UPPER SEMICONTINUITY OF THE ATTRACTOR FOR

LATTICE DYNAMICAL SYSTEMS OF PARTLY

DISSIPATIVE REACTION-DIFFUSION SYSTEMS
AHMED Y. ABDALLAH
Received 12 July 2004 and in revised form 7 December 2004
We investigate the existence of a global attractor and its upper semicontinuity for the
innite-dimensional lattice dynamical system of a partly dissipative reaction-diusion
system in the Hilbert space l
2
l
2
. Such a system is similar to the discretized FitzHugh-
Nagumo system in neurobiology, which is an adequate justication for its study.
1. Introduction
Lattice dynamical systems arise in various application elds, for instance, in chemical
reaction theory, material science, biology, laser systems, image processing and pattern
recognition, and electrical engineering (cf. [6, 7, 13]). In each eld, they have their own
forms, but in some other cases, they appear as spatial discretizations of partial dierential
equations (PDEs). Recently, many authors studied various properties of the solutions for
several lattice dynamical systems. For instance, the chaotic properties have been investi-
gated in [1, 7, 8, 10, 11, 17], and the travelling solutions have been carefully studied in
[2, 3, 7, 8, 9, 22].
From [18], we know that it is dicult to estimate the attractor of the solution semi-
ow generated by the initial value problem of dissipative PDEs on unbounded domains
because, in general, it is innite dimensional. Therefore, it is signicant to study the lat-
tice dynamical systems corresponding to the initial value problemof PDEs on unbounded
domains because of the importance of such systems and they can be regarded as an ap-
proximation to the corresponding continuous PDEs if they arise as spatial discretizations
of PDEs.
The main idea of this work is originated from [4, 16]. In [4, 19, 20], the researchers
proved the existence of global attractors for dierent lattice dynamical systems and they
investigated the nite-dimensional approximations of these global attractors. In fact,
Bates et al. [4] studied rst-order lattice dynamical systems, and Zhou [20] gave a gen-
eralization of the result given by [4]. Zhou [19] studied a second-order lattice dynamical
system and investigated the upper semicontinuity of the global attractor.
Copyright 2005 Hindawi Publishing Corporation
Journal of Applied Mathematics 2005:3 (2005) 273288
DOI: 10.1155/JAM.2005.273
274 Attractor for LDSs of PDRDSs
For a positive integer r, consider the Hilbert space
l
2
=
_
u =
_
u
i
_
iZ
r
: i =
_
i
1
, i
2
, . . . , i
r
_
Z
r
, u
i
R,
_
iZ
r
u
2
i
<
_
(1.1)
whose inner product and norm are given by, for all u =(u
i
)
iZ
r , v =(v
i
)
iZ
r l
2
,
u, v) =
_
iZ
r
u
i
v
i
, |u| =u, u)
1/2
=
_
_
iZ
r
u
2
i
_
1/2
. (1.2)
We will study the following lattice dynamical system of a partly dissipative reaction-
diusion system:
u
i
+(Au)
i
+ f
1
_
u
i
_
+g
1
_
u
i
_
+h
1
_
u
i
, v
i
_
=q
1,i
,
v
i
+ f
2
_
v
i
_
+g
2
_
v
i
_
h
2
_
u
i
, v
i
_
=q
2,i
,
(1.3)
i =(i
1
, i
2
, . . . , i
r
) Z
r
, t > 0, with the initial conditions
u
i
(0) =u
i,0
, v
i
(0) =v
i,0
, (1.4)
where is a positive constant, and are real constants such that
> 0, (1.5)
the operator A : l
2
l
2
is dened by, for all u =(u
i
)
iZ
r l
2
, i =(i
1
, i
2
, . . . , i
r
),
(Au)
i
=2ru
(i
1
,i
2
,...,i
r
)
u
(i
1
1,i
2
,...,i
r
)
u
(i
1
,i
2
1,...,i
r
)
u
(i
1
,i
2
,...,i
r
1)
u
(i
1
+1,i
2
,...,i
r
)
u
(i
1
,i
2
+1,...,i
r
)
u
(i
1
,i
2
,...,i
r
+1)
,
(1.6)
and for j =1, 2, s, s
1
, s
2
R,
q
j
=
_
q
j,i
_
iZ
r
l
2
, (1.7)
f
j
(s), g
j
(s) C
1
(R, R), f
j
(0) =g
j
(0) =0, (1.8)
h
j
_
s
1
, s
2
_
C
1
_
R
2
, R
_
, h
j
(0, 0) =0, (1.9)

