
the
m
os
t
i
m
po
rtant
obj
e
ct
ives
in
order
picki
ng
syst
em
s
is
m
ini
m
iz
ing
the
m
axi
m
al
co
m
pleti
on
ti
m
e
(m
akesp
a
n)
,
w
hi
ch
al
lows
re
duci
ng
workin
g
ti
m
e
fo
r
orde
r
pickers
,
i
m
pr
ovin
g
pro
fit
m
arg
ins
f
or
war
e
house
operati
on
s
,
reduc
ing
delive
ry
le
ad
ti
m
es
and
im
pr
ov
in
g
c
us
t
om
er
serv
ic
es
[
11
]
.
Ther
e
are
only
a
fe
w
st
ud
ie
s
con
sideri
ng
m
akes
pan
as
th
e
ir
ob
j
ect
ive
to
m
ini
m
iz
e
the
s
erv
ic
e
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
So
lv
in
g
t
he or
de
r batchi
ng an
d
se
quenci
ng prob
le
m
wi
th…
(
Jose Alej
andro C
ano
)
2517
tim
e
fo
r
al
l
po
ssible
batc
hes
[12]
,
th
us
,
m
i
nim
iz
e
the
m
akes
pan
s
uppo
r
ts
oth
e
r
ob
j
ect
ives
li
ke
m
ini
m
iz
ing
total
tard
ine
ss
[13, 1
4]
.
The
orde
r
batc
hing
pro
blem
i
s
con
si
der
e
d
N
P
-
Ha
r
d
when
the
num
b
er
of
custom
er
or
de
r
s
per
batc
h
is
gr
eat
er
tha
n
two
[
15]
,
w
hi
ch
m
eans
it
i
s
i
m
po
ssible
to
obta
in
a
po
l
ynom
ia
l
-
tim
e
so
luti
on
f
or
it
[16]
,
therefo
re,
t
his
ty
pe
of
pr
ob
l
e
m
req
ui
res
to
be
s
olv
e
d
us
i
ng
ap
pro
xim
ate
m
e
tho
ds
su
c
h
as
m
et
aheu
r
ist
ic
s
[17,
18]
,
am
ong
wh
ic
h
the
pa
rtic
le
sw
arm
optim
iz
at
ion
[19]
,
ant
c
ol
on
y
op
ti
m
iz
ation
[
20
]
,
ge
netic
al
gorithm
s
(GA)
[
21
]
,
am
ong
oth
e
rs,
c
a
n
be
m
entione
d.
Sp
eci
fical
ly
,
gro
up
-
or
ie
nte
d
gen
et
ic
al
gor
it
h
m
s
(GGA)
s
upport
the
su
cce
ssf
ul
ap
plica
ti
on
t
o
groupin
g
prob
le
m
s
because
crit
ic
al
inf
or
m
at
ion
fr
om
the
ch
r
om
os
om
e
i
s
pr
ese
r
ved
a
nd
is
cor
rectl
y
transf
e
rr
e
d
in
the
cro
ss
over
op
e
rators
[
22
]
.
T
hus,
m
aking
us
e
of
a
gro
up
-
or
i
ente
d
encodin
g sche
m
e is sensiti
ve t
o
the
gro
up f
e
at
ur
es
of the
O
BSPM
P whe
re
a g
e
ne rep
rese
nts a
batch
i
ns
t
ead
of
a
custom
er
order.
Mo
reover
,
the
stu
dies
of
[22
-
24]
ha
ve
i
m
ple
m
ented
GGAs
to
so
l
ve
the
orde
r
ba
tc
hin
g
pro
blem
in
war
eh
ouses
an
d
orde
r
pic
king
syst
e
m
s,
fo
r
wh
i
ch
they
are
us
e
d
as
a
ref
e
re
nc
e
to
j
oi
ntly
so
lve
the
OBSPMP
i
n
t
his
stu
dy,
wh
i
ch,
unli
ke
the
m
od
el
s
pro
posed
in
the
l
it
eratur
e
,
incl
ud
e
s
the
assi
gn
m
ent
of
batches
to
m
ulti
ple p
ic
ki
ng
de
vices
with
heteroge
neous
l
oadi
ng
ca
pacit
y.
To
s
olv
e
the
O
BSPM
P,
He
nn
[8]
,
Sc
ho
lz
et
al.
[
25
]
,
a
nd
V
an
Gils
et
al.
[
10
]
ha
ve
pro
posed
va
riable
neig
hborh
ood
desce
nt
an
d
va
riable
nei
ghbor
hood
sea
rch
appr
oach
es
to
m
ini
m
iz
e
ta
rdi
ness
an
d
t
otal
orde
r
pick
ti
m
e. Matu
sia
k
et
al.
[
26]
propose
d
a
n
a
dap
ti
ve
lar
ge n
ei
ghbor
hood se
arch
al
gorithm
co
nsi
der
i
ng p
ic
ker
s
with
div
e
rse
s
kill
s
to
m
ini
m
iz
e
the
total
orde
r
processi
ng
ti
m
e.
Zhang
et
al.
[
9]
pr
opose
d
a
ru
le
-
ba
sed
appr
oach
to
m
ini
m
iz
e
the
m
axi
m
u
m
com
ple
ti
on
ti
m
e
of
al
l
batc
he
s
f
or
onli
ne
e
nv
i
ronm
ents.
Howe
ver,
popula
ti
on
-
bas
ed
m
et
aheu
rist
ic
s
li
ke
ge
neti
c
al
gorithm
s
ha
ve
no
t
bee
n
f
ound
in
the
li
te
ratur
e
to
s
olve
the
OBSPMP
,
as
well
as
no
OB
SPMP
m
od
el
s
c
on
si
der
i
ng
pi
cking
de
vices
with
heter
oge
neous
loa
d
ca
pacit
y,
wh
ic
h
is a
ch
a
r
act
erist
ic
o
f
m
od
e
r
n ware
house
s and
distrib
ution cente
rs.
Ther
e
f
or
e,
this
pa
per
ai
m
s
to
prese
nt
f
or
th
e
first
ti
m
e
the
ap
plica
ti
on
of
a
GGA
f
or
th
e
OBS
PMP,
consi
der
i
ng
pi
ckin
g
dev
ic
es
with
a
heter
ogeneous
l
oad
ca
pacit
y
to
m
ini
m
iz
e
the
m
axim
u
m
co
m
pleti
on
tim
e
(m
akesp
an
).
T
he
rem
ai
nd
er
of
t
his
pa
pe
r
i
s
orga
nized
as
fo
ll
ows
.
Sect
ion
2
int
rod
uc
es
the
f
eat
ur
e
s
an
d
assum
ption
s
f
or
the
OBS
PMP
.
In
s
ect
io
n
3,
a
GGA
to
s
olve
the
OBSPM
P
is
pr
ese
nted
.
Sect
ion
4
i
ntr
oduces
the
ex
per
im
ents
to
determ
ine
the
pe
rfor
m
ance
of
t
he
pro
pos
ed
m
od
el
.
Sect
ion
5
com
par
es
the
pe
rfor
m
ance
of
the GG
A wit
h si
x ben
c
hm
ark
s,
s
howing sa
vi
ng
s
for t
he
m
a
kes
pan
. Concl
us
io
ns
a
re
disc
us
se
d
in
s
ect
io
n 6.
2.
