I
AE
S
I
nte
rna
t
io
na
l J
o
urna
l o
f
Ar
t
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
5
,
No
.
3
,
Sep
tem
b
er
2
0
1
6
,
p
p
.
95
~
1
0
4
I
SS
N:
2252
-
8938
95
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
jo
u
r
n
a
l.c
o
m/o
n
lin
e/in
d
ex
.
p
h
p
/I
J
AI
Na
tural I
mm
une
Sy
ste
m
Respo
nse
A
s Co
m
plex
e
A
da
ptive
Sy
ste
m
Using
Lea
rning
F
u
zz
y
Cog
nitive M
a
ps
A
hm
ed
T
li
l
i
,
S
a
l
i
m
Ch
ik
h
i
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter S
c
ien
c
e
a
n
d
i
ts
A
p
p
li
c
a
ti
o
n
s
,
Ne
w
F
a
c
u
lt
y
o
f
In
f
o
r
m
a
ti
o
n
T
e
c
h
n
o
lo
g
ies
a
n
d
Co
m
m
u
n
ica
ti
o
n
A
b
d
e
lh
a
m
id
M
e
h
ri
Un
iv
e
rsity
,
C
o
n
sta
n
ti
n
e
,
A
lg
e
ria
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
n
4
,
2
0
1
6
R
ev
i
s
ed
A
u
g
7
,
2
0
1
6
A
cc
ep
ted
A
u
g
u
s
t 2
3
,
2
0
1
6
In
th
e
Na
tu
ra
l
Im
m
u
n
e
S
y
ste
m
s
NIS
,
a
d
a
p
ti
v
e
a
n
d
e
m
e
rg
e
n
t
b
e
h
a
v
io
rs
re
su
lt
f
ro
m
th
e
b
e
h
a
v
io
rs
o
f
e
a
c
h
c
e
ll
a
n
d
th
e
ir
i
n
tera
c
ti
o
n
s
w
it
h
o
th
e
r
c
e
ll
s
a
n
d
e
n
v
iro
n
m
e
n
t.
M
o
d
e
li
n
g
a
n
d
S
im
u
latin
g
NIS
re
q
u
ires
a
g
g
re
g
a
ti
n
g
th
e
se
c
o
g
n
it
iv
e
in
tera
c
ti
o
n
s
b
e
tw
e
e
n
th
e
in
d
iv
id
u
a
l
c
e
ll
s
a
n
d
t
h
e
e
n
v
iro
n
m
e
n
t.
In
las
t
y
e
a
rs
th
e
F
u
z
z
y
Co
g
n
it
iv
e
M
a
p
s
(F
CM
)
h
a
s
b
e
e
n
sh
o
w
n
to
b
e
a
c
o
n
v
e
n
ien
t
to
o
l
f
o
r
m
o
d
e
li
n
g
,
c
o
n
tr
o
ll
i
n
g
a
n
d
sim
u
latin
g
c
o
m
p
lex
s
y
ste
m
s.
In
th
is
p
a
p
e
r,
a
n
e
w
t
y
p
e
o
f
le
a
rn
in
g
f
u
z
z
y
c
o
g
n
it
iv
e
m
a
p
s
(LF
CM
)
h
a
v
e
b
e
e
n
p
r
o
p
o
se
d
a
s
a
n
e
x
ten
sio
n
o
f
trad
it
io
n
a
l
F
C
M
f
o
r
m
o
d
e
li
n
g
c
o
m
p
le
x
a
d
a
p
ti
v
e
s
y
ste
m
is
d
e
s
c
rib
e
d
.
O
u
r
a
p
p
r
o
a
c
h
is
su
m
m
a
rize
d
in
tw
o
m
a
jo
r
id
e
a
s:
T
h
e
f
irst
o
n
e
is
to
i
n
c
re
a
se
th
e
re
in
f
o
rc
e
m
e
n
t
lea
rn
in
g
c
a
p
a
b
il
it
ies
o
f
th
e
F
CM
b
y
u
sin
g
a
n
a
d
a
p
tatio
n
o
f
Q
-
lea
rn
in
g
tec
h
n
iq
u
e
a
n
d
t
h
e
se
c
o
n
d
o
n
e
is
to
f
o
ste
r
d
iv
e
rsit
y
o
f
c
o
n
c
e
p
t'
s
sta
te
s
w
it
h
in
th
e
F
CM
b
y
a
d
o
p
ti
n
g
a
n
IF
-
T
HEN
ru
le
b
a
se
d
s
y
ste
m
.
T
h
r
o
u
g
h
m
o
d
e
li
n
g
a
n
d
sim
u
latin
g
re
sp
o
n
se
o
f
n
a
tu
ra
l
imm
u
n
e
sy
ste
m
,
w
e
sh
o
w
th
e
e
f
fe
c
ti
v
e
n
e
s
s
o
f
th
e
p
ro
p
o
se
d
a
p
p
ro
a
c
h
in
m
o
d
e
li
n
g
CA
S
s.
K
ey
w
o
r
d
:
C
o
m
p
le
x
A
d
ap
tiv
e
S
y
s
te
m
Fu
zz
y
C
o
g
n
it
iv
e
Ma
p
s
I
m
m
u
n
e
S
y
s
te
m
R
e
s
p
o
n
s
e
R
ein
f
o
r
ce
m
e
n
t
L
ea
r
n
i
n
g
Co
p
y
rig
h
t
©
2
0
1
6
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
A
h
m
e
d
T
l
i
li
,
Dep
ar
t
m
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
an
d
its
A
p
p
licatio
n
s
,
Ne
w
Facu
l
t
y
o
f
I
n
f
o
r
m
atio
n
T
ec
h
n
o
lo
g
ies a
n
d
C
o
m
m
u
n
ica
ti
o
n
A
b
d
el
h
a
m
id
Me
h
r
i U
n
iv
er
s
it
y
,
C
o
n
s
tan
ti
n
e,
A
l
g
er
ia
.
E
m
ail:
a
g
en
t2
5
0
0
0
@
y
a
h
o
o
.
co
m
1.
I
NT
RO
D
UCT
I
O
N
A
C
o
m
p
le
x
A
d
ap
ti
v
e
S
y
s
te
m
(
C
A
S)
[
2
1
]
is
d
ef
i
n
ed
as
a
co
l
lectio
n
o
f
en
t
ities
(
ag
e
n
ts
),
w
i
th
s
i
m
p
le
r
u
les
o
f
b
eh
a
v
io
r
,
m
er
g
ed
in
a
d
y
n
a
m
ic
an
d
u
n
k
n
o
w
n
e
n
v
ir
o
n
m
e
n
t
an
d
ab
le
to
ad
ap
t
to
it
b
y
lear
n
in
g
ex
p
er
ien
ce
s
.
T
h
e
o
v
er
all
ad
ap
tatio
n
to
t
h
e
e
n
v
ir
o
n
m
e
n
t
ap
p
ea
r
s
th
r
o
u
g
h
t
h
e
lo
ca
l
b
eh
a
v
i
o
r
o
f
en
titi
e
s
t
h
at
i
s
ad
ap
tiv
e.
Fo
u
n
d
i
n
n
at
u
r
e,
m
a
n
y
b
io
lo
g
ical
an
d
s
o
cial
s
y
s
te
m
s
ar
e
s
i
m
ilar
to
th
e
C
AS:
t
h
e
i
m
m
u
n
e
s
y
s
te
m
,
b
ir
d
f
lo
ck
s
,
t
h
e
ce
ll,
i
n
s
ec
t
co
lo
n
ies,
b
r
ain
,
ec
o
n
o
m
ic
m
ar
k
ets
etc.
.
.
.
All
t
h
ese
s
y
s
te
m
s
a
r
e
ch
ar
ac
ter
ized
b
y
th
eir
t
w
o
k
e
y
co
n
ce
p
ts
,
n
a
m
el
y
th
e
e
m
er
g
e
n
ce
o
f
g
lo
b
al
b
eh
av
io
r
,
w
h
ich
i
s
d
u
e
to
o
f
th
e
lack
o
f
ce
n
tr
alize
d
co
n
tr
o
l a
n
d
m
ea
s
u
r
i
n
g
s
e
lf
-
o
r
g
an
iza
tio
n
ad
ap
tatio
n
to
th
e
e
n
v
ir
o
n
m
en
t b
y
r
elativ
e
lear
n
i
n
g
.
I
n
th
e
Nat
u
r
al
I
m
m
u
n
e
S
y
s
t
e
m
s
(
NI
S),
e
m
er
g
e
n
t
b
eh
a
v
i
o
r
s
r
esu
lt
f
r
o
m
t
h
e
b
eh
a
v
io
r
s
o
f
ea
c
h
in
d
iv
id
u
al
ce
ll
a
n
d
th
eir
i
n
t
er
ac
tio
n
s
w
it
h
t
h
e
en
v
ir
o
n
m
en
t.
Mo
d
elin
g
NI
S
r
eq
u
ir
es
in
co
r
p
o
r
atin
g
t
h
ese
ad
ap
tiv
e
in
ter
ac
tio
n
s
a
m
o
n
g
th
e
in
d
i
v
id
u
al
ce
ll
s
an
d
th
e
en
v
ir
o
n
m
e
n
t.
Mo
d
elin
g
ap
p
r
o
ac
h
es
f
o
r
NI
S
ar
e
g
r
o
u
p
ed
in
to
t
w
o
ca
te
g
o
r
ies:
m
at
h
e
m
atica
l
m
o
d
els
g
en
er
all
y
ta
k
e
th
e
f
o
r
m
o
f
p
ar
tial
d
i
f
f
er
en
tial
eq
u
at
io
n
s
,
an
d
c
ell
-
b
ased
m
o
d
els
s
i
m
u
la
te
ea
ch
in
d
iv
id
u
al
ce
ll
b
eh
av
io
r
an
d
in
ter
ac
tio
n
s
b
et
w
ee
n
th
e
m
e
n
ab
li
n
g
t
h
e
o
b
s
er
v
atio
n
o
f
t
h
e
e
m
er
g
e
n
t
b
eh
av
io
r
.
T
h
is
s
t
u
d
y
f
o
c
u
s
e
s
o
n
t
h
e
ce
l
l
-
b
ased
m
o
d
els
o
f
NI
S,
an
d
m
ai
n
l
y
,
t
h
e
tech
n
ical
asp
ec
t
o
f
t
h
e
f
u
zz
y
r
u
le
-
b
ased
s
i
m
u
latio
n
m
et
h
o
d
f
o
r
NI
S
is
d
esc
r
ib
ed
.
Ho
w
to
i
m
p
le
m
en
t
th
e
ce
l
l
b
eh
av
io
r
s
a
n
d
t
h
e
i
n
ter
ac
tio
n
s
w
ith
t
h
e
e
n
v
ir
o
n
m
e
n
t
in
to
th
e
co
m
p
u
tat
io
n
al
d
o
m
ain
is
d
is
cu
s
s
ed
.
T
h
e
s
y
s
te
m
b
eh
av
io
r
s
d
escr
ib
ed
in
t
h
is
p
a
p
er
ar
e
d
if
f
er
en
tiatio
n
s
m
ec
h
a
n
is
m
b
et
w
ee
n
s
el
f
an
d
n
o
-
s
el
f
ce
lls
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
IJ
-
AI
Vo
l.
5
,
No
.
3
,
Sep
tem
b
er
2
0
1
6
:
95
–
104
96
L
ast
l
y
to
g
et
a
b
etter
u
n
d
er
s
tan
d
i
n
g
w
h
y
N
I
S
is
co
n
s
id
er
ed
as
co
m
p
le
x
ad
ap
tiv
e
s
y
s
te
m
th
e
f
o
llo
w
in
g
p
o
in
t
s
m
a
y
b
e
r
elev
an
t:
1.