j
f

j
(s), (1.10)
g
j
_
s
_
s 0, (1.11)
h
1
_
s
1
, s
2
_
s
1
=h
2
_
s
1
, s
2
_
s
2
, (1.12)
where
j
is a positive constant.
Ahmed Y. Abdallah 275
For example, one can consider f
j
(s) =
j
s, where
j
is a positive constant, g
j
(s) =

n
j
i=0

i j
s
2i+1
, where n
j
is a nonnegative integer and
i j
is a nonnegative constant for each
i = 0, 1, . . . , n
j
, h
1
(s
1
, s
2
) = s
m
1
1
1
s
m
2
2
, and h
2
(s
1
, s
2
) = s
m
1
1
s
m
2
1
2
, where m
1
, m
2
2 are con-
stants. In fact, for f
1
(s) =s, f
2
(s) =s, g
2
(s) =0, h
1
(s
1
, s
2
) =s
2
, and h
2
(s
1
, s
2
) =s
1
, where
and are positive constants, the system given by (1.3) and (1.4) can be regarded as
a spatial discretization of the following partly dissipative reaction-diusion system with
continuous spatial variable x R
r
and t R
+
:
u
t
u+u+g
1
(u) +v =q
1
, v
t
+v u =q
2
. (1.13)
The system (1.13) describes the signal transmission across axons and is a model of
FitzHugh-Nagumo equations in neurobiology, (cf. [5, 12, 15] and the references therein).
The existence of global attractors of the system given by (1.13) has been proved in a
bounded domain (cf. [14]) and in R
r
(cf. [16]).
2. Preliminaries
We can write the operator A in the following form:
A =A
1
+A
2
+ +A
r
(2.1)
such that for j =1, 2, . . . , r, and u =(u
i
)
iZ
r l
2
,
_
A
j
u
_
i
=2u
(i
1
,i
2
,...,i
r
)
u
(i
1
,i
2
,...,i
j1
,i
j
1,i
j+1
,...,i
r
)
u
(i
1
,i
2
,...,i
j1
,i
j
+1,i
j+1
,...,i
r
)
. (2.2)
For j =1, 2, . . . , r, dene the operators B
j
, B

j
: l
2
l
2
as follows: for u =(u
i
)
iZ
r l
2
,
_
B
j
u
_
i
=u
(i
1
,i
2
,...,i
j1
,i
j
+1,i
j+1
,...,i
r
)
u
(i
1
,i
2
,...,i
r
)
,
_
B

j
u
_
i
=u
(i
1
,i
2
,...,i
j1
,i
j
1,i
j+1
,...,i
r
)
u
(i
1
,i
2
,...,i
r
)
.
(2.3)
Then, it follows that for j =1, 2, . . . , r,
A
j
=B

j
B
j
=B
j
B

j
, (2.4)
there exists a constant C
0
=C
0
(r) such that
|Au|
2
C
0
|u|
2
,
_
_
B
j
u
_
_
2
=
_
_
B

j
u
_
_
2
4|u|
2
, u l
2
, (2.5)
_
B
j
u, v
_
=
_
u, B

j
v
_
, u, v l
2
. (2.6)
It is clear that A, A
j
, B
j
, and B

j
, j = 1, 2, . . . , r, are bounded linear operators from l
2
into l
2
.
We can represent the initial value problem, (1.3) and (1.4), in the following form:
u+Au+ f
1
(u) +g
1
(u) +h
1
(u, v) =q
1
,
v + f
2
(v) +g
2
(v) h
2
(u, v) =q
2
,
u(0) =
_
u
i,0
_
iZ
r
, v(0) =
_
v
i,0
_
iZ
r
,
(2.7)
276 Attractor for LDSs of PDRDSs
where u = (u
i
)
iZ
r , v = (v
i
)
iZ
r , Au = (Au
i
)
iZ
r , and for j = 1, 2, f
j
(u) = ( f
j
(u
i
))
iZ
r ,
g
j
(u) =(g
j
(u
i
))
iZ
r , h
j
(u, v) =(h
j
(u
i
, v
i
))
iZ
r , q
j
=(q
j,i
)
iZ
r .
Consider the Hilbert space E :=l
2
l
2
, endowed with the inner product and norm as
follows: for
j
=(u
( j)
, v
( j)
)
T
=((u
( j)
i
), (v
( j)
i
))
T
iZ
r E, j =1, 2,
_