OBSPM
P FE
ATU
RES
A
N
D ASS
U
MPTI
ONS
The
m
ai
n
featur
es
an
d
assum
ptions
of
th
e
orde
r
picki
ng
sy
stem
f
or
the
O
BSPM
P
are
th
e
fo
ll
owin
g;
In
i
)
T
he
or
der
picking
pro
blem
is
based
on
lo
w
-
le
vel
p
ic
ke
r
-
to
-
par
ts
syst
em
s,
ii
)
the
war
e
hous
e
config
ur
at
io
n
i
s
base
d
on
a
pa
rall
el
-
ai
sle
sin
gle
-
bl
ock
wa
reho
us
e,
ii
i
)
e
ach
cu
stom
er
order
is
c
om
po
se
d
of
sever
al
it
em
s,
iv
)
m
ulti
ple
pi
ckin
g
de
vices
with
heter
oge
ne
ous
load
ca
pa
ci
ty
and
co
ns
ta
nt
horizo
ntal
sp
ee
d
are
al
lowe
d,
a
nd
batc
h
siz
e
do
e
s
not
exce
e
d
th
e
ca
pacit
y
of
picki
ng
de
vi
ces
(v
)
eac
h
ba
t
ch
is
assig
ne
d
to
a
picking
de
vice
and
it
fo
ll
ow
s
the
S
-
s
hap
e
routin
g
he
ur
ist
ic
to
retriev
e
al
l
the
it
e
m
s
of
the
batch,
vi
)
the
com
pleti
on
time
of
a
batch
is
equ
al
to
the
co
m
ple
ti
on
tim
e
of
the
orde
rs
a
ssign
e
d
to
it
,
vii
)
the
serv
ic
e
tim
e
of
a
batc
h
is
e
qu
a
l
to
the
tra
vel
t
i
m
e,
m
easur
ed
as
the
tra
vele
d
distance
di
vid
ed
the
sp
e
ed
of
the
pic
king
dev
ic
e
and
viii
)
a
picking
de
vice
can
ha
nd
le
the
ne
xt
batch
on
ly
wh
e
n
a
previ
ous
batc
h
is
finish
e
d.
The
warehous
e
config
ur
at
io
n
i
s
based
on
a
par
al
le
l
-
ai
sle
sing
le
-
bl
oc
k
wa
reho
us
e
as
des
cribe
d
in
Fig
ure
1,
wh
e
re
st
or
a
ge
locat
ion
s
,
ai
sle
s
widt
h
a
nd
cro
ss
-
ai
sle
s
dim
ension
s
are
il
lustrate
d,
as
well
as
an
ex
a
m
ple
of
t
he
s
-
sh
a
pe
routin
g
strat
e
gy
u
se
d
to
so
l
ve
the
picke
r
r
out
ing
pro
blem
.
Figure
1
.
W
a
re
hous
e
d
im
ension
s
a
nd S
-
s
ha
pe
rou
ti
ng st
rategy
1u
1u
1u
D
e
p
o
t
1u
1u
1u
1u
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
.
3
,
J
une
2021 :
25
16
-
2524
2518
Give
n
a
set
of
batches
∈
,
a
set
of
c
us
tom
er
orders
∈
,
a
set
of
st
or
a
ge
l
oca
ti
on
s
,
∈
,
a
su
bse
t
of
sto
ra
ge
locat
io
ns
⊂
,
a
set
of
posit
ion
s
to
sc
he
du
l
e
a
batc
h
in
a
picki
ng
de
vic
e,
an
d
a
set
of
picking
de
vice
s
∈
, t
he
m
at
he
m
at
ic
al
f
or
m
ulatio
n o
f
t
he OBS
PMP is
descr
i
be
d
as
foll
ows:
Parameters
=
Ca
pacit
y req
ui
red f
or ord
e
r
o
=
Trav
el
ti
m
e fr
om
p
os
it
ion
i
to
j
for pic
king
de
vice
e
=
Ma
xim
u
m
capacit
y of
picking
dev
ic
e
e
=
{
1
if
an
item
of
t
he
o
rd
er
s
re
tr
ie
ve
d
from
p
osi
t
ion
0
ot
he
rwi
se
Decisi
on v
ar
ia
bles
=
{
1
if
orde
r
is
assi
gn
ed
to
batch
0
ot
he
rwi
se
=
{
1
if
bat
c
h
is
a
ss
igne
d
to
th
e
picki
ng
devi
ce
0
othe
rwi
se
=
{
1
if
batch
vi
sits
ju
st
af
t
er
vi
sit
in
g
0
othe
rwi
se
=
{
1
if
ba
t
c
h
vi
sits
to
re
t
ri
ev
e
an
item
0
ot
he
rwi
se
=
{
1
if
batch
is
sch
edul
ed
in
p
osi
t
ion
0
othe
rwi
se
=
{
1
if
ba
t
c
h
is
ass
igne
d
to
po
sitio
n
in
pick
in
g
d
ev
ice
0
othe
rwi
se
Com
pleti
on
ti
m
e fo
r
a
batch
sche
du
le
d
in
posit
ion
k
in
p
ic
king
dev
ic
e
e
Com
pleti
on
ti
m
e fo
r
or
der
o
∈
{
}
(1)
∑
∈
≤
1
∀
∈
,
∈
(2)
∑
∑
∑
∈
∗
∈
∈
=
1
∀
∈
(3)
∑
∗
∗
∈
≤
∀
∈
,
∈
,
∈
(4)
≥
∗
∀
∈
,
∈
,
∈
(5)
∑
∈
,
≠
=
∀
∈
,
∈
(6)
∑
∈
,
≠
=
∀
∈
,
∈
(7)
∑
∈
,
∈
∖
≥
∀
∈
,
⊂
(8)
∑
(
1
∗
∑
∑
∗
≠
∈
≠
∈
)
∈
≤
1
∀
∈
(9)
−
1
+
∑
(
∗
∑
∑
∗
≠
∈
≠
∈
)
∈
≤
∀
∈
,
∈
∖
{
1
}
(10)
=
∑
∑
∑
∗
∗
∈
∈
∈
∀
∈
(11)
,
≥
0
∀
∈
,
∈
,
∈
(12)
,
,
,
∈
{
0
,
1
}
∀
∈
,
∈
,
∈
,
∈
,
,
∈
(13)
The
obj
ect
ive
functi
on
in
(1)
m
ini
m
iz
es
the
m
axi
m
u
m
com
ple
ti
on
tim
e
(m
akesp
an
).
C
on
st
raints
i
n
(2)
ens
ures
th
at
at
m
os
t
a
b
at
ch
is
schedu
le
d
in
a
sp
eci
f
ic
po
sit
ion
i
n
a
picking
de
vi
ce.
Const
raints
in
(3)
gu
a
ra
ntees the
assignm
ent o
f e
ach cust
om
er o
r
der
t
o
a b
at
c
h.
In
(4)
li
m
it
s
the capacit
y of t
he
batche
s as
sign
e
d
to
each
pic
king
dev
ic
e
acc
ordin
g
to
t
he
loa
ding
ca
pacit
y
o
f
t
he
picking
dev
ic
es
.
Co
ns
t
raints
in
(
5
-
8)
e
m
bo
dy
the
TS
P
form
ulati
on
.
In
(
9)
a
nd
(
10)
m
easur
es
the
com
pletio
n
ti
m
e
fo
r
t
he
batc
h
sc
he
du
l
ed
i
n
po
sit
io
ns
1
a
nd
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
So
lv
in
g
t
he or
de
r batchi
ng an
d
se
quenci
ng prob
le
m
wi
th…
(
Jose Alej
andro C
ano
)
2519
k
in
each
picki
ng
dev
ic
e
res
pe
ct
ively
.
In
(11)
cal
culat
es
the
com
pleti
on
tim
e
of
an
ord
er
as
the
com
p
le
ti
on
tim
e
of
the
batch
wh
e
re
it
is
assigne
d.