NI
S is
a
d
ec
en
tr
alize
d
s
y
s
te
m
.
2.
NI
S
m
ec
h
an
i
s
m
is
a
co
g
n
iti
v
e
task
.
3.
T
h
er
e
ar
e
s
im
ilar
it
ies b
et
w
ee
n
NI
S a
n
d
ad
ap
tiv
e
s
o
cial
in
s
ec
ts
co
lo
n
ies,
e.
g
.
a
n
t
s
a
n
d
b
ee
s
.
4.
T
h
e
o
v
er
all
b
eh
av
io
r
ad
ap
tatio
n
o
f
NI
S
ap
p
ea
r
s
th
r
o
u
g
h
t
h
e
l
o
ca
l b
eh
av
io
r
ad
ap
tatio
n
o
f
e
ac
h
ce
ll
.
I
n
th
e
litt
ér
at
u
r
e
,
ag
e
n
t
-
b
a
s
ed
m
o
d
el
s
(
A
B
M)
an
d
ce
ll
u
lar
au
to
m
ata
(
C
A
)
ar
e
t
w
o
o
f
th
e
co
m
m
o
n
l
y
u
s
ed
m
et
h
o
d
o
lo
g
ies
f
o
r
m
o
d
e
lin
g
o
f
NI
S
[
1
6
,
17
].
T
h
e
A
B
M
ar
e
cr
iticized
f
o
r
th
eir
co
m
p
lex
it
y
,
b
y
a
g
ain
s
t
th
e
C
As ar
e
also
cr
iticized
f
o
r
lack
o
f
e
n
v
ir
o
n
m
e
n
t.
R
ec
en
t
l
y
,
m
a
n
y
s
t
u
d
ies
h
a
v
e
u
s
i
n
g
F
C
Ms
a
n
d
t
h
eir
m
u
ltip
l
e
ex
te
n
s
io
n
s
[
1
-
2
,
1
9
]
,
to
m
o
d
el
co
m
p
lex
s
y
s
te
m
s
w
h
er
e
C
A
S
s
ar
e
a
s
p
ec
ial
ca
s
e,
an
d
h
a
v
e
g
i
v
e
n
e
n
co
u
r
ag
i
n
g
r
esu
l
ts
[
3
-
5]
.
I
n
th
is
p
ap
er
w
e
p
r
ese
n
t
a
n
ap
p
r
o
ac
h
f
o
r
m
o
d
eli
n
g
C
A
S
s
b
ased
o
n
o
n
e
co
n
n
ec
tio
n
o
f
th
e
f
u
zz
y
co
g
n
it
iv
e
m
ap
s
(
FC
M
)
th
eo
r
y
a
n
d
ad
p
ativ
e
r
ein
f
o
r
ce
m
e
n
t
lear
n
in
g
alg
o
r
it
h
m
Q
-
L
ea
r
n
i
n
g
,
a
ls
o
lin
ea
r
w
ei
g
h
t
ad
ap
tatio
n
m
et
h
o
d
b
ased
o
n
h
eb
b
ian
lear
n
i
n
g
al
g
o
r
ith
m
[
7
]
d
ev
elo
p
ed
f
o
r
n
eu
r
al
n
et
w
o
r
k
s
is
u
s
ed
to
tr
ain
FC
M
b
y
u
p
d
ates
o
n
l
y
t
h
e
in
itia
l
w
ei
g
h
ts
o
f
FC
M.
2.
RE
S
E
ARCH
M
E
T
H
O
D
T
h
e
Natu
r
al
I
m
m
u
n
e
S
y
s
te
m
(
NI
S)
is
o
n
e
o
f
t
h
e
m
o
s
t
ad
v
an
ce
d
an
d
co
m
p
le
x
ad
ap
ti
v
e
b
io
lo
g
ical
s
y
s
te
m
s
.
I
n
o
r
d
er
to
m
ai
n
tai
n
i
n
d
ep
en
d
en
ce
a
n
d
h
elp
i
n
av
o
id
in
g
au
to
i
m
m
u
n
i
t
y
,
a
l
iv
in
g
o
r
g
an
i
s
m
m
u
s
t
p
r
ev
en
t
in
v
as
io
n
b
y
n
u
m
er
o
u
s
m
icr
o
o
r
g
an
is
m
s
a
n
d
h
ar
m
f
u
l
s
u
b
s
tan
ce
s
f
r
o
m
t
h
e
en
v
i
r
o
n
m
e
n
t,
an
d
m
u
s
t
h
an
d
le
t
h
o
s
e
t
h
at
d
o
e
n
ter
.
NI
S
h
a
s
a
d
o
u
b
le
o
b
j
ec
tiv
e
s
i
n
ce
it
h
a
s
to
m
ax
i
m
i
ze
h
ar
m
f
u
l
a
n
ti
g
en
s
eli
m
i
n
atio
n
an
d
at
th
e
s
a
m
e
ti
m
e
m
i
n
i
m
i
ze
h
ar
m
to
s
elf
(
au
to
i
m
m
u
n
i
t
y
)
.
I
m
m
u
n
it
y
i
n
cl
u
d
es
f
u
n
cti
o
n
s
to
d
is
tin
g
u
i
s
h
b
et
w
ee
n
s
el
f
a
n
d
n
o
s
el
f
co
m
p
o
n
en
ts
(
ce
ll
s
)
,
an
d
to
r
e
m
o
v
e
th
e
latter
.
T
h
e
i
m
m
u
n
e
r
esp
o
n
s
e
p
r
o
ce
s
s
i
s
d
ep
icted
in
Fig
u
r
e
5
.
Fig
u
r
e
1
.
Natu
r
al
i
m
m
u
n
e
r
es
p
o
n
s
e
m
ec
h
an
is
m
s
I
m
m
u
n
e
r
esp
o
n
s
es
ar
e
i
m
p
le
m
en
ted
b
y
s
p
ec
ial
ce
lls
ca
lle
d
l
y
m
p
h
o
c
y
tes.
T
h
ese
ce
ll
s
ar
e
f
u
r
n
i
s
h
ed
w
it
h
a
s
o
r
t
o
f
a
n
te
n
n
a
s
to
r
ec
o
g
n
ize
n
o
s
el
f
c
h
e
m
ical
s
tr
u
c
tu
r
es
(
a
n
ti
g
en
d
eter
m
i
n
an
t
s
)
n
o
t
in
cl
u
d
ed
a
m
o
n
g
th
e
s
el
f
co
m
p
o
n
e
n
t
s
.
B
ce
lls
an
d
T
ce
lls
ar
e
tw
o
m
aj
o
r
ty
p
es
o
f
l
y
m
p
h
o
c
y
tes
a
n
d
ar
e
d
er
iv
ed
f
r
o
m
h
e
m
ato
p
o
ietic
s
te
m
ce
lls
i
n
th
e
b
o
n
e
m
ar
r
o
w
.
B
ce
lls
ar
e
in
v
o
lv
ed
i
n
t
h
e
h
u
m
o
r
al
i
m
m
u
n
e
r
esp
o
n
s
e.
T
h
ey
w
o
r
k
c
h
ief
l
y
b
y
s
ec
r
etin
g
s
u
b
s
tan
ce
s
ca
lled
an
tib
o
d
ies
in
to
th
e
b
o
d
y
‟
s
f
l
u
id
s
.
Mo
r
eo
v
er
,
T
ce
lls
ar
e
in
v
o
lv
ed
in
ce
ll
-
m
ed
iate
d
i
m
m
u
n
e
r
esp
o
n
s
e
a
n
d
ca
n
b
e
s
u
b
d
iv
id
ed
in
to
t
w
o
g
r
o
u
p
s
:
th
e
h
elp
er
T
ce
ll
(
T
h
)
an
d
th
e
s
u
p
p
r
ess
o
r
T
ce
ll
(
T
s
)
.
A
ctiv
a
tio
n
o
f
T
ce
lls
b
y
an
ti
g
en
-
p
r
ese
n
t
i
n
g
ce
lls
(
A
P
C
s
)
in
l
y
m
p
h
n
o
d
es
i
s
a
k
e
y
st
ar
tin
g
ev
en
t
in
m
a
n
y
n
atu
r
al
i
m
m
u
n
e
r
esp
o
n
s
es.
T
h
ce
ll
s
ar
e
p
ar
tic
u
lar
l
y
i
m
p
o
r
t
an
t
i
n
th
e
i
m
m
u
n
e
s
y
s
te
m
.
B
ec
au
s
e
th
e
T
h
ce
ll
s
ca
n
n
o
t
r
ec
o
g
n
ize
an
tig
e
n
d
ir
e
ctl
y
,
t
h
e
a
n
ti
g
en
h
a
v
e
to
b
e
p
r
o
ce
s
s
ed
b
y
s
o
m
e
o
t
h
er
ac
ce
s
s
o
r
y
ce
l
ls
(
An
ti
g
e
n
-
P
r
esen
tin
g
C
ell
s
(
A
P
C
)
)
in
ad
v
an
ce
.
T
h
e
ac
tiv
atio
n
o
f
T
h
c
ells
d
ep
en
d
s
o
n
th
e
in
ter
ac
tio
n
o
f
T
ce
ll
r
ec
ep
to
r
s
(
T
C
R
s
)
,
w
h
ic
h
ar
e
m
o
lec
u
les
f
o
u
n
d
o
n
t
h
e
s
u
r
f
ac
e
o
f
T
ce
lls
t
h
at
ar
e
g
e
n
er
all
y
r
e
s
p
o
n
s
i
b
le
f
o
r
r
ec
o
g
n
izi
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
I
SS
N:
2252
-
8938
N
a
tu
r
a
l I
mmu
n
e
S
ystem
R
esp
o
n
s
e
A
s
C
o
mp
lexe
A
d
a
p
tive
S
ystem
Usi
n
g
Lea
r
n
in
g
F
u
z
z
y
…(
A
h
m
e
d
T
l
il
i
)
97
p
ep
tid
es
b
o
u
n
d
to
Ma
j
o
r
His
t
o
co
m
p
atib
ilit
y
C
o
m
p
le
x
(
M
H
C
)
m
o
lec
u
le
s
,
w
it
h
p
ep
tid
es
t
h
at
ar
e
d
i
s
p
la
y
ed
o
n
th
e
s
u
r
f
ac
e
o
f
an
ti
g
e
n
-
p
r
ese
n
tin
g
ce
lls
.
T
ce
ll
r
ec
ep
to
r
s
s
ca
n
th
e
s
u
r
f
ac
e
o
f
an
ti
g
e
n
p
r
esen
ti
n
g
ce
ll
s
f
o
r
s
p
ec
if
ic
p
ep
tid
es
b
o
u
n
d
to
m
o
lecu
les
o
f
th
e
MH
C
.
I
f
t
h
e
s
p
ec
if
ic
p
ep
tid
es
ar
e
f
o
u
n
d
,
th
e
T
h
ce
ll
is
ac
tiv
ated
,
an
d
s
ec
r
etes
in
ter
le
u
k
in
(
I
L
+)
an
d
o
th
er
v
ar
io
u
s
c
h
e
m
ical
s
i
g
n
als.
T
h
e
s
ec
r
eted
in
ter
leu
k
i
n
p
lay
s
an
i
m
p
o
r
tan
t
r
o
le
in
ac
tiv
ati
n
g
B
ce
lls
.
O
n
th
e
o
t
h
er
h
a
n
d
,
th
e
s
u
p
p
r
ess
o
r
T
ce
ll
(
T
s
)
ca
n
s
ec
r
ete
s
u
p
p
r
ess
i
n
g
s
i
g
n
a
l
in
ter
le
u
k
i
n
(
I
L
-
)
to
in
h
i
b
it t
h
e
ac
tio
n
o
f
i
m
m
u
n
e
r
esp
o
n
s
e.