1
,
2
_
E
=
_
u
(1)
, u
(2)
_
+
_
v
(1)
, v
(2)
_
,
||
E
=, )
1/2
E
, E.
(2.8)
Now, the system given by (2.7) is equivalent to the following initial value problem in
the Hilbert space E :=l
2
l
2
:
+C() =F(), (0) =
_
u
0
, v
0
_
T
, (2.9)
where =(u, v)
T
, C() =(Au, 0)
T
, F() =(G
1
(), G
2
())
T
,
G
1
() =q
1
f
1
(u) g
1
(u) h
1
(u, v),
G
2
() =q
2
f
2
(v) g
2
(v) +h
2
(u, v).
(2.10)
For a given function f (s
1
, s
2
) C
1
(R
2
, R), let D
1
f (a, b) represent the partial derivative
of f with respect to the rst independent variable, s
1
, at (s
1
, s
2
) =(a, b), and let D
2
f (a, b)
represent the partial derivative of f with respect to the second independent variable, s
2
, at
(s
1
, s
2
) =(a, b). From (1.9) and the mean value theorem, it follows that given u =(u
i
)
iZ
r ,
v =(v
i
)
iZ
r l
2
, there exist
1i
,
2i
(0, 1), for each i Z
r
, such that
_
_
h
1
(u, v)
_
_
2
=
_
iZ
r
_
h
1
_
u
i
, v
i
__
2
=
_
iZ
r
_
D
1
h
1
_

1i
u
i
, v
i
_
u
i
+D
2
h
1
_
0,
2i
v
i
_
v
i
_
2
2
_
max
a|u|
max
b|v|
_
D
1
h
1
(a, b)
_
2
_
|u|
2
+2
_
max
b|v|
_
D
2
h
1
(0, b)
_
2
_
|v|
2
.
(2.11)
Thus, for u, v l
2
, we have h
1
(u, v) l
2
. Similarly one can show that for u, v l
2
, h
2
(u, v)
l
2
. By using (1.8) and the mean value theorem, it is easy to show that f
1
, f
2
, g
1
, and g
2
map l
2
into l
2
. From the above discussion, it is obvious that F maps E into E.
3. The existence of an absorbing set
First we will prove that there exists a unique local solution of the system given by (2.9)
in E. Suppose that (1.7), (1.8), and (1.9) are satised. Let G be a bounded set in E, and
Ahmed Y. Abdallah 277

j
=(u
( j)
, v
( j)
) =((u
( j)
i
), (v
( j)
i
))
iZ
r G, j =1, 2, then
_
_
F
_

1
_
F
_

2
__
_
2
E
=
_
_
G
1
_

1
_
G
1
_

2
__
_
2
+
_
_
G
2
_

1
_
G
2
_

2
__
_
2
2
_
_
f
1
_
u
(1)
_
f
1
_
u
(2)
__
_
2
+4
_
_
g
1
_
u
(1)
_
g
1
_
u
(2)
__
_
2
+4
2
_
_
h
1
_
u
(1)
, v
(1)
_
h
1
_
u
(2)
, v
(2)
__
_
2
+2
_
_
f
2
_
v
(1)
_
f
2
_
v
(2)
__
_
2
+4
_
_
g
2
_
v
(1)
_
g
2
_
v
(2)
__
_
2
+4
2
_
_
h
2
_
u
(1)
, v
(1)
_
h
2
_
u
(2)
, v
(2)
__
_
2
.
(3.1)
Using (1.9) and the mean value theorem, it follows that there exist
3i
,
4i
(0, 1), for each
i Z
r
, and L
1
=L
1
(G) such that
_
_
h
1
_
u
(1)
, v
(1)
_
h
1
_
u
(2)
, v
(2)
__
_
2
=
_
iZ
r
_
h
1
_
u
(1)
i
, v
(1)
i
_
h
1
_
u
(2)
i
, v
(2)
i
__
2
=
_
iZ
r
_
_
_
D
1
h
1
_
u
(1)
i
+
3i
_
u
(2)
i
u
(1)
i
_
, v
(1)
i
__
u
(1)
i
u
(2)
i
_
+D
2
h
1
_
u
(2)
i
, v
(1)
i
+
4i
_
v
(2)
i
v
(1)
i
___
v
(1)
i
v
(2)
i
_
_
_
_
2
2
_
max
a|u
(1)
|+|u
(2)
|,b|v
(1)
|
_
D
1
h
1
(a, b)
_
2
_
_
_
u
(1)
u
(2)
_
_
2
+2
_
max
a|u
(2)
|,b|v
(1)
|+|v
(2)
|
_
D
2
h
1
(a, b)
_
2
_
_
_
v
(1)
v
(2)
_
_
2
L
1
_
_
_
u
(1)
u
(2)
_
_
2
+
_
_
v
(1)
v
(2)
_
_
2
_
.
(3.2)
Thus
_
_
h
1
_
u
(1)
, v
(1)
_
h
1
_
u
(2)
, v
(2)
__
_
2
L
1
_
_

2
_
_
2
. (3.3)
We can obtain similar results, as (3.3), for f
1
, f
2
, g
1
, g
2
, and h
2
by using (1.8), (1.9), and
the mean value theorem. In such a case, using (3.1), there exists L
2
=L
2
(G) such that
_
_
F
_