Co
nst
raints
in
(12
)
and
(
13)
deter
m
ine
the
no
n
-
ne
gativit
y
and
dom
ai
n
of
the
va
riable
s.
The
ord
er
ba
tc
hin
g
pro
blem
is
con
side
red
NP
-
Ha
rd
wh
e
n
the
nu
m
ber
of
custom
er
or
de
rs
pe
r
batch
is
gr
eat
er
than
tw
o
[
15
]
,
the
refor
e
any
exten
sio
n
of
this
prob
l
e
m
is
con
side
red
N
P
-
Hard
as
well
[13,
27,
28]
.
The
n,
t
he
OB
SPMP
is
co
nsi
der
e
d
a
n
NP
-
hard
pro
blem
,
so
it
can
not
be
s
olv
e
d
us
i
ng
e
xact
so
luti
on
m
et
ho
ds
at
le
ast
fo
r
la
rg
e
instance
s
[25]
,
thus
m
et
aheurist
ic
s
lik
e
ge
netic
al
go
rithm
s
can
prov
i
de
sat
isfact
or
y
so
l
utions
in
short
com
pu
ti
ng
ti
m
es
for
com
bin
at
or
ia
l
prob
le
m
s
relat
ed
to
log
i
sti
cs
and
opera
ti
on
s
m
anag
em
ent
[2
9, 30]
. In
this
stud
y,
we
intr
oduce a G
GA
t
o pro
vid
e
high
-
qu
al
it
y solutio
ns
in
s
hort c
om
pu
ti
ng
tim
es f
or
t
he O
BSPM
P,
sat
isf
yi
ng
the
ope
rati
ng
re
qu
irem
en
ts of real
war
e
hous
es
.
3.
GROUPE
D
G
ENET
IC AL
GORIT
HM
S
The
pr
opos
e
d
gro
up
-
ori
ente
d
encodin
g
sc
he
m
e
rep
rese
nts
the
assig
nm
ent
of
order
s
t
o
ba
tc
hes
an
d
the
sequ
e
nci
ng
of
batche
s
in
picking
de
vice
s.
Du
e
to
eac
h
gen
e
re
pr
ese
nt
s
a
batch
in
a
picking
de
vice;
the
chrom
os
om
es
are
of
va
riable
le
ng
th
.
I
n
the
encodin
g
sc
he
m
e
sh
own
in
Figure
2,
each
gen
e
is
c
om
po
se
d
of
the
num
ber
of
the
pic
king
de
vice,
t
he
batch
assig
ned
to
th
e
pic
king
dev
i
ce,
a
nd
the
c
ust
om
er
order
s
gro
upe
d
in the bat
c
h.
Figu
re
2. G
GA enc
od
i
ng sc
he
m
e
To
pro
du
ce
t
he
init
ia
l
popu
l
at
ion
of
siz
e
P
,
we
f
ollow
a
n
orde
r
gr
oup
proce
dure
t
hat
us
es
a
n
order
pool
to
place
orde
rs
that
ha
ve
no
t
ye
t
bee
n
a
ssign
e
d
to
a
ba
tc
h.
F
or
eac
h
ge
ne,
a
picking
dev
ic
e
is
ra
nd
om
l
y
chosen
,
an
d
a
batch
is
ope
ne
d
f
or
t
his
picki
ng
de
vice,
the
n,
a
n
orde
r
is
r
andom
ly
cho
sen
f
ro
m
the
ord
er
pool
and
is
assig
ne
d
to
the
op
e
n
batch.
Anothe
r
or
de
r
is
random
ly
cho
se
n
fro
m
the
or
der
pool
and
is
assig
ned
to
the
ope
ned
bat
ch
if
t
he
capac
it
y
req
uirem
ent
of
th
e
ord
er
i
s
le
ss
than
or
e
qu
al
t
o
the
a
va
il
able
capaci
ty
of
t
he
batch,
oth
er
wi
se,
ano
t
her
or
de
r
is
ran
dom
ly
cho
se
n
from
t
he
orde
r
pool.
If
none
of
the
or
de
rs
in
the
order
pool
fit
on
the o
pe
n
batc
h,
the
n
the
batch
is
cl
os
ed
.
The
ord
er
gro
up
pr
oce
dure
en
ds
w
he
n
al
l
cu
stom
er
orders
are
gro
uped
in
to
batches
.
T
he
fitness
f
un
ct
i
on
represe
nts
the
ob
j
ect
ive
f
un
ct
io
n
of
the
OBSPMP,
w
hich
is
m
ini
m
iz
ing
tot
al
com
pleti
on
tim
e,
so
the
m
a
xim
u
m
co
m
ple
ti
on
ti
m
e
or
total
com
pletio
n
tim
e
(
m
akesp
an)
is
identic
al
to
the
tim
e
req
uired
to
colle
ct
a
give
n
set
of
cust
om
er
order
s
.
Gi
ven
t
he
set
of
orders,
o
=
1,
.
..
,
O
,
le
t
C
o
be
the
com
ple
ti
on
ti
m
e
of the
orde
r
o
, so t
he
fitne
ss
f
un
ct
io
n
t
o
m
inim
iz
e w
it
h
the
GGA
is
Max
o
ϵ
O
{
C
o
}.
The
sel
ect
ion
of
pa
re
ntal
chrom
os
om
es
fo
r
the
cro
ss
ov
e
r
op
erat
o
r
is
ba
sed
on
the
li
near
sel
ect
ion
rankin
g
m
et
ho
d;
t
his
m
et
h
od
assig
ns
t
he
highest
sel
ect
ing
probab
il
ity
to
chro
m
os
o
m
es
with
bette
r
perform
ance,
prom
oting
the
cr
os
sin
g
bet
ween
pa
re
nts
with
high
-
qual
it
y
gen
et
ic
in
f
or
m
at
ion
.
T
he
n,
t
he
nu
m
ber
of p
ai
r
s of p
a
re
nts is
determ
ined
acc
ordin
g
to
the c
r
os
s
ov
e
r rat
e (
Cr
)
a
nd
pa
ren
ts
are chose
n usi
ng the
roulet
te
wh
eel
sel
ect
ion
.
Fig
ur
e
3
il
lustrate
s
the
crosso
ve
r
operat
or
that
beg
i
ns
with
t
he
sel
ect
ion
of
tw
o
cro
ssi
ng
poi
nts
delim
it
ing
the
crossi
ng
sect
i
on
(
s
te
p
1)
.
T
h
e
exc
hange
of
the
cr
os
si
ng
se
ct
ion
betwee
n
each
pair
of
par
e
nts
m
ay
le
ad
to
i
nf
easi
ble
so
lut
ion
s
w
he
n
an
order
a
pp
ea
rs
twic
e
on
a
ch
r
om
os
om
e
(
st
ep
2).
A
correct
ion
m
echan
ism
is
app
li
ed
to
fi
x
infea
sible
offs
pr
i
ng,
rem
ov
ing
old
gen
e
s
co
ntaini
ng
orde
rs
ap
pe
a
rin
g
on
th
e
ne
w
ge
nes,
a
nd
th
en
updates
the
se
qu
e
ncin
g
of
ba
tc
hes
in
each
picking
dev
ic
e
(step
3).
Orde
rs
that
hav
e
not
ye
t
been
assig
ne
d
to
the
ch
r
om
os
om
e
beco
m
e
p
art
of
the
orde
r
pool,
an
d
the
order
gro
up
procedu
re
is ap
plied to
com
ple
te
each
c
hrom
os
om
e
(
ste
p 4).
The
survi
val
m
echan
ism
ensu
res
that
el
it
e
in
div
id
uals
prevail
in
each
gen
e
rati
on
acc
ordin
g
t
o
t
he
su
r
viv
al
rate
(
Sr
)
.