Fig
u
r
e
2
.
Ou
r
C
o
m
p
le
x
ad
ap
tiv
e
ar
tif
ic
ial
i
m
m
u
n
e
s
y
s
te
m
(
C
AA
I
S)
m
o
d
eled
in
t
h
e
b
ac
k
g
r
o
u
n
d
o
f
th
e
L
F
C
M.
Mo
to
r
co
n
ce
p
t T
h
,
af
ter
ac
tiv
atio
n
p
r
o
ce
s
s
,
h
a
s
t
w
o
p
o
s
s
ib
le
ac
tio
n
s
SE
L
F a
n
d
N
O
-
SE
L
F a
ct
io
n
as
r
esp
o
n
s
e
to
en
v
ir
o
n
m
e
n
t.
P
r
o
p
o
s
ed
f
r
a
m
e
w
o
r
k
f
o
r
m
o
d
elin
g
a
n
d
s
i
m
u
lati
n
g
n
at
u
r
a
l
i
m
m
u
n
e
s
y
s
te
m
r
esp
o
n
s
e
a
s
co
m
p
le
x
ad
ap
tiv
e
s
y
s
te
m
is
s
u
m
m
ar
ize
d
in
F
ig
u
r
e
3
.
T
h
e
m
ain
id
ea
i
s
a
co
n
n
ec
tio
n
b
et
w
ee
n
t
h
e
f
u
zz
y
co
g
n
iti
v
e
m
ap
s
an
d
r
ein
f
o
r
ce
m
e
n
t le
ar
n
i
n
g
.
T
h
e
f
ir
s
t s
tep
is
to
m
o
d
el
t
h
e
s
y
s
te
m
in
t
h
e
b
ac
k
g
r
o
u
n
d
o
f
t
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
,
i.e
.
th
e
FC
M.
T
h
en
w
e
d
escr
i
b
e
th
e
r
e
f
lecto
r
s
co
n
ce
p
ts
,
i
n
t
er
m
ed
iates
co
n
ce
p
ts
a
n
d
m
o
to
r
s
co
n
ce
p
ts
,
a
n
d
i
n
th
e
las
t
p
o
in
t
o
f
th
e
f
ir
s
t
s
tep
,
w
e
d
eter
m
in
e
a
n
d
d
escr
ib
e
th
e
r
elatio
n
o
r
li
n
k
b
et
w
ee
n
c
o
n
ce
p
ts
w
it
h
t
h
eir
v
alu
e
s
w
ei
g
h
t.
I
n
t
h
e
s
ec
o
n
d
s
tep
,
th
e
s
y
s
te
m
is
ca
p
ab
le
o
f
lear
n
in
g
f
r
o
m
e
x
p
er
ien
ce
w
it
h
th
e
e
n
v
ir
o
n
m
e
n
t
u
s
i
n
g
a
r
e
w
ar
d
-
p
u
n
is
h
m
e
n
t
p
r
o
ce
d
u
r
e,
ca
lled
r
ein
f
o
r
ce
m
e
n
t
lear
n
i
n
g
b
ased
ad
ap
tiv
e
Q
-
lear
n
i
n
g
alg
o
r
it
h
m
(
Fig
u
r
e
2
)
.
Fig
u
r
e3
.
F
r
a
m
e
w
o
r
k
o
f
t
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
R
ein
f
o
r
ce
m
en
t
lear
n
in
g
is
ca
lled
o
n
-
lin
e
m
eth
o
d
in
m
ac
h
in
e
lear
n
in
g
th
eo
r
y
an
d
th
e
in
ter
ac
tio
n
b
etw
ee
n
th
e
lear
n
in
g
NI
S
m
o
d
eled
an
d
w
ith
en
v
ir
o
n
m
en
t,
Fig
u
r
e
2
,
is
its
m
ain
s
o
u
r
ce
o
f
in
tellig
en
ce
.
A
m
o
s
t
u
s
ed
R
L
alg
o
r
ith
m
is
Q
-
L
ea
r
n
in
g
[
9
]
in
m
ac
h
in
e
lear
n
in
g
,
w
o
r
k
s
b
y
lear
n
in
g
an
ac
tio
n
-
s
tate
v
alu
e
f
u
n
ctio
n
th
at
ex
p
r
ess
es
cu
m
u
lativ
e
r
ew
ar
d
o
f
tak
in
g
a
g
iv
en
ac
tio
n
in
a
g
iv
en
s
tate.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
IJ
-
AI
Vo
l.
5
,
No
.
3
,
Sep
tem
b
er
2
0
1
6
:
95
–
104
98
Fig
u
r
e
4
.
T
h
e
r
ein
f
o
r
ce
m
en
t le
ar
n
in
g
m
o
d
el.
2
.
1
.
F
uzzy
Co
g
nitiv
e
M
a
ps
T
h
e
ter
m
co
g
n
iti
v
e
m
ap
(
C
M)
ap
p
ea
r
s
f
o
r
th
e
f
ir
s
t
ti
m
e
in
1
9
4
8
's
i
n
ar
ticle
b
y
E
.
T
o
l
m
an
[
1
0
]
co
g
n
iti
v
e
m
ap
s
in
r
at
s
a
n
d
m
e
n
to
d
escr
ib
e
t
h
e
ab
s
tr
ac
t
m
e
n
tal
r
ep
r
esen
tatio
n
o
f
s
p
ac
e
b
u
i
lt
b
y
r
ats
tr
ain
ed
to
n
av
i
g
ate
i
n
t
h
e
lab
y
r
i
n
t
h
.
T
h
e
ter
m
F
C
M
(
F
u
zz
y
C
o
g
n
iti
v
e
Ma
p
)
w
as
i
n
tr
o
d
u
ce
d
in
1
9
8
6
b
y
B
.
Ko
s
k
o
[
2
]
,
to
d
escr
ib
e
a
s
i
m
p
le
e
x
te
n
s
io
n
o
f
C
Ms
b
y
th
e
co
m
b
in
a
tio
n
o
f
f
u
zz
y
lo
g
ic
a
n
d
ar
tif
ic
ial
n
eu
r
al
n
et
w
o
r
k
s
.
T
h
is
r
o
b
u
s
t
co
m
b
in
a
tio
n
g
i
v
en
F
C
Ms
a
s
tr
u
ct
u
r
e
s
i
m
i
lar
to
ar
tif
icial
r
ec
u
r
r
en
t
n
eu
r
al
n
et
w
o
r
k
s
(
A
r
tif
ic
ial
R
ec
u
r
r
en
t
Ne
u
r
al
Net
w
o
r
k
AR
NN.
F
C
M
s
(
Fi
g
u
r
e
3
)
ca
n
d
escr
ib
e
th
e
co
m
p
le
x
b
eh
a
v
io
r
o
f
e
n
titi
e
s
.
T
h
e
y
ar
e
r
ep
r
esen
ted
as
d
ir
ec
ted
g
r
ap
h
s
w
h
o
s
e
n
o
d
es
ar
e
co
n
ce
p
ts
(
class
i
f
ied
in
to
th
r
ee
t
y
p
es
:
s
en
s
o
r
y
,
m
o
to
r
an
d
ef
f
ec
to
r
s
)
an
d
t
h
e
ar
c
s
r
ep
r
esen
t
ca
u
s
al
r
elatio
n
s
h
ip
s
b
et
w
e
e
n
t
h
ese
co
n
ce
p
ts
.
E
ac
h
ar
c
f
r
o
m
o
n
e
co
n
ce
p
t
C
i
to
o
n
e
co
n
ce
p
t
C
j
is
a
s
s
o
ciate
d
w
it
h
a
w
e
ig
h
t
w
ij
r
ef
lec
tin
g
a
r
elatio
n
s
h
ip
o
f
in
h
ib
itio
n
(
w
ij
<0
)
o
r
ex
citatio
n
(w
ij
>
0
)
.
E
ac
h
co
n
ce
p
t
is
a
s
s
o
ciate
d
w
it
h
a
d
eg
r
ee
o
f
ac
ti
v
atio
n
,
r
ep
r
ese
n
t
'
s
t
h
e
s
tate
at
ti
m
e
t,
an
d
ca
n
b
e
m
o
d
i
f
ied
o
v
er
ti
m
e.
T
h
e
d
y
n
a
m
ics
o
f
an
F
C
M
ca
n
b
e
s
u
m
m
ar
ized
in
o
n
e
c
y
cle
(
f
r
o
m
t
t
o
t
+1
)
b
y
u
p
d
ati
n
g
th
e
ac
ti
v
atio
n
s
v
ec
to
r
.
Fig
u
r
e
5
.
An
FC
M
a
s
a
g
r
ap
h
T
h
e
f
o
llo
w
i
n
g
g
i
v
es
a
f
o
r
m
a
l
d
escr
ip
tio
n
o
f
an
FC
M
[
6
]
.
K
d
en
o
tes
o
n
e
o
f
th
e
r
in
g
s
or
,
b
y
δ
o
n
e
o
f
th
e
n
u
m
b
er
s
0
o
r
1
,
f
o
r
V
o
n
e
o
f
th
e
s
ets
{0
,
1
},
{
-
1
.
0
,
1
},
o
r
[
-
δ
,
a]
.
L
et
(
n
,
t
0
)
∈
I
N²
an
d
k
∈
*
+
.
An
F
C
M
F is
a
s
ix
f
o
ld
(
C
,
K
,
W
,
A
,
f
a
,
R
)
:
a.
C
= {
C
1
,…,C
n
}
i
s
th
e
s
et
o
f
n
co
n
ce
p
ts
f
o
r
m
in
g
t
h
e
n
o
d
es o
f
a
g
r
ap
h
.
b.
K
⊂
C
×
C
is
t
h
e
s
et
o
f
ar
cs (
C
i
, C
j
)
f
r
o
m
C
i
to
C
j
.
c.
W
:
C
×C
→K
(C
i
,
C
j
)
→
W
ij
is
a
f
u
n
ctio
n
o
f
C
×C
to
IR
ass
o
ciatin
g
a
w
eig
h
t
W
ij
to
a
p
air
o
f
co
n
ce
p
ts
(C
i
,
C
j
)
,
w
ith
W
ij
=
0
if
(C
i
,
C
j
)
∉
A
,
o
r
W
ij
eq
u
al
to
th
e
w
eig
h
t
o
f
th
e
ed
g
e
if
(C
i
,
C
j
)
∈
A
.
No
te
th
at
W
(
C
×
C
)
=
(
W
ij
)
∈
K
n
×
n
is
a
m
atr
ix
o
f
M
n
(
I
R
)
.
a.
A
:
C
→V
n
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
I
SS
N:
2252
-
8938
N
a
tu
r
a
l I
mmu
n
e
S
ystem
R
esp
o
n
s
e
A
s
C
o
mp
lexe
A
d
a
p
tive
S
ystem
Usi
n
g
Lea
r
n
in
g
F
u
z
z
y
…(
A
h
m
e
d
T
l
il
i
)
99
C
i
→a
i
is
a
f
u
n
ctio
n
th
a
t
m
ap
s
ea
ch
co
n
ce
p
t
C
i
to
t
h
e
s
eq
u
en
ce
o
f
i
ts
ac
ti
v
atio
n
d
eg
r
ee
at
th
e
m
o
m
e
n
t
t
∈
I
N,
a
i
(
t)
∈
V
is
its
d
eg
r
ee
o
f
ac
tiv
at
io
n
at
th
e
m
o
m
e
n
t
t
.