1
_
F
_

2
__
_
2
E
L
2
_
_

2
_
_
2
. (3.4)
Thus F is locally Lipschitz from E into E. In such a case from the standard theory of
ordinary dierential equations, we get the following lemma.
Lemma 3.1. If (1.7), (1.8), and (1.9) are satised, then for any initial data (0) = (u
0
,
v
0
)
T
E, there exists a unique local solution (t) = (u(t), v(t))
T
of (2.9) such that
C
1
([0, T), E) for some T > 0. If T < , then lim
tT
||
2
E
=.
Assume that (1.5), (1.7), (1.8), (1.9), (1.10), (1.11), and (1.12) are satised. Now we
are ready to prove that the solution of the system given by (2.9) exists globally and there
exists an absorbing set.
278 Attractor for LDSs of PDRDSs
Let =(u, v)
T
E be a solution of (2.9). If we consider the inner product of (2.9) with
in E, taking into account (2.4) and (2.6), we obtain
1
2
d
dt
|u|
2
+
r
_
j=1
_
_
B
j
u
_
_
2
+
_
f
1
(u), u
_
+
_
g
1
(u), u
_
+
_
h
1
(u, v), u
_
=
_
q
1
, u
_
, (3.5)
1
2
d
dt
|v|
2
+
_
f
2
(v), v
_
+
_
g
2
(v), v
_

_
h
2
(u, v), v
_
=
_
q
2
, v
_
, (3.6)
If we multiply (3.5) by , (3.6) by , and we sum up the results, taking into account
(1.5) and (1.12), we nd that

2
d
dt
|u|
2
+

2
d
dt
|v|
2
+
r
_
j=1
|B
j
u|
2
+
_
f
1
(u), u
_
+
_
f
2
(v), v
_
+
_
g
1
(u), u
_
+
_
g
2
(v), v
_
=
_
q
1
, u
_
+
_
q
2
, v
_
.
(3.7)
By using (1.8), (1.10), and the mean value theorem, for each i Z
r
, there exists a constant

5i
(0, 1) such that
_
f
1
(u), u
_
=
_
iZ
r
_
f
1
_
u
i
_
u
i
_
=
_
iZ
r
_
f

1
_

5i
u
i
_
u
2
i
_

1
|u|
2
. (3.8)
Thus we have
_
f
1
(u), u
_

1
|u|
2
,
_
f
2
(v), v
_

2
|v|
2
. (3.9)
From (1.11), we obtain
_
g
1
(u), u
_
=
_
iZ
r
_
g
1
_
u
i
_
u
i
_
0,
_
g
2
(v), v
_
0. (3.10)
Now if we substitute (3.9) and (3.10) into (3.7), we nd that