T
he
im
m
i
gr
at
io
n
rate
(
Ir
)
def
i
nes
th
e
num
ber
of
ne
w
ind
ivi
du
al
s
to
create
usi
ng
th
e
orde
r
gro
up
procedu
re
to
pro
vid
e
div
e
rsity
to
the
popu
la
ti
on
a
nd
pr
e
ve
nt
pr
e
m
at
ur
e
conve
r
gen
ce
.
The
m
utati
on
op
e
rato
r
is
im
plem
ented
in
a
nu
m
ber
of
i
nd
i
viduals
de
f
ined
by
the
m
uta
ti
on
rate
(
Mr
)
.
T
he
m
utati
on
proce
dure
sta
rt
s
sel
ect
ing
two
gen
es
rand
oml
y,
then
the
sel
ect
ed
ge
nes
ar
e
rem
ov
ed
an
d
the
order
s
of
t
hese
gen
e
s
bec
om
e
avail
able
int
o
the
order
po
ol.
The
se
quence
of
batch
es
in
e
ach
picking
de
vice
is
up
dated
a
nd
the
rem
ai
nin
g
orders
are
assigne
d
to
ne
w
ge
nes
us
i
ng
the
order
group
pr
ocedu
re.
Last
ly
,
wh
e
n
G
ge
ne
rati
ons
G
e
n
e
1
G
e
n
e
2
G
e
n
e
3
G
e
n
e
4
G
e
n
e
5
P
i
c
k
i
n
g
d
e
v
i
c
e
1
3
3
2
2
Ba
t
c
h
n
u
m
b
e
r
1
1
2
1
2
4
9
2
1
8
5
3
10
7
6
O
rd
e
rs
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
.
3
,
J
une
2021 :
25
16
-
2524
2520
are
sat
isfie
d
t
he
gen
et
ic
al
gor
it
h
m
will
stop
.
Figure
4
sho
w
s
the
flo
wc
har
t
of
t
he
G
G
A
s
umm
arizi
ng
th
e
ste
ps
of the e
voluti
onary
process
.
Figure
3. Cr
os
s
ov
e
r o
per
at
or
f
or the
G
GA
Figure
4. The
fl
ow
c
har
t
of the
prop
os
ed
GG
A
P
i
c
ki
ng
de
vi
c
e
1
3
3
2
2
2
1
1
3
1
27
Ba
t
c
h n
um
be
r
1
1
2
1
2
1
1
2
1
2
38
4
9
2
1
8
2
9
1
7
3
1
5
3
10
10
3
5
6
4
12
7
6
4
8
5
23
6
5
P
i
c
ki
ng
de
vi
c
e
1
1
3
3
2
2
2
1
3
7
12
Ba
t
c
h n
um
be
r
1
2
1
2
1
2
1
1
1
8
22
4
1
7
2
1
8
2
9
9
9
36
5
5
6
10
10
3
3
10
7
7
8
6
4
Pi
c
k
i
n
g
d
e
v
i
c
e
C
a
p
a
c
i
t
y
P
i
c
ki
ng
de
vi
c
e
1
3
3
2
3
1
50
Ba
t
c
h n
um
be
r
1
1
2
1
1
2
45
1
7
2
2
9
3
40
5
6
10
3
8
P
i
c
ki
ng
de
vi
c
e
1
3
3
2
1
2
3
2
3
1
Ba
t
c
h n
um
be
r
1
1
2
1
2
1
1
2
2
1
1
7
2
9
4
2
9
5
7
1
5
6
10
3
10
3
8
4
8
6
O
rde
rs
O
f
f
s
p
r
i
n
g
1
S
te
p
4
O
f
f
s
p
r
i
n
g
2
O
r
d
e
r
s
O
r
d
e
r
s
i
z
e
S
te
p
1
S
te
p
2
S
te
p
3
O
rde
rs
O
rde
rs
Ch
r
o
m
o
m
e
1
Ch
r
o
m
o
m
e
2
O
rde
rs
S
tart
S
e
t
p
o
p
u
latio
n
siz
e
(
P
),
n
u
m
b
e
r
o
f
g
e
n
e
ra
ti
o
n
s (
G
),
c
ro
ss
o
v
e
r
ra
te (
Cr
),
m
u
tatio
n
ra
te (
Mr
),
su
rv
iv
a
l
ra
te
(
Sr
),
im
m
ig
ra
ti
o
n
ra
te
(
Ir
)
Cra
te
th
e
c
h
ro
m
o
so
m
e
p
in
th
e
g
e
n
e
ra
ti
o
n
g
f
o
ll
o
w
in
g
th
e
o
rd
e
r
g
ro
u
p
p
ro
c
e
d
u
re
S
e
lec
t
P
x
Sr
e
li
te ch
ro
m
o
so
m
e
s
f
ro
m
g
e
n
e
ra
ti
o
n
g
f
o
r
th
e
n
e
x
t
g
e
n
e
ra
ti
o
n
(
g
+
1
)
¿
g
=
G
?
En
d
Co
m
p
u
te
th
e
f
it
n
e
ss
f
u
n
c
ti
o
n
f
o
r
e
a
c
h
c
h
ro
m
o
so
m
e
p
No
No
Ye
s
¿
p
=
P
?
p
=
p
+
1
Ye
s
g
=
1
;
p
=
1
Us
e
th
e
li
n
e
a
r
ra
n
k
in
g
se
lec
ti
o
n
m
e
th
o
d
to
ra
n
d
o
m
ly
se
lec
t
P
x
2
Cr
c
h
ro
m
o
so
m
e
s
f
ro
m
g
e
n
e
ra
ti
o
n
g
to
b
e
p
a
irs
o
f
p
a
re
n
ts
G
e
n
e
ra
te
P
x
Cr
o
f
f
sp
rin
g
b
y
u
sin
g
th
e
c
ro
ss
o
v
e
r
o
p
e
ra
to
r
a
n
d
a
d
d
th
e
m
to
g
e
n
e
ra
ti
o
n
(
g
+1
)
Cre
a
te
P
x
Ir
n
e
w
c
h
ro
m
o
so
m
e
s
b
y
a
p
p
ly
in
g
th
e
m
ig
ra
ti
o
n
o
p
e
ra
to
r
Ra
n
d
o
m
ly
se
lec
t
P
x
Mr
c
h
ro
m
o
so
m
e
s
f
ro
m
g
e
n
e
ra
ti
o
n
(g
+
1
),
a
n
d
a
p
p
ly
th
e
m
u
tatio
n
o
p
e
ra
to
r
g
=
g
+
1
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
So
lv
in
g
t
he or
de
r batchi
ng an
d
se
quenci
ng prob
le
m
wi
th…
(
Jose Alej
andro C
ano
)
2521
4.
E
X
PERI
MEN
TS
In
orde
r
t
o
determ
ine
the
pe
r
form
ance
of
th
e
G
GA,
se
ver
a
l
exp
e
rim
ents
are
car
ried
out.
The
res
ults
of
t
he
G
G
A
a
re
com
par
e
d
with
six
be
nc
hm
ark
ru
le
s
c
al
le
d
FCFS
-
L
H,
FCF
S
-
HL,
SLO
S
-
L
H,
SL
OS
-
HL
,
LSO
S
-
L
H,
a
nd
LSOS
-
HL.
Th
e
ab
br
e
viati
on
LH
m
eans
that
batche
s
are
fir
st
assigne
d
t
o
t
he
picking
de
vi
ce
1,
then
to
t
he
pic
king
de
vice
2
a
nd
s
o
on
un
ti
l
t
he
la
st
picki
ng
dev
ic
e,
afte
r
w
hich,
the
ne
xt
batches
are
ass
ign
e
d
to
the
picki
ng
dev
ic
e
2
an
d
s
o
on
unti
l
assigni
ng
al
l
batch
es
to
picki
ng
de
vices.