W
e
No
te
a
(
t)
=
[
(
a
i
(
t)
)
i
∈
[[1,
n
]
]
T
th
e
v
ec
to
r
o
f
ac
tiv
at
io
n
s
at
t
h
e
m
o
m
en
t
t
.
b.
f
a
∈
(
IR
n
)
N
is
a
s
eq
u
en
ce
o
f
v
ec
to
r
s
o
f
f
o
r
ce
d
ac
tiv
atio
n
s
s
u
c
h
as
f
o
r
i
∈
[
1
,
n
]
an
d
t
≥
t
0
is
th
e
f
o
r
ce
d
ac
tiv
atio
n
o
f
th
e
co
n
ce
p
t C
i
at
m
o
m
e
n
t t.
c.
(
R)
is
a
r
ec
u
r
r
en
ce
r
elat
io
n
s
h
i
p
o
n
t ≥
t
0
b
et
w
ee
n
a
i
(
t +
1
)
,
a
i
(t
)
an
d
f
ai
(
t
)
f
o
r
i
∈
[
1
,
n
]
in
d
icatin
g
t
h
e
d
y
n
a
m
ics o
f
th
e
m
ap
F
.
(
R
)
:
∀
i
∈
[
1
,
n
]
,
∀
t
≥
t
0,
a
i
(t
0
)
=
0
a
i
(
t+1
)
=
σ
[
g
i
(
ƒ
ai
(
t)
,
∑
j
∈
[1,
n]
W
ij
a
j
(
t
))
]
Fig
u
r
e
6
.
C
o
g
n
itiv
e
m
ap
s
s
tan
d
ar
d
izin
g
f
u
n
ctio
n
.
T
h
e
Mo
d
e
r
e
p
r
esen
ted
b
y
th
e
f
u
n
ctio
n
is
to
r
ed
u
ce
t
h
e
v
al
u
e
o
f
co
n
ce
p
t
s
w
it
h
i
n
th
e
r
an
g
e
o
f
v
al
u
es
tak
e
n
a
s
t
h
e
ar
ea
an
d
ca
n
b
e
ei
th
er
b
i
n
ar
y
,
ter
n
ar
y
a
n
d
s
i
g
m
o
id
.
T
h
e
v
al
u
e
o
f
ea
c
h
co
n
ce
p
t
i
s
ca
lcu
lated
w
it
h
o
r
ig
i
n
al
f
o
r
m
u
la
p
r
o
p
o
s
ed
b
y
Ko
s
k
o
[
2
]
:
O
th
er
alter
n
at
iv
e
s
ar
e
to
tak
e
in
to
ac
co
u
n
t
t
h
e
p
ast
h
is
to
r
y
o
f
co
n
ce
p
ts
an
d
j
o
in
tl
y
p
r
o
p
o
s
ed
th
e
f
o
llo
w
in
g
eq
u
at
io
n
:
W
T
h
e
A
lg
o
r
it
h
m
1
s
h
o
w
s
t
h
e
s
tep
s
to
f
o
llo
w
f
o
r
th
e
ca
lc
u
latio
n
o
f
th
e
n
ex
t
i
n
p
u
t
v
ec
to
r
at
ea
ch
ite
r
atio
n
.
A
l
g
o
r
ith
m
1
:
C
alc
u
lat
io
n
o
f
t
h
e
o
u
tp
u
t
v
ec
to
r
Step
1
:
R
ea
d
th
e
in
p
u
t v
ec
to
r
an
d
w
ei
g
h
t
m
atr
i
x
W
.
Step
2
: Calcu
late
t
h
e
o
u
tp
u
t v
ec
to
r
W
Step
3
: A
p
p
l
y
t
h
e
tr
an
s
f
er
f
u
n
ctio
n
to
th
e
o
u
tp
u
t
v
ec
to
r
Step
4
:
v
er
if
y
th
e
co
n
d
itio
n
s
o
f
ter
m
in
at
io
n
o
f
t
h
e
alg
o
r
it
h
m
2
.
2
.
B
a
s
ics Re
info
rc
e
m
ent
L
ea
rn
ing
(
RL
)
T
h
e
Ma
r
k
o
v
Dec
is
io
n
P
r
o
ce
s
s
es
(
MD
P
)
d
ef
i
n
es
t
h
e
f
o
r
m
al
f
r
a
m
e
w
o
r
k
o
f
r
ei
n
f
o
r
ce
m
en
t
lear
n
in
g
[
1
3
]
.
Mo
r
e
f
o
r
m
all
y
,
an
MD
P
p
r
o
ce
s
s
is
d
ef
in
ed
b
y
:
a.
S,
a
f
in
ite
s
et
o
f
s
ta
tes.
s
Є
S
b.
A
,
a
f
in
i
te
s
et
o
f
ac
tio
n
s
i
n
s
tat
e
s
.
a
Є
A
(
s
)
c.
r
,
a
r
ew
ar
d
f
u
n
ctio
n
.
r
(
s
,
a)
Є
R
d.
P
,
th
e
p
r
o
b
ab
ilit
y
o
f
tr
a
n
s
i
t
io
n
f
r
o
m
o
n
e
s
tate
to
an
o
t
h
er
d
ep
en
d
in
g
o
n
t
h
e
s
elec
t
ed
ac
tio
n
.
P
(
s
'|
s
,
a)
=
P
a
(
s
,
s
'
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
IJ
-
AI
Vo
l.
5
,
No
.
3
,
Sep
tem
b
er
2
0
1
6
:
95
–
104
100
T
h
e
p
r
o
b
lem
is
to
f
in
d
a
n
o
p
ti
m
al
p
o
lic
y
o
f
ac
t
io
n
s
t
h
at
ac
h
ie
v
es
th
e
g
o
al
b
y
m
ax
i
m
izi
n
g
t
h
e
r
e
w
ar
d
s
,
s
tar
ti
n
g
f
r
o
m
an
y
in
itial
s
tate.
A
t
ea
ch
iter
atio
n
,
th
e
ag
e
n
t
b
ein
g
in
t
h
e
s
tate
ch
o
o
s
es
an
ac
tio
n
,
a
cc
o
r
d
in
g
to
th
e
s
e
o
u
tp
u
ts
t
h
e
en
v
ir
o
n
m
e
n
t
s
en
d
s
eit
h
er
a
w
ar
d
o
r
a
p
en
alt
y
to
th
e
a
g
en
t
s
h
o
w
n
b
y
t
h
e
f
o
llo
w
in
g
f
o
r
m
u
la:
r
k
=
h
(
s
k
,
a
k
, s
k+
1
)
.
T
o
f
in
d
th
e
to
tal
co
s
t,
w
h
ic
h
i
s
r
ep
r
esen
ted
b
y
t
h
e
f
o
r
m
u
la
Σ
h
(
s
k
,a
k
,s
k
+
1
)
,
th
e
co
s
t
s
ar
e
a
cc
u
m
u
lated
at
ea
ch
iter
atio
n
o
f
th
e
s
y
s
te
m
.
I
n
[
8
]
th
e
ex
p
ec
ted
r
ew
ar
d
is
w
e
ig
h
ted
b
y
th
e
p
ar
a
m
eter
γ
an
d
b
ec
o
m
es
Σ
γ
h
(s
i
,a
i
,s
i+
1
)
w
it
h
0
≤
γ
≤
1
.
T
h
e
R
L
i
s
to
f
in
d
a
p
o
licy
o
r
an
o
p
ti
m
al
s
tr
ate
g
y
π
*
,
a
m
o
n
g
th
e
d
if
f
er
e
n
t
π
p
o
s
s
ib
le
s
tr
ateg
ie
s
in
t
h
e
s
elec
tio
n
o
f
t
h
e
ac
tio
n
.
C
o
n
s
id
er
in
g
th
at
a
n
o
p
ti
m
al
p
o
lic
y
π
e
x
is
t
s
,
an
d
th
e
n
th
e
B
ell
m
a
n
[
1
1
]
o
p
tim
a
lit
y
eq
u
atio
n
i
s
s
ati
s
f
ie
d
:
V
π
*
= V
*
(s
i
)
=
m
a
x
{
R
(
s
i
,
a)
+
δ
(
∑P
(
s
i
→s
i+
1
,
a)
V
*
(s
i
+
1
)}
∀
s
∈
S
E
q
u
atio
n
(
2
)
s
ets
th
e
v
al
u
e
f
u
n
ctio
n
o
f
t
h
e
o
p
ti
m
a
l
p
o
lic
y
t
h
at
r
ein
f
o
r
ce
m
e
n
t
lear
n
i
n
g
w
i
ll
s
ee
k
to
ass
es
s
:
V
*
(
s
)
=
m
ax
V
π
(
s
)
I
n
Q
-
L
ea
r
n
i
n
g
a
lg
o
r
it
h
m
tech
n
iq
u
e
[
1
3
]
,
th
e
a
g
en
t,
Fo
r
an
y
p
o
licy
π
a
n
d
a
n
y
s
tate
s
∈
S
,
t
h
e
v
a
lu
e
o
f
tak
i
n
g
ac
tio
n
a
i
n
s
ta
te
s
u
n
d
e
r
p
o
licy
π
,
d
en
o
ted
Q
π
(
s
,
a
)
,
is
th
e
e
x
p
ec
ted
d
is
co
u
n
ted
f
u
t
u
r
e
r
e
w
ar
d
s
tar
ti
n
g
i
n
s
,
tak
i
n
g
a
,
an
d
h
e
n
ce
f
o
r
t
h
f
o
l
lo
w
i
n
g
π
.
I
n
th
is
ca
s
e
t
h
e
f
u
n
ctio
n
(
3
)
ca
n
also
b
e
ex
p
r
ess
ed
f
o
r
a
s
tate
-
ac
tio
n
p
air
:
Q
*
(
s
,
a
)
=
m
a
x
Q
π
(
s
,
a
)
Q
-
lear
n
in
g
i
s
o
n
e
o
f
t
h
e
m
o
s
t
p
o
p
u
lar
r
ein
f
o
r
ce
m
e
n
t
lear
n
i
n
g
m
et
h
o
d
s
d
ev
elo
p
ed
b
y
W
atk
in
s
(
1
9
8
9
)
an
d
is
b
a
s
ed
o
n
T
D(
0
)
.
I
t
in
v
o
lv
e
s
f
in
d
i
n
g
s
ta
te
-
ac
t
io
n
q
u
alitie
s
r
at
h
er
th
a
n
j
u
s
t
s
tate
v
alu
e
s
.
Q
-
L
ea
r
n
i
n
g
alg
o
r
ith
m
tech
n
iq
u
e
is
to
in
tr
o
d
u
ce
a
q
u
alit
y
f
u
n
ctio
n
Q
r
ep
r
esen
ts
a
v
al
u
e
f
o
r
ea
ch
s
tate
-
ac
tio
n
p
air
an
d
Q
π
(
s
,
a)
is
to
s
tr
en
g
th
e
n
esti
m
ate
w
h
e
n
s
tar
ti
n
g
f
r
o
m
s
tate
s
,
ex
ec
u
ti
n
g
ac
tio
n
a
b
y
f
o
llo
w
in
g
a
p
o
licy
π
:
Q
π
(
s
,
a)
=
E
Σγ
r
i
a
n
d
Q
*
(
s
,
a)
i
s
t
h
e
o
p
ti
m
al
s
tate
-
ac
tio
n
p
air
b
y
f
o
llo
w
i
n
g
p
o
lic
y
π
*
i
f
Q
*
(
s
,
a)
=
m
a
x
Q
π
(
s
,
a)
an
d
if
w
e
r
ea
ch
th
e
Q
*
(
s
i
,
a
i
)
f
o
r
ea
ch
p
air
s
tate
-
ac
t
io
n
th
e
n
w
e
s
a
y
th
at
t
h
e
ag
e
n
t
ca
n
r
ea
ch
th
e
g
o
al
s
tar
tin
g
f
r
o
m
an
y
i
n
it
ial
s
tate.