2
d
dt
|u|
2
+

2
d
dt
|v|
2
+
1
|u|
2
+
2
|v|
2

_
q
1
, u
_
+
_
q
2
, v
_

2
1
_
_
q
1
_
_
2
+

1

2
|u|
2
+

2
2
_
_
q
2
_
_
2
+

2

2
|v|
2
.
(3.11)
Thus if we choose
=min
_
, ,
1
,
2

_
, (3.12)
Ahmed Y. Abdallah 279
then it follows that
d
dt
||
2
E
+||
2
E

_
_
q
1
_
_
_
2
+
_

_
_
q
2
_
_
_
2
. (3.13)
From the Gronwall lemma, we obtain
||
2
E
e
t
_
_

0
_
_
2
E
+
1 e
t

2
_
_

_
_
q
1
_
_
_
2
+
_

_
_
q
2
_
_
_
2
_
, (3.14)
lim
t
||
2
E

_
_
q
1
_
_
_
2
+
_

_
_
q
2
_
_
_
2
. (3.15)
Inequality (3.15) implies that the solution semigroup S(t)
t0
of (2.9) exists globally and
possesses a bounded absorbing set in E. In such a case, the maps
S(t) : (0) E S(t)(0) =(t) E, t 0, (3.16)
generate a continuous semigroup S(t)
t0
on E. Now from Lemma 3.1 and (3.15), we
are ready to present the following lemma.
Lemma 3.2. If (1.5), (1.7), (1.8), (1.9), (1.10), (1.11), and (1.12) are satised, then for
any initial data in E, the solution (t) of (2.9) exists globally for all t 0. That is,
C
1
([0, ), E). Moreover, there exists a bounded ball O = O
E
(0, r
0
) in E, centered at 0 with
radius r
0
, such that for every bounded set G of E, there exists T(G) 0 such that
S(t)GO, t T(G), (3.17)
where r
2
0
> ((/)|q
1
|)
2
+((/)|q
2
|)
2
. Therefore, there exists a constant T
0
0 depend-
ing on O such that
S(t)O O, t T
0
. (3.18)
4. The existence of the global attractor
To prove the existence of the global attractor for the solution semigroup S(t)
t0
of (2.9),
we need to prove the asymptotic compactness of S(t)
t0
. Along the lines of [4], the
key idea of showing the asymptotic compactness for such a lattice system is to establish
uniform estimates on Tail Ends of solutions.
Lemma 4.1. If (1.5), (1.7), (1.8), (1.9), (1.10), (1.11), and (1.12) are satised and (0) =
(u
0
, v
0
) O, where O is the bounded absorbing ball given by Lemma 3.2, then for any > 0,
there exist positive constants T() and K() such that the solution (t) = (u(t), v(t))
T
=
(
i
(t))
iZ
r =((u
i
(t)), (v
i
(t)))
T
iZ
r E of (2.9) satises
_
|i|
m
K()
_
_

i
(t)
_
_
2
E
=
_
|i|
m
K()
_
u
2
i
(t) +v
2
i
(t)
_
(4.1)
for all t T(), where |i|
m
=max
1jr
i
j
for i =(i
1
, i
2
, . . . , i
r
) Z
r
.
280 Attractor for LDSs of PDRDSs
Proof. Consider a smooth increasing function C
1
(R
+
, R) such that
(s) =0, 0 s < 1,
0 (s) 1, 1 s < 2,
(s) =1, s 2,
(4.2)
and there exists a constant M
0
such that

(s) M
0
for all s R
+
.
Let L be an arbitrary positive integer. Set w
i
= (|i|
m
/L)u
i
, z
i
= (|i|
m
/L)v
i
, w =
(w
i
)
iZ
r , z = (z
i
)
iZ
r , and = (w, z)
T
. Following [4], we take the inner product of (2.9)
with in E, then it follows that
_
iZ
r

_
|i|
m
L
__
1
2
d
dt
u
2
i
+ f
1
_
u
i
_
u
i
+g
1
_
u
i
_
u
i
+h
1
_
u
i
, v
i
_
u
i
_
+
_
iZ
r
r
_
j=1
_
B
j
u
_
i
_
B
j
w
_
i
=
_
iZ
r

_
|i|
m
L
_
q
1,i
u
i
,
(4.3)
_
iZ
r

_
|i|
m
L
__
1
2
d
dt
v
2
i
+ f
2
_
v
i
_
v
i
+g
2
_
v
i
_
v
i
h
2
_
u
i
, v
i
_
v
i
_
=
_
iZ
r

_
|i|
m
L
_
q
2,i
v
i
. (4.4)
If we multiply (4.3) by , (4.4) by , and we sum up the results, taking into account
(1.5) and (1.12), we get
_
iZ
r

_
|i|
m
L
_
_
_
_

2
d
dt
u
2
i
+

2
d
dt
v
2
i
+ f
1
_
u
i
_
u
i
+ f
2
_
v
i
_
v
i
+g
1
_
u
i
_
u
i
+g
2
_
v
i
_
v
i
_
_
_
+
_
iZ
r
r
_
j=1
_
B
j
u
_
i
_
B
j
w
_
i
=
_
iZ
r

_
|i|
m
L
_
_
q
1,i
u
i
+q
2,i
v
i
_
.
(4.5)
Using (1.8), (1.10), and the mean value theorem, it follows that for each i Z
r
,
f
1
_
u
i
_
u
i

1
u
2
i
, f
2
_
v
i
_
v
i

2
v
2
i
. (4.6)
From (1.11), we obtain
g
1
_
u
i
_
u
i
0, g
2
_
v
i
_
v
i
0. (4.7)
Recalling (69) of [21], taking into account that |B
j
| 2, j =1, 2, . . . , r, one can see that

_
iZ
r
r
_
j=1
_
B
j
u
_
i
_
B
j
w
_
i

16nM
0
L
r
2
0
, t T
0
. (4.8)
Ahmed Y. Abdallah 281
Now if we substitute (4.6), (4.7), and (4.8), into (4.5), we nd that for t T
0
,
_
iZ
r