T
he
ab
br
e
viati
on
HL
m
eans
that
batc
hes
a
r
e
assig
ne
d
firs
t
to
the
la
st
picking
de
vice,
t
hen
to
the
pe
nult
i
m
at
e
picking
de
vice
an
d
s
o
on
un
ti
l
the
first
pi
cking
de
vice,
after
w
hich
,
th
e
nex
t
batc
hes
are
assig
ned
t
o
the
penultim
ate
picking
de
vic
e
and
so
on
un
ti
l
assigni
ng
al
l
batc
he
s
to
pic
king
de
vices.
Mo
re
over
,
FC
FS
m
eans
that
the
assi
gn
m
ent
of
ord
ers
to
batches
is
base
d
on
t
he
fi
rst
-
c
om
e
-
first
-
ser
ve
d
r
ule.
S
LO
S
m
eans
that
the
assignm
ent
of
orde
rs
to
batc
hes
is
perform
ed
fro
m
the
sm
a
ll
e
st
-
siz
ed
order
to
the
la
rg
e
st
-
siz
ed
order.
Likewise
,
L
SO
S
m
eans
t
hat
the
assignm
ent o
f order
s to
batc
he
s is p
er
form
e
d
f
ro
m
the largest
-
siz
ed or
der
to the s
m
al
le
s
t
-
siz
ed
orde
r.
Fi
gure
5
il
lustrate
s the soluti
ons
pro
vide
d by the
pro
pose
d ben
c
hm
ark
heurist
ic
s for t
he OBSPM
P
.
Figure
5. Be
nc
hm
ark
h
e
ur
ist
i
cs for
the
OBS
PMP
The
e
xp
e
ri
m
ents
are
co
nf
i
gur
ed
wit
h
the
pa
ram
et
ers
descri
bed
i
n
Ta
ble
1.
By
com
bin
i
ng
dif
fer
e
nt
values
f
or
O
,
m,
a
nd
w
,
24
pro
blem
cl
asses
are
gen
e
rated
,
an
d
10
i
ns
ta
nc
es
are
cal
c
ulate
d
f
or
each
pr
ob
le
m
.
Ther
e
f
or
e,
we
pr
ovi
de
240
instances
in
total
.
The
pa
ra
m
et
ers
o
f
the
GGA
are
P
=
20,
G
=
40,
Cr
=0.85,
Mr
=0.0
5,
Sr
=0
.1
,
a
nd
Ir
=0
.05
.
The
ex
pe
rim
e
nts
are
car
ried
ou
t
on
a
n
In
te
l C
or
e
i5
-
23
00
CPU
at
2.8
G
H
z
and
8 GB RAM.
T
he
al
go
rithm
is
i
m
ple
m
ented wit
h Visual Ba
sic
.
Table
1.
Orde
r pickin
g si
m
ula
ti
on
e
nv
i
ronm
e
nts
Para
m
ete
rs
Valu
es
Total n
u
m
b
er
o
f
c
u
sto
m
e
r
o
rders
(
O
)
1
0
,
3
0
,
5
0
Nu
m
b
e
r
o
f
ite
m
s
per order
(
m
)
U[
1
,
5
],
U[
5
,
1
5
]
W
ar
eh
o
u
se lay
o
u
t
(sto
rage locatio
n
s)
(
w
)
4
0
0
,
9
0
0
,
1
2
5
0
,
2
0
0
0
Ro
u
tin
g
strategy
S
-
sh
ap
e
Sto
rage po
licy
Ran
d
o
m
-
b
ased
Cap
acity
of
the p
ic
k
in
g
dev
ice
1
,
2
,
3
3
0
,
4
0
,
5
0
Sp
eed o
f
the p
ick
i
n
g
dev
ices
4
5
m
/
m
in
5.
RESU
LT
S
AND DI
SCUS
S
ION
The
res
ults
of
the
G
GA
are
com
par
ed
wit
h
th
e
six
ben
c
hm
ark
s,
cal
c
ulati
ng
sa
vings
f
or
the
t
otal
com
pleti
on
tim
e.
As
s
how
n
in
Table
2,
GGA
saves
on
a
ver
a
ge
bet
ween
14.3%
an
d
23.5%
of
total
com
pleti
on
tim
e
wh
en
co
m
par
ed
to
t
he
ben
c
hm
ark
he
ur
ist
ic
s.
The
pro
posed
G
G
A
ind
ee
d
offers
s
avin
gs
of
up
to
40.
7%
P
i
c
ki
ng
de
vi
c
e
1
2
3
1
2
3
2
1
3
2
1
27
Ba
t
c
h n
um
be
r
1
1
1
2
2
1
1
1
2
2
2
38
1
2
5
8
9
1
2
5
8
9
3
1
3
10
7
3
6
7
4
12
4
4
10
5
23
6
6
5
7
12
P
i
c
ki
ng
de
vi
c
e
1
2
3
1
2
3
2
1
3
2
8
22
Ba
t
c
h n
um
be
r
1
1
1
2
2
1
1
1
2
2
9
36
3
8
1
9
2
3
8
1
9
2
10
7
6
5
6
5
10
10
Pi
c
k
i
n
g
d
e
v
i
c
e
C
a
p
a
c
i
t
y
4
4
1
50
7
7
2
45
3
40
P
i
c
ki
ng
de
vi
c
e
1
2
3
1
2
3
2
1
3
2
Ba
t
c
h n
um
be
r
1
1
1
2
2
1
1
1
2
2
2
9
1
5
7
2
9
1
8
7
4
8
10
5
4
10
6
3
6
3
L
S
O
S
-H
L
L
S
O
S
-H
L
O
rde
rs
O
r
d
e
r
s
O
r
d
e
r
s
i
z
e
O
rde
rs
F
CF
S
-L
H
F
CF
S
-H
L
S
L
O
S
-L
H
S
L
O
S
-H
L
O
rde
rs
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
.
3
,
J
une
2021 :
25
16
-
2524
2522
in
total
com
ple
ti
on
tim
e
wh
en
com
par
ed
to
S
LOS
-
HL
w
hen
con
si
der
i
ng
10
orde
rs,
it
em
s
per
o
r
der
b
et
w
een 1
and 5, a
nd a la
yout
wi
th
2,000 s
t
or
a
ge posi
ti
on
s
.
Fo
r
al
l
the
eva
luate
d
insta
nce
s,
the
GGA
sa
ves
on
ave
rage
18.
3%
the
t
ot
al
com
pletio
n
tim
e
wh
e
n
com
par
ed
to
t
he
six
ben
c
hma
rk
s;
in
t
his
m
ann
e
r,
th
e
pro
po
s
ed
al
go
rith
m
i
m
pr
ov
es
th
e
eff
ic
ie
ncy
of
order
picking
sig
nifi
cantl
y.
From
the
par
ti
cula
r
re
su
lt
s
of
Table
3,
G
GA
gen
e
r
at
es
ave
rag
e
sa
vings
of
26.6%
w
he
n
O
=1
0.
F
or
in
sta
nces
w
he
n
O
=3
0
an
d
O
=5
0,
GGA
pro
vid
es
a
verage
savi
ngs
of
15.
5%
an
d
12
.
7%
resp
ect
ively
.