T
h
e
v
al
u
e
o
f
Q
is
u
p
d
ated
b
y
t
h
e
f
o
llo
w
in
g
eq
u
atio
n
:
Q
+1
(s
i
,a
i
)
=
Q
(s
i
,a
i
)
+
α
[
h
(
s
i
,a
i
,s
i+
1
)
+
γ
ar
g
m
ax
(
Q
(s
i+
1,
a)
)
–
Q
(s
i
,a
i
)
]
2.
3
.
T
he
a
da
pta
t
io
n o
f
L
ea
r
nin
g
F
uzzy
Co
g
nitiv
e
M
a
ps
T
h
e
r
atio
n
ale
o
f
th
e
p
r
o
p
o
s
ed
im
m
u
n
e
r
esp
o
n
s
e
in
s
p
ir
ed
L
FC
M
is
to
f
o
s
ter
lear
n
in
g
ca
p
ab
ilit
y
an
d
m
em
o
r
y
ac
q
u
is
itio
n
o
f
th
e
L
FC
M.
T
o
s
h
o
w
h
o
w
th
ese
tw
o
is
s
u
es
h
av
e
b
ee
n
ad
d
r
ess
ed
,
th
e
C
o
m
p
lex
ad
ap
tiv
e
ar
t
if
icial
I
m
m
u
n
e
s
y
s
tem
h
as
b
ee
n
co
n
s
id
er
ed
an
d
m
o
d
eled
in
th
e
b
ac
k
g
r
o
u
n
d
o
f
p
r
esen
tin
g
L
FC
M
[
1
5
]
.
I
n
im
m
u
n
e
r
esp
o
n
s
e
th
e
ab
ilit
y
to
m
em
o
r
ize
m
o
s
t
p
r
ev
io
u
s
ly
en
co
u
n
ter
ed
an
tig
en
s
b
y
B
ce
lls
,
en
ab
les
it
to
m
o
u
n
t
a
m
o
r
e
ef
f
ec
tiv
e
r
ea
ctio
n
in
an
y
f
u
tu
r
e
en
co
u
n
t
er
s
.
T
h
is
m
ec
h
an
is
m
in
th
e
n
atu
r
al
im
m
u
n
e
s
y
s
tem
is
u
s
u
ally
d
esig
n
ed
as
th
e
ab
ilit
y
o
f
ad
ap
tiv
e
lear
n
in
g
an
d
im
m
u
n
e
m
em
o
r
y
ac
q
u
is
itio
n
.
T
h
is
is
th
e
b
asis
o
f
m
ath
em
atica
l
ad
ap
tatio
n
o
f
th
e
Q
-
L
ea
r
n
in
g
alg
o
r
ith
m
in
th
e
s
en
s
e
o
f
in
s
tr
u
ctin
g
th
e
ag
en
t
to
co
n
s
id
er
o
p
tim
ally
its
h
is
to
r
y
,
ie
th
e
v
alu
e
o
f
Q
to
aim
to
m
em
o
r
ize
th
e
s
tate
v
is
ited
b
y
th
e
ag
en
t
.
in
o
th
er
s
w
o
r
d
s
,
o
n
ce
th
e
B
ce
ll
id
en
tif
ies
th
e
in
ter
leu
k
in
s
u
b
s
tan
ce
f
r
o
m
th
e
T
h
ce
ll
co
n
ce
p
t,
it
d
iv
id
es
in
to
an
tib
o
d
y
s
y
n
th
etic
ce
lls
,
an
d
f
in
aly
s
ec
r
etes t
h
e
an
tib
o
d
y
(
A
b
)
.
T
h
e
C
ASs
ar
e
d
is
ti
n
g
u
is
h
ed
f
r
o
m
o
t
h
er
s
y
s
te
m
s
b
y
t
h
eir
d
y
n
a
m
ic
i
m
p
r
o
v
e
m
e
n
ts
i
n
c
u
r
r
en
t
p
o
lic
y
f
o
r
ea
ch
in
ter
ac
tio
n
w
ith
t
h
e
e
n
v
ir
o
n
m
e
n
t.
So
th
i
s
is
a
lo
ca
l
co
n
s
tr
u
ct
io
n
th
a
t
d
o
es
n
o
t
r
eq
u
ir
e
an
ass
e
s
s
m
e
n
t
o
f
th
e
o
v
er
all
s
tr
ateg
y
.
T
h
is
o
b
s
er
v
atio
n
lead
s
u
s
to
o
v
er
lo
o
k
t
h
e
v
a
lu
e
o
f
t
h
e
q
u
a
lit
y
f
u
n
c
tio
n
Q
i
n
s
tep
(
i+1
)
.
T
h
is
tr
an
s
late
s
m
ath
e
m
at
icall
y
b
y
:
Q
n
(s
i+
1
,
a)
=
0
an
d
th
er
ef
o
r
e
eq
u
atio
n
(
6
)
o
f
th
e
f
u
n
ctio
n
Q
b
ec
o
m
es
a
s
f
o
llo
w
s
:
Q
(s
i
, a
i
)
= Q
(s
i
,
a
i
)
+ α
[
r
i
-
Q
(s
i
,
a
i
)
]
T
h
e
v
alu
e
o
f
Q
en
ab
le
s
y
s
te
m
to
m
o
u
n
t
a
m
o
r
e
e
f
f
ec
ti
v
e
ac
tio
n
i
n
a
n
y
f
u
t
u
r
e
e
n
co
u
n
t
er
ed
s
tate
alr
ea
d
y
v
is
ited
.
So
th
e
Q
v
al
u
e
is
d
e
s
ig
n
ed
to
in
s
tr
u
ct
t
h
e
a
g
en
t
to
co
n
s
id
er
o
p
ti
m
all
y
i
ts
h
is
to
r
ical
p
as
t.
I
f
th
e
ag
en
t is
i
n
a
s
tate
a
lr
ea
d
y
v
is
i
t
ed
,
w
it
h
a
Q
v
al
u
e
i
n
t
h
e
tab
le
o
f
v
al
u
es,
i
t
w
il
l b
e
d
ir
ec
tl
y
e
x
p
lo
ited
to
m
o
v
e
to
th
e
n
ex
t
s
ta
te,
o
th
er
w
i
s
e
it
w
i
ll
e
x
p
lo
r
e
th
e
p
o
s
s
ib
le
ac
t
io
n
s
i
n
t
h
i
s
s
ta
te
ac
co
r
d
in
g
to
th
eir
r
esp
ec
ti
v
e
p
r
o
b
a
b
ilit
ies T
h
e
f
o
ll
o
w
in
g
p
s
eu
d
o
co
d
e
p
r
o
v
id
es a
n
u
p
d
ate
o
f
th
e
v
alu
e
o
f
Q
f
u
n
ctio
n
:
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
I
SS
N:
2252
-
8938
N
a
tu
r
a
l I
mmu
n
e
S
ystem
R
esp
o
n
s
e
A
s
C
o
mp
lexe
A
d
a
p
tive
S
ystem
Usi
n
g
Lea
r
n
in
g
F
u
z
z
y
…(
A
h
m
e
d
T
l
il
i
)
101
I
f
r
=
1
/ /
Aw
ar
d
Q
(s
i
, a
i
)
= Q
(s
i
,
a
i
)
+ α
[
1
-
Q
(s
i
,a
i
)]
I
f
r
=
0
/ /
P
en
alt
y
Q
(s
i
,
a
i
)
= (
1
-
α
)
Q
(s
i
,
a
i
)
I
n
o
u
r
ap
p
r
o
ac
h
,
if
th
e
s
tate
s
ar
e
r
ep
r
esen
ted
af
ter
f
u
zz
y
f
ic
atio
n
b
y
t
h
e
co
n
ce
p
ts
i
n
p
u
ts
o
r
s
en
s
o
r
y
co
n
ce
p
ts
,
th
e
o
u
tp
u
t
v
ec
to
r
is
r
ep
r
esen
ted
b
y
t
h
e
s
e
t
o
f
o
u
t
p
u
t
co
n
ce
p
ts
o
r
ef
f
ec
to
r
s
co
n
ce
p
ts
th
at
r
ep
r
esen
t
ac
tio
n
s
to
p
er
f
o
r
m
in
t
h
e
en
v
i
r
o
n
m
e
n
t
af
ter
d
ef
u
zz
y
f
icat
io
n
.
T
h
e
m
o
to
r
s
co
n
ce
p
ts
ar
e
th
e
d
ec
is
io
n
-
m
a
k
in
g
m
ec
h
a
n
i
s
m
.
T
h
e
ex
p
lo
r
atio
n
o
f
th
e
ac
tio
n
s
i
s
ac
co
m
p
an
ied
b
y
a
n
u
p
d
ate
o
f
t
h
eir
p
r
o
b
ab
ilit
ies
ac
co
r
d
in
g
to
th
e
li
n
ea
r
s
ch
e
m
e
[
9
]
:
I
f
r
=
1
/ /
Aw
ar
d
P
(s
i
, a
i
)
=
P
(s
i
,
a
i
)
+
β (
1
-
P
(s
i
,
a
i
))
I
f
r
=
0
/ /
P
en
alty
P
(s
i
, a
i
)
=
(
1
-
β
)
P
(s
i
,
a
i
)
2
.
4
O
pera
t
io
na
l
m
ec
ha
ni
s
m
m
o
del
T
h
e
m
ec
h
a
n
is
m
to
id
en
tify
t
h
e
n
at
u
r
e
o
f
t
h
e
a
n
ti
g
en
a
n
d
th
e
s
elec
tin
g
o
f
ac
tio
n
to
c
o
n
s
id
er
is
s
u
m
m
ar
ized
b
y
t
h
e
f
u
zz
y
r
u
l
e
b
ased
s
y
s
te
m
.
A
s
et
o
f
I
F
-
T
HE
N
lin
g
u
is
t
ic
r
u
le
s
,
w
it
h
th
e
i
n
p
u
t
s
a
n
d
th
e
o
u
tp
u
ts
ar
e
co
m
p
o
s
ed
o
f
f
u
zz
y
s
tate
m
e
n
t
s
,
is
t
h
e
ess
e
n
tial p
a
r
t o
f
th
e
f
u
zz
y
r
u
le:
I
F a
s
et
o
f
co
n
d
itio
n
s
ar
e
s
atis
f
ied
T
HE
N
a
s
et
o
f
r
esu
lts
ca
n
b
e
in
f
er
r
ed
I
n
th
i
s
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
,
th
e
w
ei
g
h
t
s
,
w
ij
,
ar
e
d
y
n
a
m
ic
an
d
ca
n
b
e
m
o
d
if
ied
ac
co
r
d
in
g
to
r
ein
f
o
r
ce
m
en
t
lear
n
i
n
g
al
g
o
r
ith
m
to
p
er
m
it
th
e
n
et
w
o
r
k
t
o
b
e
tr
ain
ed
b
y
ex
p
er
ie
n
ce
[
2
0
]
.
B
ased
o
n
th
e
th
eo
r
etica
l
asp
ec
ts
d
escr
ib
e
d
ab
o
v
e,
th
e
p
s
eu
d
o
co
d
e
o
f
A
l
g
o
r
ith
m
2
s
u
m
m
ar
i
ze
s
o
u
r
ap
p
r
o
ac
h
.