_
|i|
m
L
__

2
d
dt
u
2
i
+

2
d
dt
v
2
i
+
1
u
2
i
+
2
v
2
i
_

_
iZ
r

_
|i|
m
L
_
_
q
1,i
u
i
+q
2,i
v
i
_
+
16nM
0
L
r
2
0

_
iZ
r

_
|i|
m
L
__

1
q
2
1,i
+

1

4
u
2
i
+

2
q
2
2,i
+

2

4
v
2
i
_
+
16nM
0
L
r
2
0
.
(4.9)
Thus if we choose
=min
_
, ,
1
,
2

_
, (4.10)
then it follows that for t T
0
,
_
iZ
r

_
|i|
m
L
__
d
dt
_
_

i
(t)
_
_
2
E
+
_
_

i
(t)
_
_
2
E
_
2
_
iZ
r

_
|i|
m
L
_
_
_

q
1,i
_
2
+
_

q
2,i
_
2
_
+
32nM
0
L
r
2
0
2
_
|i|
m
L
_
_

q
1,i
_
2
+
_

q
2,i
_
2
_
+
32nM
0
L
r
2
0
.
(4.11)
Since q
1
, q
2
l
2
, then for a given > 0, we can x L such that
2
_
|i|
m
L
_
_

q
1,i
_
2
+
_

q
2,i
_
2
_
+
32nM
0
L
r
2
0


2
, (4.12)
and in such a case we get that
_
iZ
r

_
|i|
m
L
__
d
dt
_
_

i
(t)
_
_
2
E
+
_
_

i
(t)
_
_
2
E
_

2
, t T
0
. (4.13)
From the Gronwall lemma, we obtain
_
iZ
r
_

_
|i|
m
L
_
_
_

i
(t)
_
_
2
E
_
e
t
_
iZ
r
_

_
|i|
m
L
_
_
_

i
(0)
_
_
2
E
_
+

2
, t T
0
. (4.14)
Since (0) =(u
0
, v
0
)
T
O, we have
_
_
(0)
_
_
E
r
0
. (4.15)
Therefore,
_
iZ
r
_

_
|i|
m
L
_
_
_

i
(t)
_
_
2
E
_
r
2
0
e
t
+

2
, t T
0
. (4.16)
282 Attractor for LDSs of PDRDSs
But again for > 0, there exists a constant T
1
=T
1
() such that
r
2
0
e
t

2
, t T
1
. (4.17)
Using (4.16) and (4.17) with K() =2L, T() =maxT
0
, T
1
, we obtain
_
|i|
m
K()
_
_

i
_
_
2
E
=
_
|i|
m
K()
_

_
|i|
m
L
_
_
_

i
_
_
2
E
_

_
iZ
r
_

_
|i|
m
L
_
_
_

i
_
_
2
E
_
, t T().
(4.18)
The proof is completed.
Lemma 4.2. If (1.5), (1.7), (1.8), (1.9), (1.10), (1.11), and (1.12) are satised, then the solu-
tion semigroup S(t)
t0
of (2.9) is asymptotically compact in E, that is, if
n
is bounded
in E and t
n
, then S(t
n
)
n
is precompact in E.
Proof. By using Lemmas 3.2 and 4.1, above, the proof of this lemma will be similar to
that of [19, Lemma 3.2].
Theorem 4.3. If (1.5), (1.7), (1.8), (1.9), (1.10), (1.11), and (1.12) are satised, then the
solution semigroup S(t)
t0
of (2.9) possesses a global attractor in E.
Proof. From the existence theorem of global attractors, (cf. [18, Lemmas 2 and 4]), we
get the result.
5. Upper semicontinuity of the global attractor
Here we will study the upper semicontinuity of the global attractor of the solution
semigroup S(t)
t0
of (2.9), in the sense that is approximated by the global attractors
of nite-dimensional versions of (2.9), as was done in [4] for a simpler system.
Let n be a positive integer, and
Z
r
n
=
_
i Z
r
: |i|
m
n
_
. (5.1)
For i =(i
1
, i
2
, . . . , i
r
) Z
r
n
, consider w =(w
i
)
|i|
m
n
R
(2n+1)
r
. For convenience, we reorder
the subscripts of components of w as follows:
w =
_
w
(n,n,...,n,n)
, w
(n,n,...,n,n+1)
, . . . , w
(n,n,...,n,n)
,
w
(n,n,...,n+1,n)
, w
(n,n,...,n+1,n+1)
, . . . ,
w
(n,n,...,n+1,n)
, . . . , w
(n,n,...,n,n)
, w
(n,n,...,n,n+1)
, . . . , w
(n,n,...,n,n)
_
T
.
(5.2)
Let
X =w =
_
w
i
_
|i|
m
n
: w R
(2n+1)
r
, (5.3)
where subscripts of components of w are ordered as in (5.2).
Ahmed Y. Abdallah 283
Consider the (2n + 1)
r
-dimensional ordinary dierential equations with the initial
data in X:
w
i
+(Aw)
i
+ f
1
_
w
i
_
+g
1
_
w
i
_
+h
1
_
w
i
, z
i
_
=q
1,i
,
z
i
+ f
2
_
z
i
_
+g
2
_
z
i
_
h
2
_
w
i
, z
i
_
=q
2,i
,
w
i
(0) =w
i,0
, z
i
(0) =z
i,0
i Z
r
n
, t > 0.
(5.4)
The system (5.4) can be written as
w+
`
Aw+
`
f
1
(w) + ` g
1
(w) +
`
h
1
(w, z) = ` q
1
,
z +
`
f
2
(z) + ` g
2
(z)
`
h
2
(w, z) = ` q
2
,
w(0) =
_
w
i,0
_
|i|
m
n
, z(0) =
_
z
i,0
_
|i|
m
n
,
(5.5)
where w =(w
i
)
|i|
m
n
, z =(z
i
)
|i|
m
n
X, for k =1, 2, . . . , r,
w
(i
1
,...,n,i
k
+1,...,i
r
)
=w
(i
1
,...,n+1,i
k
+1,...,i
r
)
,
w
(i
1
,...,n,i
k
+1,...,i
r
)
=w
(i
1
,...,n1,i
k
+1,...,i
r
)
,
(
`
Aw)
(i
1
,i
2
,...,i
r
)
=2rw
(i
1
,i
2
,...,i
r
)
w
(i
1
1,i
2
,...,i
r
)
w
(i
1
,i
2
1,...,i
r
)
w
(i
1
,i
2
,...,i
r
1)
w
(i
1
+1,i
2
,...,i
r
)
w
(i
1
,i
2
+1,...,i
r
)
w
(i
1
,i
2
,...,i
r
+1)
,
(5.6)
and for j =1, 2,
`
f
j
(w) =
_
f
j
_
w
i
__
|i|
m
n
, ` g
j
(w) =
_
g
j
_
w
i
__
|i|
m
n
,