Althou
gh
m
ini
m
iz
ing
the
tot
al
picking
ti
m
e
is
not
the
obj
ect
iv
e
f
unct
ion
in
t
he
pro
po
s
e
d
al
gorithm
,
fo
r
the
evaluated
instances
the
GGA
offers
aver
a
ge
savi
ngs
of
4.6%
on
picki
ng
ti
m
e
wh
e
n
com
par
ed
t
o
t
he
ben
c
hm
ark
heurist
ic
s,
s
o
t
he
GGA
i
n
a
ddit
ion
t
o
im
pr
ov
i
ng
cu
stom
e
r
se
rv
ic
e
t
hro
ugh
the
com
pleti
on
tim
e
al
so
i
m
pr
ov
e
s
the
op
e
rat
ing
costs
of
the
or
der
picking.
F
ur
t
her
m
or
e,
th
e
aver
a
ge
com
pu
ti
ng
tim
es
of
the
al
gorithm
wh
en
O
=1
0,
O
=3
0,
a
nd
O
=5
0
are
0.
04,
1.6
0,
an
d
9.8
9
m
inu
te
s
res
pecti
vely
,
wh
ic
h
are
viable
for
th
e
da
il
y op
erati
ons
p
la
nnin
g o
f w
areho
us
es a
nd
distrib
ution ce
nters.
Table
2.
T
otal
com
pleti
on
tim
e savi
ng
s
for G
GA
FCFS
-
LH
FCFS
-
HL
SLOS
-
LH
SLOS
-
HL
LSOS
-
LH
LSOS
-
HL
Av
erage
1
4
.3%
2
2
.4%
1
6
.3%
2
3
.5%
1
4
.4%
1
8
.8%
Maxi
m
u
m
2
7
.9%
3
8
.7%
3
5
.5%
4
0
.7%
2
7
.9%
3
3
.4%
Mini
m
u
m
5
.3%
9
.6%
5
.7%
1
2
.3%
2
.6%
5
.4%
Table
3.
A
ver
a
ge
sa
ving
on
to
ta
l com
pleti
on
tim
e fo
r GG
A
O
FCFS
-
LH
FCFS
-
HL
SLOS
-
LH
SLOS
-
HL
LSOS
-
LH
LSOS
-
HL
10
2
0
,94
%
3
0
,21
%
2
5
,66
%
3
0
,62
%
2
3
,75
%
2
8
,68
%
30
1
1
,15
%
2
0
,64
%
1
2
,54
%
2
1
,45
%
1
1
,32
%
1
5
,65
%
50
1
0
,82
%
1
6
,48
%
1
0
,63
%
1
8
,33
%
8
,15
%
1
2
,02
%
6.
CONCL
US
I
O
N
This
stu
dy
ad
dresses
the
orde
r
batchi
ng
a
nd
sequ
e
ncin
g
prob
le
m
with
m
ulti
ple
picke
rs
(O
BS
PMP)
to
m
ini
m
iz
e
the
m
akesp
an.
We
dealt
with
the
assig
nm
ent
of
order
s
to
batches
an
d
with
the
assig
nm
ent
of
batches
to
pic
king
dev
ic
es
with
heter
og
e
ne
ous
loa
d
ca
pa
ci
ty
.
By
m
ea
ns
of
num
erical
ex
per
im
ents,
it
was
pro
ved
that
the
GGA
offe
rs
s
olu
ti
ons
supe
rior
to
th
os
e
pro
vid
e
d
by
r
ule
-
base
d
he
ur
ist
ic
s.
Conseq
ue
ntly
,
the
GGA
pro
vid
es
ave
rag
e
sa
vin
gs
of
18.
3%
on
m
akesp
a
n
com
par
ed
to
six
widely
use
d
he
ur
ist
ic
s
in
real
war
e
hous
e
e
nvir
on
m
ents,
an
d
the
s
olu
ti
on
s
are
obta
ine
d
in
feasible
c
om
pu
ta
ti
on
al
tim
es
that
al
lo
w
their
app
li
cat
io
n
in
daily
war
e
house
operati
ons.
Ther
e
f
or
e,
im
p
lem
enting
thes
e
so
luti
ons
ca
n
i
m
pr
ove
war
e
house
perform
ance
sign
ific
a
ntly
by
i
m
pr
ovin
g
pr
of
it
m
arg
ins,
reducin
g
op
e
r
at
ing
tim
e
fo
r
the
orde
r
pic
ker
s
,
i
m
pr
ovin
g
c
us
tom
er
serv
ic
e
a
nd
re
du
ci
ng
pi
ckin
g
se
rv
ic
e
t
i
m
es.
Fu
tu
re
re
search
co
uld
a
ddress
t
he
j
oi
nt
orde
r
batchin
g,
se
quencin
g,
assig
nm
ent
an
d
pick
er
r
outi
ng
pro
blem
con
side
ring
3D
wa
reho
us
es,
onli
ne
cu
stom
er
orders a
nd p
ic
king
d
evices
wi
th h
et
er
ogene
ou
s
vel
ocity
.
REFERE
NCE
S
[1]
J.
A.
Cano,
Ale
xande
r
Alber
to
Corre
a
-
Espin
al,
and
Rodrigo
Andrés
Góm
ez
-
Monto
y
a
,
“
An
eva
l
uat
ion
of
pi
cki
n
g
routi
ng
pol
ic
i
es
to
improve
ware
house
eff
iciency
,
”
In
te
rnat
ional
Journal
of
Industrial
E
ngine
ering
an
d
Manage
ment,
vo
l.
8
,
no
.
4
,
pp
.
22
9
-
238,
2017
.
[2]
E.
Ardjm
and,
Om
id
Sanei
Baj
gira
n
,
and
E
y
a
d
Yous
sef
,
“
U
s
ing
li
st
-
base
d
s
imulat
ed
ann
eal
ing
and
genetic
al
gorit
hm
for
or
der
batching
an
d
pic
ker
rou
ti
ng
in
put
wal
l
ba
sed
pic
king
s
y
st
ems
,
”
Appl
i
ed
Soft
Computing
Journal,
vo
l. 75, pp. 106
-
119,
20
19.
[3]
S.
Henn
and
V.
Schm
id,
“
Metaheuri
sti
cs
for
or
der
batching
an
d
seque
nci
ng
in
m
anua
l
orde
r
p
ic
king
s
y
stems
,
”
Computers and
I
ndustrial
Engi
n
e
ering,
vo
l. 66, n
o.
2
,
pp
.
338
-
35
1,
2013
.
[4]
J.
A.
Cano
,
Al
exa
nder
Corre
a
-
Espina
l
,
and
R
odrigo
Góm
ez
-
Monto
y
a
.
,
“
Us
ing
genetic
al
go
rit
hm
s
for
orde
r
bat
ch
ing
in
m
ult
i
-
pa
rallel
-
ai
sl
e
pic
ker
-
to
-
p
art
s
sy
st
ems
,
”
Inter
nati
onal
Journal
of
Appl
ie
d
Dec
ision
Scienc
es,
vol.
13
,
no
.
4
,
pp
.
417
-
434
,
2020
.
[5]
J.
A.
Cano
,
“
Order
Picki
ng
Opti
m
iz
at
ion
B
ase
d
on
a
Pick
er
Rout
ing
Heuri
sti
c:
M
ini
m
iz
ing
Tot
a
l
Tra
ve
le
d
Dist
an
ce
in
W
are
houses,
”
Handbook
of
Re
search
on
th
e
App
li
ca
ti
ons
of
Int
ernati
onal
Tr
anspor
tat
ion
and
Logis
tics
f
or
World
Tr
ade,
G.
Ç.
C
ey
hun
,
Ed.
PA
,
USA:
IGI
Gl
obal,
pp
.
74
-
96
,
2020.
[6]
B.
Mené
nd
ez,
e
t
al
.