A
l
g
o
r
ith
m
2
:
P
s
eu
d
o
co
d
e
o
f
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
Step
1
:
R
ea
d
th
e
v
ec
to
r
k
an
d
w
eig
h
t
m
atr
ix
W
Step
2
:
C
alcu
late
th
e
o
u
tp
u
t
v
ec
to
r
k
:
k
k
k
W
Step
3
:
A
p
p
ly
th
e
tr
an
s
f
er
f
u
n
ctio
n
to
th
e
o
u
tp
u
t v
ec
to
r
k
St
ep
4
:
A
m
o
n
g
th
e
ac
tiv
e
co
n
ce
p
ts
ch
o
o
s
e
th
e
o
n
e
th
at
h
as
th
e
h
ig
h
est
v
alu
e
o
f
th
e
f
u
n
ctio
n
Q
,
if
n
o
t
p
r
o
b
ab
ilit
y
Step
5
:
ca
lcu
late
th
e
n
ew
o
u
tp
u
t
v
ec
to
r
(
o
u
tp
u
t
co
n
ce
p
ts
)
k
Step
6
:
Dep
en
d
in
g
o
n
th
e
r
esp
o
n
s
e
to
th
e
en
v
ir
o
n
m
en
t
:
If
r
=
1
/
/ A
w
ar
d
(
Up
d
atin
g
th
e
p
r
o
b
ab
ilit
y
P
ij
an
d
th
e
Q
v
alu
e
)
Q
(
s
i
, a
i
)
=
Q
(
s
i
,
a
i
)
+ α
[
1
–
Q
(s
i
,
a
i
)]
W
C
i
,C
j
)
=
W
C
i
,C
j
)
P
(a
i
)
= P
(a
i
)
+
β
[
1
-
P
(a
i
)]
If
r
=
o
/ /
P
en
alty
(
Up
d
atin
g
t
h
e
p
r
o
b
ab
ilit
y
P
ij
,
th
e
w
eig
h
t o
f
th
e
co
n
n
ec
tio
n
a
n
d
th
e
v
a
lu
e
of
Q)
Q
k
(s
i
, a
i
)
)
= (
1
-
α
)
Q
k
(s
i
, a
i
)
W
C
i
,C
j
)
=
W
C
i
,C
j
)
+
η [
1
-
W
C
i
,C
j
)]
P
(a
i
)
= (
1
-
β
)
P
(a
i
)
Step
7
:
I
f
th
e
ter
m
in
a
tio
n
co
n
d
itio
n
s
ar
e
r
ea
lized
Sto
p
.
Oth
er
w
i
s
e
g
o
to
Step
2
.
3.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
T
o
ev
alu
ate
th
e
p
er
f
o
r
m
a
n
c
e
o
f
o
u
r
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
,
th
e
s
i
m
u
latio
n
o
f
t
h
e
s
y
s
te
m
w
a
s
i
m
p
le
m
en
ted
i
n
M
A
T
L
AB
,
w
h
ic
h
co
m
p
r
i
s
es
F
u
zz
if
icatio
n
an
d
d
ef
u
zz
i
f
ic
atio
n
w
it
h
F
C
M
m
o
d
eli
n
g
[
2
2
]
.
T
ab
le
2
s
h
o
w
s
w
eig
h
t
v
al
u
es
b
et
w
ee
n
co
n
c
ep
ts
af
ter
d
ef
f
u
zi
f
icatio
n
p
r
o
ce
s
s
i
n
th
e
b
i
m
o
d
al
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
IJ
-
AI
Vo
l.
5
,
No
.
3
,
Sep
tem
b
er
2
0
1
6
:
95
–
104
102
m
o
d
e.
T
h
e
m
ai
n
p
u
r
p
o
s
e
o
f
t
h
e
i
m
m
u
n
e
s
y
s
te
m
is
to
r
ec
o
g
n
ize
all
ce
ll
s
w
i
th
i
n
th
e
b
o
d
y
a
n
d
ca
teg
o
r
ize
t
h
o
s
e
ce
lls
as
s
el
f
o
r
n
o
n
-
s
el
f
.
A
cti
v
atio
n
o
f
T
ce
lls
b
y
a
n
ti
g
en
-
p
r
esen
ti
n
g
ce
ll
s
(
A
P
C
s
)
,
w
it
h
t
h
e
ac
ce
s
s
o
r
y
co
n
ce
p
t
MH
C
,
i
n
l
y
m
p
h
n
o
d
es
is
a
k
e
y
i
n
itia
tin
g
e
v
e
n
t
in
n
at
u
r
al
i
m
m
u
n
e
r
esp
o
n
s
e
s
.
I
n
t
h
is
ca
s
e
t
h
e
T
h
co
n
ce
p
t
(
T
h
ce
lls
)
is
co
n
s
id
er
ed
as
t
h
e
m
o
to
r
co
n
ce
p
t
a
n
d
t
h
e
a
ll
o
th
er
s
co
n
ce
p
ts
ar
e
co
n
s
i
d
er
ed
as
ac
ce
s
s
o
r
y
co
n
ce
p
ts
.
T
ce
lls
ar
e
ab
le
alo
n
e
to
d
if
f
er
en
tiate
b
et
w
ee
n
s
el
f
an
d
n
o
s
e
lf
(
an
ti
g
e
n
s
)
ce
ll
s
.
T
ce
ll
r
ec
ep
to
r
s
s
ca
n
th
e
s
u
r
f
ac
e
o
f
A
P
C
f
o
r
s
p
ec
if
i
c
p
ep
tid
es
b
o
u
n
d
to
m
o
lecu
le
s
o
f
t
h
e
MH
C
.
I
f
t
h
e
s
p
ec
i
f
ic
p
ep
tid
es
ar
e
f
o
u
n
d
,
th
e
T
h
ce
l
l
i
s
ac
ti
v
ated
,
s
o
th
e
n
o
-
s
el
f
ac
t
io
n
i
s
e
x
ec
u
ted
a
n
d
t
h
e
a
n
tib
o
d
y
w
i
ll
b
e
s
ec
r
eted
,
o
th
er
w
is
e
th
e
an
ti
g
en
i
s
r
ec
o
g
n
ized
as
s
el
f
ce
ll
th
e
n
th
e
s
elf
ac
t
io
n
is
e
x
ec
u
ted
an
d
th
e
i
m
m
u
n
e
r
esp
o
n
s
e
is
ter
m
i
n
ated
.
T
ab
le
1
.
C
o
n
ce
p
t D
escr
ip
tio
n
o
f
T
h
e
C
AA
I
S i
n
L
F
C
M
B
ac
k
g
r
o
u
n
d
C
o
n
c
e
p
t
s
D
e
scri
p
t
i
o
n
Ag
A
n
t
i
g
e
n
s (v
i
r
u
s
a
n
d
b
a
c
t
e
r
i
a
)
A
P
C
A
n
t
i
g
e
n
P
r
e
se
n
t
i
n
g
C
e
l
l
s
M
H
C
M
a
j
o
r
H
i
st
o
c
o
m
p
a
t
i
b
i
l
i
t
y
C
o
mp
l
e
x
mo
l
e
c
u
l
e
Th
T
h
e
H
e
l
p
e
r
T
c
e
l
l
I
L
+
T
h
e
i
n
t
e
r
l
e
u
k
i
n
p
o
si
t
i
v
e
si
g
n
a
l
se
c
r
e
t
e
d
b
y
T
h
c
e
l
l
B
B
C
e
l
l
Ab
A
n
t
i
b
o
d
y
p
r
o
d
u
c
e
d
b
y
B
c
e
l
l
s
Ts
T
h
e
su
p
p
r
e
sso
r
T
c
e
l
l
IL
-
T
h
e
i
n
t
e
r
l
e
u
k
i
n
n
e
g
a
t
i
v
e
si
g
n
a
l
se
c
r
e
t
e
d
b
y
T
s c
e
l
l
T
h
e
n
u
m
b
er
o
f
co
n
ce
p
ts
h
as
b
ee
n
r
ed
u
ce
d
to
9
co
n
ce
p
ts
th
u
s
to
av
o
id
th
e
co
m
p
lex
i
t
y
o
f
t
h
e
C
AA
I
S
m
o
d
eled
in
t
h
is
L
F
C
M
t
y
p
e
a
n
d
f
o
r
th
e
p
r
o
p
o
s
ed
tech
n
iq
u
e
to
b
e
m
o
r
e
clea
r
to
n
o
s
p
ec
ia
lis
t
r
ea
d
er
s
w
e
u
s
e
f
u
zz
y
f
ied
b
in
ar
y
m
o
d
e.
C
o
n
ce
p
ts
Ag
a
n
d
A
b
ar
e
th
e
F
ac
to
r
-
co
n
ce
p
ts
(
s
e
n
s
o
r
y
co
n
c
ep
ts
an
d
ef
f
ec
to
r
s
co
n
ce
p
ts
r
esp
ec
tiv
el
y
)
,
w
h
ic
h
r
ep
r
esen
t
th
e
in
p
u
t
an
d
o
u
tp
u
t
co
n
ce
p
t
(
in
ter
m
o
f
i
n
ter
ac
tio
n
w
it
h
th
e
en
v
ir
o
n
m
e
n
t)
.
T
ab
le
2
.
W
eig
h
t
Valu
es B
et
wee
n
C
o
n
ce
p
t
s
in
t
h
e
B
i
m
o
d
al
Mo
d
e
C
o
n
c
e
p
t
s
Ag
A
P
C
M
H
C
Th
I
L
+
B
Ab
Ts
IL
-
Ag
0
+1
0
0
0
+1
0
0
0
A
P
C
0
0
+1
0
0
0
0
0
0
M
H
C
0
0
0
+1
0
0
0
0
0
Th
-
1
0
0
0
+1
0
0
0
0
I
L
+
0
0
0
0
0
+1
0
0
0
B
0
0
0
0
0
0
+1
0
0
Ab
-
1
0
0
0
0
0
0
+1
0
Ts
0
0
0
0
0
0
0
0
+1
IL
-
0
0
0
-
1
0
0
0
0
0
T
h
e
W
m
a
tr
ix
li
n
k
ass
o
ciate
d
t
o
th
is
m
o
d
el
ca
n
b
e
w
r
itte
n
as
f
o
llo
w
s
:
T
h
e
FC
M
(
Fi
g
u
r
e
2
)
h
a
s
t
welv
e
ed
g
e
s
an
d
n
i
n
e
co
n
ce
p
t
s
w
it
h
li
n
k
s
e
x
citato
r
y
(
+1
)
o
f
'A
g
'
to
'A
P
C
'
,
'A
P
C
'
to
'
MH
C
'
,
'
MH
C
'
to
'
T
h
'
,
'
T
h
'
to
'
I
L
+
'
,
'
I
L
+
'
to
'
B
'
,
'
B
'
to
'A
b
'
,
'A
b
'
to
'
T
s
'
a
n
d
'
T
s
'
to
'
I
L
-
'
,
a
n
d
lin
k
ed
in
h
ib
ito
r
(
-
1
)
o
f
'A
b
'
to
'Ag
'
,
'
I
L
-
'
to
'
T
h
'
a
n
d
'
T
h
'
to
'Ag
'
.
T
h
e
co
n
ce
p
t
is
ac
tiv
e
if
its
v
a
lu
e
is
eq
u
al
to
1
,
o
th
er
w
i
s
e
it
is
in
ac
ti
v
e
(
b
in
ar
y
m
o
d
e)
.
I
t
is
g
iv
e
n
an
in
itial
ac
t
iv
atio
n
v
ec
to
r
A
=
(
1
0
0
0
0
0
0
0
0
)
.
T
ab
le
I
s
h
o
w
‟
s
t
h
e
v
a
lu
e
s
P
(
a
i
)
o
f
th
e
p
r
o
b
ab
ilit
ies
o
f
ac
tio
n
s
an
d
v
alu
e
s
o
f
t
h
e
f
u
n
ct
io
n
Q
u
p
d
ated
at
ea
ch
iter
atio
n
.