h
j
(w, z) =
_
h
j
_
w
i
, z
i
__
|i|
m
n
, ` q
j
=
_
q
j,i
_
|i|
m
n
.
(5.7)
For w =(w
i
)
|i|
m
n
X, dene the linear operators
`
B
j
,

B

j
,
`
A
j
: X X, j =1, 2, . . . , r, by
_
`
B
j
w
_
i
=w
(i
1
,...,i
j
+1,...,i
r
)
w
(i
1
,...,i
j
,...,i
r
)
,
_

j
w
_
i
=w
(i
1
,...,i
j
1,...,i
r
)
w
(i
1
,...,i
j
,...,i
r
)
,
_
`
A
j
w
_
i
=2w
(i
1
,...,i
j
,...,i
r
)
w
(i
1
,...,i
j
1,...,i
r
)
w
(i
1
,...,i
j
+1,...,i
r
)
,
(5.8)
then
`
A=
`
A
1
+
`
A
2
+ +
`
A
r
,
`
A
j
=
`
B
j

B

j
=

B

j
`
B
j
, j =1, 2, . . . , r. (5.9)
For w
( j)
=(w
( j)
i
)
|i|
m
n
X, j =1, 2, i =(i
1
, i
2
, . . . , i
r
) Z
r
n
, dene
_
w
(1)
, w
(2)
_
X
=
_
|i|
m
n
w
(1)
i
w
(2)
i
,
_
_
w
(1)
_
_
X
=
_
w
(1)
, w
(1)
_
1/2
X
. (5.10)
284 Attractor for LDSs of PDRDSs
In such a case, it is clear that X = (X, | |
X
) is a Hilbert space, and
`
E = X X is a
Hilbert space with the following inner product and norm: for Y
j
=(w
( j)
, z
( j)
)
T
=((w
( j)
i
),
(z
( j)
i
))
T
|i|
m
n

`
E, j =1, 2,
_
Y
1
, Y
2
_
`
E
=
_
w
(1)
, w
(2)
_
X
+
_
z
(1)
, z
(2)
_
X
,
_
_
Y
1
_
_
`
E
=
_
_
_
w
(1)
_
_
2
X
+
_
_
z
(1)
_
_
2
X
_
1/2
. (5.11)
It is easy to check that problem (5.5) can be formulated to the following rst-order
system in the Hilbert space
`
E:

Y +
`
C(Y) =
`
F(Y), Y(0) =
_
w(0), z(0)
_
T

`
E, (5.12)
where
Y =(w, z)
T
,
`
C(Y) =(
`
Aw, 0)
T
,
`
F(Y) =
_
`
G
1
(Y),
`
G
2
(Y)
_
T
,
`
G
1
(Y) = ` q
1