,
“
Gene
r
al
V
ari
ab
le
Ne
ighbo
rhood
Sear
ch
fo
r
the
Orde
r
Batc
hing
and
Sequ
e
nci
ng
Problem,”
European
Journ
al
of
Operationa
l
R
ese
arch,
vol
.
263,
no
.
1
,
pp
.
8
2
-
93,
2017
.
[7]
M.
Bustil
lo
,
et
a
l.
,
“
An
al
gori
th
m
for
bat
ch
ing,
seque
nci
ng
and
pic
king
op
erati
o
ns
in
a
w
are
hou
se,
”
In
te
rnation
al
Confe
renc
e
on
I
ndustrial
Engi
n
e
ering
and
S
ystem
s Manage
ment
,
IE
EE IE
SM
20
15,
2015
,
pp
.
84
2
-
849.
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
So
lv
in
g
t
he or
de
r batchi
ng an
d
se
quenci
ng prob
le
m
wi
th…
(
Jose Alej
andro C
ano
)
2523
[8]
S.
Henn,
“
Order
bat
chi
ng
and
se
quenc
ing
for
the
m
ini
m
iz
at
ion
of
the
tot
a
l
ta
rdin
e
ss
in
pic
ker
-
to
-
p
art
ware
houses,
”
Fl
e
xi
bl
e
S
erv
ice
s
and
Manufactu
ring J
ournal,
vol
.
27
,
no
.
1
,
pp
.
8
6
-
114,
2015
.
[9]
J.
Zha
ng
,
et
al
.
,
“
On
-
li
ne
orde
r
batching
and
s
eque
nc
ing
probl
em
with
m
ult
iple
pi
cke
rs:
A
h
ybrid
rul
e
-
base
d
al
gorit
hm
,
”
Applied
Ma
the
mati
ca
l
Mode
ll
ing
,
vo
l. 45, no. 1, pp. 27
1
-
284,
2017
.
[10]
T.
Van
Gi
ls,
e
t
a
l.
,
“
Form
ula
ti
ng
and
Solving
th
e
Inte
gra
te
d
B
at
ch
ing,
Rout
ing,
an
d
Picke
r
Schedu
l
ing
Problem
in
a
Rea
l
-
li
fe
Spare
Parts
W
are
house
,
”
European Jou
rnal
of
Operat
io
nal
R
ese
arch,
vo
l.
277
,
no
.
3
,
pp
.
814
-
830,
2019
.
[11]
S.
Henn,
“
Algorit
hm
s
for
on
-
line
ord
er
ba
tc
hi
ng
in
an
ord
er
pic
king
ware
ho
use,
”
Comput
ers
and
Operatio
ns
Re
search,
vol
.
3
9,
no
.
11
,
pp
.
25
49
-
2563,
2012
.
[12]
E.
Ardjm
and,
et
al.,
“
Minim
iz
in
g
orde
r
pi
cki
ng
m
ake
span
with
m
ult
ipl
e
p
ic
k
ers
in
a
wave
p
ic
k
ing
ware
hous
e
,
”
Inte
rnational
Jo
urnal
of Produc
t
ion
E
conomic
s,
vol.
206
,
no
.
1
,
p
p.
169
-
183
,
201
8.
[13]
T.
L
.
Chen
,
e
t
a
l
.
,
“
An
eff
i
ci
en
t
h
y
brid
al
gor
it
hm
for
int
egr
ated
o
rde
r
batching
,
se
quenc
ing
and
ro
uti
ng
probl
em,”
Inte
rnational
Jo
urnal
of Product
ion
E
conomic
s,
vol.
159
,
pp
.
158
-
167,
2015
.
[14]
B.
Mené
nde
z,
et
al.,
“
Para
l
le
l
var
ia
bl
e
nei
gh
borhood
sea
rch
for
the
m
in
-
m
ax
orde
r
batching
proble
m
,
”
Inte
rnational
Tr
ansacti
ons i
n
Operational R
ese
ar
ch,
vo
l. 24, no. 3
,
pp
.
635
-
662
,
2
017.
[15]
N.
Gade
m
ann
an
d
S.
van
de
Ve
ld
e,
“
Order
b
at
ch
i
ng
to
m
ini
m
ize
t
ota
l
tra
v
el
ti
m
e
i
n
a
p
ara
l
le
l
-
ai
sl
e
ware
house,
”
II
E
Tr
ansacti
ons,
vo
l.
37
,
no
.
1
,
pp
.
6
3
-
75,
2005
.
[16]
M.
M.
Ta
v
akol
and
A.
Sam
i,
“
Parti
cl
e
sw
arm
opti
m
iz
a
ti
on
in
solving
Vehi
cle
Routi
ng
Probl
em,”
Bu
ll
e
ti
n
o
f
El
e
ct
rica
l
Eng
in
ee
ring a
nd
Infor
matic
s
(
BE
EI)
,
v
ol.
2
,
no
.
4
,
p
p
.
2
52
-
257,
2013
.
[17]
S.
Kala
n
ta
ri
an
d
M.
S.
Abad
e
h,
“
GA
SA
-
J
OSH:
A
H
y
brid
Evol
uti
on
ar
y
-
An
nea
l
ing
Approa
ch
for
Job
-
Sho
p
Schedul
ing
Problem,”
Bulletin
of
Elec
tri
cal E
ngi
nee
ring a
nd
Info
rm
ati
cs
(
BE
EI)
,
vol.
2
,
no
.
2
,
pp
.
132
-
140,
2013
.
[18]
T.
L.
Duong,
et
al
.
,
“
Applic
ation
of
a
new
constr
ai
nt
handling
m
et
hod
for
ec
ono
m
ic
dispat
ch
co
nsideri
ng
el
e
ct
r
i
c
m
ark
et
,
”
Bu
ll
e
tin
of
El
ec
tri
cal
Engi
ne
ering
and
Informatic
s
(BE
EI)
,
vol.
9,
no.
4,
pp.
1542
-
1549,
2020,
doi:
10.
11591/eei
.
v9i
4.
2351
.
[19]
S.
Sabika
n,
S.
W
.
Nawawi,
and
Naa
Aziz
.
,
“
Modell
ing
of
tim
e
-
to
col
li
sion
for
unm
anne
d
ae
rial
vehi
c
le
using
par
ticle
s
sw
arm
opti
m
izati
on
,
”
IAE
S
Inte
rnat
i
onal
Journal
o
f
Artificial
Intelli
gen
ce
(
IJ
-
AI)
,
vol.
9,
no
.
3
,
pp.
488
-
496
,
20
20.
[20]
A.
S.
Gi
rsang,
et
al
.
,
“
Fast
Ant
C
olon
y
Optimiz
ation
for
Cluste
rin
g,
”
Indone
sian
J
ournal
of
El
ec
tri
cal
Engi
ne
erin
g
and
Computer
S
ci
en
ce
(
IJE
ECS)
,
vol
.
12
,
no
.
1
,
p
p.
78
-
86
,
2018
.
[21]
Z.
A.
Ali,
et
al.
,
“
An
enha
nce
d
h
y
brid
gen
et
i
c
al
gorit
hm
for
solving
tra
v
el
ing
sale
sm
an
problem
,
”
Indone
sian
Journal
of
Elec
t
rical
Engi
ne
erin
g
and
Computer
Sci
en
ce
(
IJE
ECS
)
,
vol.
18
,
no
.
2
,
p
p.
1035
-
1039
,
2020.
[22]
S.
Koch
and
G.
W
äsc
her
,
“
A
Grouping
Gene
t
ic
Algorit
hm
fo
r
the
Order
Batching
Problem
in
Distribut
ion
W
are
houses,”
Jo
urnal
of Busin
es
s E
conomic
s
,
vol
.