T
a
b
le
I
I
g
iv
es
th
e
o
u
tp
u
t
v
ec
to
r
f
o
r
all
iter
atio
n
s
i
n
r
esp
o
n
s
e
to
th
e
en
v
ir
o
n
m
e
n
t.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
I
SS
N:
2252
-
8938
N
a
tu
r
a
l I
mmu
n
e
S
ystem
R
esp
o
n
s
e
A
s
C
o
mp
lexe
A
d
a
p
tive
S
ystem
Usi
n
g
Lea
r
n
in
g
F
u
z
z
y
…(
A
h
m
e
d
T
l
il
i
)
103
T
ab
le
3
.
A
ctio
n
P
r
o
b
ab
ilit
ies an
d
Q
-
F
u
n
ctio
n
Va
lu
e
s
a
i
P
(
a
i
)
Q(
s
i
, a
i
)
Valu
e
se
l
f
no
-
se
l
f
0
.
5
0
.
5
(Ag
i
,
se
l
f
)
(Ag
i
,
n
o
-
se
l
f
)
0
0
T
ab
le
4
.
Vec
to
r
Ou
tp
u
t a
t E
ac
h
I
ter
atio
n
s
I
n
p
u
t
s
v
e
c
t
o
r
O
u
t
p
u
t
v
e
c
t
o
r
I
t
e
r
a
t
i
o
n
(
1
0
0
0
0
0
0
0
0
)
1
1
0
0
0
0
0
0
0
1
1
1
1
0
0
0
0
0
0
2
1
1
1
1
0
0
0
0
0
3
1
1
1
1
1
0
0
0
0
4
A
t
iter
atio
n
n
°
4
th
e
i
m
m
u
n
e
s
y
s
te
m
i
s
f
ac
i
n
g
a
s
itu
a
tio
n
wh
er
e
it
h
a
s
t
w
o
p
o
s
s
ib
le
ac
tio
n
s
,
if
t
h
e
s
tate
is
n
o
t
e
n
co
u
n
tr
ed
,
r
ep
r
esen
ted
b
y
t
h
e
ac
ti
v
e
co
n
ce
p
ts
T
h
,
th
e
ex
ec
u
tio
n
o
f
t
h
e
ac
tio
n
SE
L
F
lead
s
to
th
e
d
ea
ctiv
atio
n
o
f
Ag
co
n
ce
p
t
a
n
d
ex
ec
u
tio
n
o
f
t
h
e
ac
tio
n
N
O
-
SE
L
F
lead
s
to
th
e
ac
t
iv
at
i
o
n
o
f
t
h
e
B
co
n
ce
p
t
(
th
r
o
u
g
h
I
L
+
co
n
ce
p
t)
ai
m
s
to
n
eu
tr
alize
th
e
an
tig
e
n
b
u
t
m
u
s
t
ch
o
o
s
e
o
n
e
a
m
o
n
g
t
h
e
m
a
n
d
th
i
s
ch
o
ice
is
gui
d
ed
ei
th
er
b
y
t
h
e
v
al
u
e
o
f
f
u
n
ctio
n
Q,
if
t
h
e
s
tate
i
s
alr
e
ad
y
v
is
ited
o
r
a
n
ti
g
en
alr
ea
d
y
en
co
u
n
ter
ed
,
o
r
b
y
th
e
v
al
u
e
o
f
th
e
p
r
o
b
ab
ilit
y
o
f
th
e
ac
tio
n
i
f
t
h
e
f
ir
s
t p
ass
i
n
t
h
is
s
ta
te.
T
h
is
m
ec
h
an
is
m
to
id
en
ti
f
y
t
h
e
n
a
tu
r
e
o
f
th
e
an
ti
g
e
n
an
d
th
e
s
elec
tin
g
ac
tio
n
to
co
n
s
id
er
is
s
u
m
m
ar
ized
b
y
th
e
f
o
llo
w
i
n
g
th
r
e
e
f
u
zz
y
r
u
le
b
ased
s
y
s
te
m
:
R
1
:
if A
g
is
Q(
A
g
,
S
E
LF)
th
en
Th
is
S
E
LF
-
a
ctio
n
R
2
:
if A
g
is
Q(
A
g
,
N
O
-
S
E
LF)
th
en
Th
is
N
O
-
S
E
LF
-
a
ctio
n
.
R
3
:
if
A
g
is
n
o
t
Q(
A
g
,
S
E
LF)
a
n
d
A
g
is
n
o
t
Q(
A
g
,
n
o
-
s
elf)
th
en
Th
is
a
ctio
n
to
p
erfo
r
m
s
elec
ted
a
cc
o
r
d
in
g
to
th
e
p
r
o
b
a
b
ilit
y.
I
n
th
is
s
y
s
tem
b
ased
o
n
r
u
les
th
e
co
n
d
itio
n
s
o
f
th
e
f
ir
s
t
tw
o
r
u
les
R
1
an
d
R
2
r
esu
lt
th
at
s
y
s
tem
h
as
m
et
th
e
an
tig
en
b
ef
o
r
e
an
d
class
if
y
it
as
a
p
ar
t
o
f
th
e
s
elf
o
r
n
o
-
s
elf
b
y
th
e
u
p
d
ate
tab
le
o
f
Q
v
alu
es,
f
o
r
m
o
r
eo
v
er
th
e
R
3
r
u
le
r
eq
u
ir
es
th
e
T
h
co
n
ce
p
t
to
ex
p
lo
it
s
p
ac
e
o
f
p
o
s
s
ib
le
ac
tio
n
s
ac
co
r
d
in
g
to
th
eir
r
esp
ec
tiv
e
p
r
o
b
ab
ilit
ies.
4.
CO
NCLU
SI
O
N
T
h
e
s
o
f
t
co
m
p
u
t
in
g
tec
h
n
iq
u
e
o
f
f
u
zz
y
co
g
n
iti
v
e
m
ap
s
f
o
r
m
o
d
elin
g
an
d
s
i
m
u
lat
in
g
co
m
p
le
x
ad
ap
tiv
e
s
y
s
te
m
h
a
s
b
ee
n
d
is
cu
s
s
ed
in
th
is
p
ap
er
.
A
n
e
w
co
n
n
ec
t
io
n
b
et
w
ee
n
f
u
zz
y
s
y
s
te
m
a
n
d
r
ein
f
o
r
ce
m
en
t
lear
n
in
g
h
as
b
ee
n
p
r
o
p
o
s
ed
f
o
r
an
al
y
zi
n
g
n
a
tu
r
al
i
m
m
u
n
e
s
y
s
te
m
r
esp
o
n
s
e.
I
n
t
h
e
ar
ti
f
ici
al
in
tell
ig
e
n
ce
f
ield
th
e
n
atu
r
al
i
m
m
u
n
e
s
y
s
te
m
N
I
S
is
ar
g
u
ed
t
h
at
it
is
a
co
m
p
l
ex
ad
ap
tiv
e
s
y
s
te
m
.
Glo
b
al
e
m
er
g
e
n
t
b
e
h
av
io
r
s
ca
n
b
e
o
b
s
er
v
ed
b
y
ap
p
ly
i
n
g
lo
ca
l
r
u
les
to
in
d
iv
id
u
al
ce
ll
s
as
d
escr
ib
ed
b
y
Ho
llan
d
in
co
m
p
le
x
ad
ap
tiv
e
s
y
s
te
m
t
h
eo
r
y
.
T
h
e
co
m
p
le
x
i
t
y
a
n
d
cr
itici
s
m
r
aised
b
y
th
e
co
m
m
u
n
it
y
i
n
th
e
ar
ea
o
f
m
o
d
eli
n
g
C
A
S
s
b
y
A
B
M
s
a
n
d
C
A
s
,
led
u
s
to
s
ee
k
an
o
t
h
er
ap
p
r
o
ac
h
,
w
h
ic
h
is
c
o
n
tain
ed
i
n
s
a
m
e
co
n
ce
p
ts
in
s
p
ir
ed
b
y
th
e
ar
ea
o
f
lif
e.
I
n
p
s
y
c
h
o
lo
g
y
b
e
h
av
io
r
is
g
e
n
er
all
y
r
elate
d
to
th
e
co
n
ce
p
ts
o
f
e
m
o
tio
n
s
,
p
er
ce
p
tio
n
s
an
d
s
e
n
s
atio
n
s
.
T
h
ese
k
e
y
co
n
ce
p
ts
o
f
lif
e
ca
n
b
e
s
u
p
p
o
r
ted
b
y
FC
M
s
.
C
ASs
ar
e
th
er
ef
o
r
e
in
th
e
f
ield
o
f
ar
tif
icial
lif
e
m
o
r
e
th
an
o
th
er
ar
ea
s
o
f
co
m
p
u
tin
g
.
T
h
e
ar
ea
o
f
F
C
M
s
,
d
esp
ite
t
h
e
i
m
p
r
o
v
e
m
e
n
t
m
ad
e
b
y
d
if
f
er
en
t
r
esear
c
h
tea
m
s
in
th
e
w
o
r
ld
,
r
e
m
ai
n
s
a
n
ar
ea
d
en
s
e,
lo
w
-
u
n
if
ied
.
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
is
p
ap
er
i
s
t
h
e
r
es
u
lt
o
f
a
r
esear
ch
p
r
o
j
ec
t
en
titl
ed
"
Mo
d
elin
g
a
n
d
i
m
p
le
m
en
t
in
g
o
f
co
m
p
le
x
ad
ap
tiv
e
s
y
s
te
m
s
"
w
h
ich
w
as
s
u
p
p
o
r
ted
b
y
th
e
MI
SC
lab
o
r
ato
r
y
o
f
ab
d
elh
a
m
id
m
eh
r
i
Un
i
v
er
s
it
y
,
A
l
g
er
ia.
W
e
th
an
k
all
th
e
lab
o
r
ato
r
y
te
a
m
s
a
n
d
r
esp
o
n
s
ib
le
f
o
r
t
h
eir
co
o
p
er
atio
n
an
d
s
u
p
p
o
r
ts
.
RE
F
E
R
E
NC
E
S
[1
]
Ax
elr
o
d
R
o
b
er
t (
1
9
7
6
)
.
'
Stru
c
t
u
r
e
o
f
d
ec
is
io
n
'
.
P
r
in
ce
to
n
u
n
i
v
er
s
it
y
p
r
ess
,
P
r
in
ce
to
n
,
Ne
w
J
er
s
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
IJ
-
AI
Vo
l.
5
,
No
.
3
,
Sep
tem
b
er
2
0
1
6
:
95
–
104
104
[2
]
Ko
s
k
o
B
.
,
Fu
zz
y
C
o
g
n
iti
v
e
M
ap
s
,
I
n
tern
a
tio
n
a
l J
o
u
r
n
a
l Ma
n
-
Ma
ch
in
e
S
tu
d
ies
,
2
4
:
6
5
-
7
5
,
1
9
8
6
[3
]
Ma
ik
el
L
eó
n
1
,
C
ir
o
R
o
d
r
ig
u
ez
1
,
Ma
r
ía
M.
Gar
cía1
,
R
af
ae
l
B
ello
1
,
an
d
Ko
en
Van
h
o
o
f
„
Fu
zz
y
C
o
g
n
iti
v
e
Ma
p
s
Fo
r
Mo
d
elin
g
C
o
m
p
lex
S
y
s
te
m
s
‟
.
© Sp
r
in
g
er
-
Ver
lag
2
0
1
0
.
[4
]
C
h
r
y
s
o
s
to
m
o
s
D.
Sty
l
io
s
an
d
P
eter
P.Gr
o
u
m
p
o
s
.
„
Mo
d
elin
g
C
o
m
p
lex
S
y
s
te
m
s
Usi
n
g
Fu
zz
y
C
o
g
n
iti
v
e
Ma
p
s
'
.
I
E
E
E
tr
a
n
s
a
ctio
n
s
o
n
s
y
s
tems,
Ma
n
a
n
d
cy
b
ern
etics
.
J
an
u
ar
y
2
0
0
4
.
[5
]
T
ah
ar
Gu
er
r
a
m
1
,
R
a
m
d
an
e
Ma
a
m
r
i2
a
n
d
Z
a
id
i
Sa
h
n
o
u
n
2
'A
T
o
o
l
f
o
r
Qu
alitat
iv
e
C
a
u
s
a
l
R
ea
s
o
n
i
n
g
O
n
C
o
m
p
le
x
S
y
s
te
m
s
'
.
I
JCS
I
I
n
te
r
n
a
tio
n
a
l
Jo
u
r
n
a
l
o
f
C
o
mp
u
te
r
S
cien
ce
I
s
s
u
es
,
Vo
l.
7
,
I
s
s
u
e
6
,
No
v
em
b
er
2
0
1
0
I
SS
N
(
On
li
n
e)
:
1694
-
0
8
1
4
.
[6
]
C
.
B
u
c
h
e,
P
.
C
h
ev
a
illi
er
,
A
.
N
´
ed
´
elec
,
M.
P
ar
en
t
h
o
¨
en
an
d
J
.
T
is
s
ea
u
„
Fu
zz
y
co
g
n
iti
v
e
m
ap
s
f
o
r
t
h
e
s
i
m
u
lat
io
n
o
f
i
n
d
iv
id
u
al
ad
ap
tiv
e
b
eh
a
v
io
r
s
‟
W
ile
y
On
lin
e
L
ib
r
ar
y
2
0
1
0
[7
]
E
lp
in
ik
i P
ap
ag
eo
r
g
io
u
an
d
P
et
er
Gr
o
u
m
p
o
s
.
„
A
W
eig
h
t
A
d
a
p
tatio
n
Me
t
h
o
d
f
o
r
Fu
zz
y
C
o
g
n
iti
v
e
Ma
p
s
to
a
P
r
o
ce
s
s
C
o
n
tr
o
l
P
r
o
b
lem
‟
.
Sp
r
in
g
er
2
0
0
4
[8
]
R
ich
ar
d
S.
S
u
tto
n
a
n
d
A
n
d
r
e
w
G.
B
ar
to
.
„
R
ei
n
f
o
r
ce
m
e
n
t
L
ea
r
n
i
n
g
:
An
I
n
tr
o
d
u
ctio
n
‟
.
A
B
r
ad
f
o
r
d
B
o
o
k
T
h
e
MI
T
P
r
ess
C
a
m
b
r
id
g
e,
M
ass
ac
h
u
s
etts
L
o
n
d
o
n
,
E
n
g
lan
d
2
0
0
5
[9
]
“
T
SP
Ho
m
e
p
ag
e”
,
h
ttp
://co
m
o
p
t.i
f
i.
u
n
i
-
h
eid
elb
er
g
.
d
/s
o
f
t
w
ar
e/T
SP
L
I
B
/in
d
ex
.
h
t
m
l
[1
0
]
E
.
T
o
lm
a
n
.
C
o
g
n
it
iv
e
m
ap
s
i
n
r
ats an
d
m
e
n
.
P
s
y
ch
o
lo
g
ical
R
ev
ie
w
v
o
l
u
m
e5
5
,
1
9
4
8
.
[1
1
]
T
h
o
m
as J
.
Sar
g
e
n
t,
D
y
n
a
m
ic
Ma
cr
o
ec
o
n
o
m
ic
T
h
eo
r
y
,
Secti
o
n
1
.
1
-
1
.
4
.
[1
2
]
W
ei
W
an
g
,
Sh
a
n
g
ce
Gao
,
Z
h
en
g
T
an
g
.
'A
C
o
mp
lex
a
r
tifi
ciel
im
mu
n
e
s
ystem'
Fo
u
r
th
I
n
ter
n
at
io
n
a
l
C
o
n
f
er
en
ce
o
n
Nat
u
r
al
C
o
m
p
u
tatio
n
.
2
0
0
8
I
E
E
E
.
[1
3
]
E
.
A
.
J
as
m
i
n
.
T
.
P
.
I
m
t
h
ia
s
A
h
a
m
ed
.
V.
P
.
J
ag
ath
y
R
aj
.
„
R
e
in
f
o
r
ce
m
e
n
t
L
ea
r
n
i
n
g
ap
p
r
o
ac
h
es
to
E
co
n
o
m
ic
Dis
p
atch
p
r
o
b
le
m
‟
.
E
ls
ev
ier
_
2
0
1
1
.
[1
4
]
E
v
a
Vo
ln
a1
'
Ne
u
r
o
ev
o
lu
t
io
n
a
r
y
o
p
ti
m
izatio
n
'
.
I
JCS
I
I
n
tern
a
tio
n
a
l
Jo
u
r
n
a
l
o
f
C
o
mp
u
ter
S
cien
ce
I
s
s
u
es
,
Vo
l.
7
,
I
s
s
u
e
2
,
No
4
,
Ma
r
ch
2
0
1
0
I
SS
N
(
On
lin
e)
: 1
6
9
4
-
0
7
8
4
.
[1
5
]
Ah
m
ed
T
lili
,
Salim
c
h
i
k
h
i
'M
o
d
elin
g
C
o
m
p
lex
A
d
ap
tiv
e
s
S
y
s
te
m
s
u
s
i
n
g
L
ea
r
n
i
n
g
F
u
zz
y
C
o
g
n
iti
v
e
Ma
p
s
'
.
I
n
tern
a
n
tio
n
a
l J
o
u
r
n
a
l
o
f Co
mp
u
ter A
p
p
lica
tio
n
I
JCA
.
Vo
lu
m
e
5
3
No
3
2
0
1
2
.
[1
6
]
T
h
o
r
n
e,
B
.
C
.
,
A
.
M.
B
aile
y
,
D.
W
.
DeSi
m
o
n
e,
a
n
d
S.
M.
P
eir
ce
.
A
g
e
n
t
-
b
a
s
ed
m
o
d
elin
g
o
f
m
u
lticell
m
o
r
p
h
o
g
e
n
ic
p
r
o
ce
s
s
es d
u
r
i
n
g
d
ev
elo
p
m
e
n
t.
B
ir
th
De
f
ec
t
s
R
es.
C
E
m
b
r
y
o
T
o
d
ay
8
1
:3
4
4
–
3
5
3
,
2
0
0
7
.
[1
7
]
C
h
a
v
ali,
A
.
K.
,
E
.
P
.
Gian
ch
an
d
a
n
i,
K.
S.
T
u
n
g
,
M.
B
.
L
a
w
r
e
n
ce
,
S.
M.
P
eir
ce
,
an
d
J
.
A
.
P
ap
in
.
C
h
ar
ac
ter
izi
n
g
e
m
er
g
e
n
t
p
r
o
p
er
ties
o
f
i
m
m
u
n
o
lo
g
ical
s
y
s
te
m
s
w
it
h
m
u
lticel
lu
l
ar
r
u
le
-
b
a
s
ed
co
m
p
u
tatio
n
al
m
o
d
eli
n
g
.
T
r
en
d
s
[1
8
]
I
m
m
u
n
o
l.
2
9
:5
8
9
–
5
9
9
,
2
0
0
8
.
[1
9
]
H.
A
t
lan
,
“
T
h
e
li
v
i
n
g
ce
ll
as
a
p
ar
ad
i
g
m
o
f
co
m
p
lex
n
at
u
r
a
l
s
y
s
te
m
s
,
”
C
o
m
p
le
x
u
s
,
v
o
l.
1
8
,
n
o
.
6
–
7
,
p
p
.
764
–
7
6
6
,
2
0
0
3
.
[2
0
]
Min
a
k
s
h
i
S
h
ar
m
a
,
Dr
.
So
u
r
ab
h
Mu
k
h
er
j
ee
.
"
Fu
zz
y
C
-
Me
an
s
,
A
N
FIS
a
n
d
Gen
etic
A
l
g
o
r
ith
m
f
o
r
Seg
m
en
tin
g
Ast
r
o
c
y
to
m
a
–
A
T
y
p
e
o
f
B
r
ain
T
u
m
o
r
".
I
A
E
S
I
n
tern
a
tio
n
a
l
Jo
u
r
n
a
l
o
f
A
r
tifi
c
ia
l
I
n
tellig
en
ce
(
I
J
-
AI)
Vo
l.
3
,
No
.
1
,
Ma
r
ch
2
0
1
4
,
p
p
.
1
6
~2
3
.
[2
1
]
HOP
FIE
L
D,
J
.
J
.
&
T
ANK,
D
.
W
.
(
1
9
8
6
)
.
C
o
m
p
u
ti
n
g
w
it
h
n
eu
r
al
n
et
w
o
r
k
s
:
a
m
o
d
el.
Sci
en
ce
2
3
3
,
6
2
5
-
633.
[2
2
]
J
ea
n
Ho
llan
d
.
'
St
u
d
y
in
g
C
o
m
p
lex
A
d
ap
tiv
es S
y
s
te
m
s
'
.
2
0
0
6
Sp
r
in
g
er
Scie
n
ce
+
B
u
s
i
n
es
s
Me
d
ia,
I
n
c.
[2
3
]
B
ea
le,
M.
an
d
H.
De
m
u
t
h
,
1
9
9
4
.
Fu
zz
y
s
y
s
te
m
s
2
7
.
A
l
-
E
m
r
an
A
.
,
P
.
Kap
u
r
D.
P
f
a
h
l
a
n
d
G.
R
u
h
e,
2
0
1
0
.
to
o
lb
o
x
:
Fo
r
u
s
e
w
i
th
M
A
T
L
A
B
P
W
S
P
u
b
lis
h
i
n
g
S
tu
d
y
i
n
g
t
h
e
i
m
p
ac
t
o
f
u
n
ce
r
tai
n
t
y
in
o
p
er
atio
n
al
C
o
m
p
an
y
.
B
I
B
L
I
O
G
R
AP
H
Y
O
F
AUT
H
O
RS
A
h
m
e
d
T
li
li
w
a
s
b
o
rn
in
Co
n
st
a
n
ti
n
e
,
A
lg
e
ria
o
n
M
a
r
2
1
,
1
9
7
7
.
He
re
c
e
i
v
e
d
th
e
M
a
g
ister
d
e
g
re
e
s
f
ro
m
El
h
a
d
j
L
a
k
h
d
a
r
U
n
iv
e
rsity
o
f
Ba
tn
a
,
A
lg
e
ri
a
,
in
2
0
0
7
.
I
n
2
0
1
0
,
h
e
jo
i
n
e
d
th
e
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
S
c
ie
n
c
e
a
n
d
it
s
A
p
p
li
c
a
ti
o
n
s,
Ne
w
s
F
a
c
u
lt
y
o
f
In
f
o
r
m
a
ti
o
n
T
e
c
h
n
o
lo
g
ies
a
n
d
Co
m
m
u
n
ica
ti
o
n
A
b
d
e
l
h
a
m
id
M
e
h
ri
U
n
iv
e
rsity
,
Co
n
sta
n
ti
n
e
a
s
a
stu
d
e
n
t
re
se
a
rc
h
e
r.
His
in
tere
stin
g
a
r
e
a
s
a
re
A
rti
f
i
c
ial
In
telli
g
e
n
c
e
,
s
o
f
twa
re
e
n
g
in
e
e
rin
g
,
Ne
u
ra
l
n
e
tw
o
rk
,
F
u
z
z
y
lo
g
ic an
d
F
u
z
z
y
I
n
f
e
re
n
c
e
S
y
ste
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.