86
,
no
.
1
,
pp
.
1
31
-
153,
2016
.
[23]
M.
Mutingi
and
C.
Mbohw
a,
“Opti
m
iz
ing
Order
Bat
ch
ing
in
Order
Picki
ng
Sy
stems
:
H
y
br
id
Grouping
Gene
t
ic
Algorit
hm
,
”
Gr
o
uping
Gene
ti
c
A
l
gorithms
,
pp.
12
1
-
140,
2017
.
[24]
J.
C.
-
H.
Pan,
et
al
.
,
“
Order
bat
ch
ing
in
a
pic
k
-
and
-
pass
ware
housing
sy
st
em
with
group
gene
tic
algorithm,”
Om
ega,
vol.
57
,
pp
.
238
-
248,
2015
.
[25]
A.
Scholz
,
et
al., “O
rde
r
pic
king w
it
h
m
ult
ipl
e
p
i
cke
rs a
nd
due
d
a
te
s
-
Sim
ult
ane
ou
s soluti
on
of
orde
r
batching
,
batc
h
assignm
ent
and
seque
nci
ng
,
and
pic
ker
ro
uti
ng
proble
m
s,”
European
Journal
of
Operational
Res
earc
h,
vol.
263
,
no.
2
,
pp
.
461
-
4
78,
2017
.
[26]
M.
Matusia
k,
et
al
.
,
“
Util
i
zi
ng
indi
vidual
pic
k
e
r
skill
s
to
impr
ove
orde
r
bat
ch
ing
in
a
ware
h
ouse,
”
Europea
n
Journal
of
Operational
Re
search
,
vol
.
263
,
no
.
3
,
pp
.
888
-
899
,
20
17.
[27]
J.
A.
Cano,
“
Form
ula
ti
ons
for
joi
nt
orde
r
pi
c
king
proble
m
s
in
low
-
le
vel
pi
c
ker
-
to
-
par
t
s
y
st
e
m
s,”
Bul
le
t
in
of
El
e
ct
rica
l
Eng
in
ee
ring a
nd
Infor
matic
s
(
BE
EI)
,
v
ol.
9
,
no
.
2
,
pp
.
8
36
-
844,
2020
.
[28]
J.
A.
Cano,
Ale
xande
r
A.
Corr
ea
-
Espin
al
,
and
Rodr
igo
Andrés
Góm
ez
-
Monto
y
a
.
,
“
Mathe
m
atic
al
progra
m
m
ing
m
odel
ing
for
jo
int
orde
r
ba
tc
hi
ng,
seque
n
ci
ng
and
pic
k
er
rou
ting
proble
m
s
in
m
anua
l
orde
r
pi
cki
ng
s
y
s
te
m
s,
”
Jo
urnal
of
King
Saud
Unive
rs
ity
-
Engi
ne
ering
Sc
i
enc
es
,
vol
.
32
,
n
o.
3
,
pp
.
219
-
22
8,
2019
.
[29]
S.
S.
Sadiq,
et
al
.
,
“
Solving
m
ult
i
-
objecti
v
e
m
aste
r
produc
tion
sche
dule
pr
oble
m
using
me
m
et
ic
al
gori
th
m
,
”
Indone
sian
Journal
of
El
ectric
al
Engi
nee
ring
an
d
Computer
Sci
enc
e
(
IJE
ECS)
,
vol.
18,
no.
2,
pp.
938
-
945,
2020,
doi:
10.
11591/i
j
eecs.
v18.i
2.p
p938
-
94
5.
[30]
A.
K.
Ari
y
an
i,
W
a
y
an
Firdaus
Mahm
udy
,
and
Yus
uf
Pri
y
o
An
ggodo
,
“
H
y
brid
Gene
tic
Algori
th
m
s
and
Sim
ula
te
d
Annea
li
ng
for
Multi
-
tri
p
Veh
icle
Routi
ng
Probl
em
with
Ti
m
e
W
indows,”
Inte
rnational
Journal
of
El
e
ct
rica
l
and
Computer
Engi
n
ee
ring
(
IJE
C
E)
,
vol.
8
,
no
.
6
,
pp
.
4713
-
4723,
201
8.
BIOGR
AP
H
I
ES
OF
A
UTH
ORS
Jose
Alej
and
ro
Can
o
obta
ine
d
his
PhD
in
En
gine
er
ing,
Indus
tr
y
and
Oganiza
ti
ons
from
the
Nati
ona
l
Univer
sit
y
o
f
Colom
bi
a,
and
obtained
his
Master
in
A
dm
ini
strat
iv
e
En
gine
er
ing
and
Bac
he
lor
in
Ind
ustria
l
Eng
ineeri
ng
from
the
sa
m
e
unive
rsit
y
.
He
is
working
as
an
As
socia
t
e
Profess
or
at
the
Univer
sit
y
of
Mede
l
li
n,
Colom
bi
a.
He
obtained
h
is PhD i
n
Engi
nee
ring,
Indust
r
y
and
Oganiz
a
ti
on
s.
His
rese
arc
h
i
nte
rest
in
cl
udes
orde
r
pic
king
s
ystems
,
logi
stic
s,
opti
m
isat
ion
,
and
m
et
ah
eur
ist
i
cs.
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
.
3
,
J
une
2021 :
25
16
-
2524
2524
Pablo
Corté
s
is
a
Ca
the
dr
at
i
c
Pr
ofe
ss
or
at t
he
Se
vil
le
Univer
si
t
y
(Spain)
wher
e
h
e
be
longs t
o
the
Depa
rtment
o
f
I
ndustria
l
Organi
za
t
ion
and
Busin
ess Mana
gement
II. His re
se
arc
h
int
er
est
inc
lud
es
comput
at
ion
al
int
e
ll
ig
en
ce
,
oper
a
ti
ons
re
sea
rch
,
tr
ansport
& logistics,
v
ertical
tra
nsporta
ti
on,
a
nd
uti
l
ity
net
wor
ks.
Emir
o
An
tonio
Camp
o
obta
ine
d
his
Master
in
Industria
l
E
ngine
er
ing
from
the
Nati
on
al
Univer
sit
y
of
Co
lombia,
Colom
bi
a
and
ob
ta
in
ed
h
is
Bac
he
lor
in
In
dustria
l
Engi
ne
e
ring
from
the
sam
e
unive
rsit
y
.
He
is
worki
ng
as
an
oper
at
ions
and
logis
ti
cs
consult
ant
for
diffe
ren
t
m
anuf
ac
turi
ng
a
nd
servic
e
com
pani
es.
His
rese
arc
h
intere
st
in
c
lude
s
oper
ations
m
ana
gement,
logi
stic
s,
opt
imis
at
ion, a
nd
m
et
a
heur
isti
cs.
Alexander
Albert
o
Corr
ea
-
Es
pin
al
is
a
Ti
tul
ar
Profess
or
at
the
Nati
onal
Univer
sit
y
of
Colom
bia
,
Colo
m
bia
.
He
ob
ta
in
ed
his
PhD
from
the
Univ
ersitat
Polit
èc
n
ica
de
C
at
a
lun
y
a,
Spain
.
He
obta
ine
d
his
MS
c
in
Industria
l
Engi
n
ee
ring
f
rom
Univer
sidad
de
los
Andes,
Colom
bia
.
His
rese
arc
h
intere
st
inc
lud
es
opti
m
i
sati
on,
op
erati
on
s
rese
arc
h
,
desig
n
of
expe
r
iment
s,
and
supp
l
y
cha
in
m
ana
geme
nt.
Evaluation Warning : The document was created with Spire.PDF for Python.