`
f
1
(w) ` g
1
(w)
`
h
1
(w, z),
`
G
2
(Y) = ` q
2

`
f
2
(z) ` g
2
(z) +
`
h
2
(w, z).
(5.13)
Similar to Section 2, one can see that, if (1.5), (1.7), (1.8), (1.9), (1.10), (1.11), and
(1.12) are satised, then (5.12) is a well-posed problem in
`
E. Thus for any Y(0)
`
E, there
exists a unique solution Y C([0, +),
`
E) C
1
((0, +),
`
E), see Lemmas 3.1 and 3.2, also
there exist maps of solutions S
n
(t) : Y(0)
`
E Y(t) =S
n
(t)Y(0)
`
E, t 0, generating a
continuous semigroup S
n
(t)
t0
on
`
E.
Similar to Lemma 3.2 and Theorem 4.3, we can prove the following Lemma.
Lemma 5.1. If (1.5), (1.7), (1.8), (1.9), (1.10), (1.11), and (1.12) are satised, then there
exists a bounded ball
`
O =
`
O
`
E
(0, r
0
) in
`
E, centered at 0 with radius r
0
such that for every
bounded set
`
G of
`
E, there exists T(
`
G) 0 such that
S
n
(t)
`
G
`
O, t T(
`
G), n =1, 2, . . . , (5.14)
where r
0
is the same constant given by Lemma 3.2, and it is independent of n. Moreover, the
semigroup S
n
(t)
t0
possesses a global attractor
n
,
n

`
O
`
E.
Here we prove the upper semicontinuity of the global attractor of the solution semi-
group S(t)
t0
of (2.9). In such a case, we should extend the element u =(u
i
)
|i|
m
n
X
to an element of l
2
such that u
i
=0 for |i|
m
> n, still denoted by u.
Lemma 5.2. If (1.5), (1.7), (1.8), (1.9), (1.10), (1.11), and (1.12) are satised, and
n
(0)

n
, then there exists a subsequence
n
k
(0) of
n
(0) and
0
such that
n
k
(0) con-
verges to
0
in E.
Proof. Consider
n
(t) =S
n
(t)
n
(0) =(u
n
(t), v
n
(t))
T

`
E to be a solution of (5.12). Since

n
(0)
n
,
n
(t)
n

`
O for all t R
+
, and again the element
n
= (
n,i
)
|i|
m
n

`
E
can be extended to the element
n
= (
n,i
)
iZ
r E, where
n,i
= (0, 0)
T
for |i|
m
> n, it
Ahmed Y. Abdallah 285
follows that
_
_

n
(t)
_
_
`
E
=
_
_

n
(t)
_
_
E
=
__
_
u
n
_
_
2
+
_
_
v
n
_
_
2
_
1/2
r
0
, t R
+
, n =1, 2, . . . .
(5.15)
From (2.5) and (5.15), we get that
_
_
`
C
_

n
(t)
__
_
2
`
E

_
_
C
_

n
(t)
__
_
2
E
=
_
_
Au
n
_
_
2
C
0
_
_
u
n
_
_
2
C
0
r
2
0
, (5.16)
for all t R
+
, n =1, 2, . . . .
Similarly, for j = 1, 2, the element ` q
j
= (q
j,i
)
|i|
m
n
R
(2n+1)
r
can be extended to the
element q
n, j
= (q
n, j,i
)
iZ
r l
2
, where q
n, j,i
= q
j,i
for |i|
m
n and q
n, j,i
= 0 for |i|
m
> n.
From (1.8) and (1.9), we know that for j =1, 2,
f
j
(0) =g
j
(0) =h
j
(0, 0) =0. (5.17)
Therefore,
_
_
`
F
_

n
(t)
__
_
2
`
E
=
_
_
F
_

n
(t)
__
_
2
E
=
_
_
q
n,1
f
1
_
u
n
_
g
1
_
u
n
_
h
1
_
u
n
, v
n
__
_
2
+
_
_
q
n,2
f
2
_
v
n
_
g
2
_
v
n
_
+h
2
_
u
n
, v
n
__
_
2
2
_
_
q
1
_
_
2
+4
_
_
f
1
_
u
n
__
_
2
+8
_
_
g
1
_
u
n
__
_
2
+8
2
_
_
h
1
_
u
n
, v
n
__
_
2
+2
_
_
q
2
_
_
2
+4
_
_
f
2
_
v
n
__
_
2
+8
_
_
g
2
_
v
n
__
_
2
+8
2
_
_
h
2
_
u
n
, v
n
__
_
2
.
(5.18)
Using (1.9), (5.15), and the mean value theorem, there exist
6i
,
7i
(0, 1) for each i Z
r
such that
_
_
h
1
_
u
n
, v
n
__
_
2
=
_
iZ
r
_
h
1
_
u
n,i
, v
n,i
__
2
=
_
iZ
r
_
D
1
h
1
_

6i
u
n,i
, v
n,i
_
u
n,i
+D
2
h
1
_
0,
7i
v
n,i
_
v
n,i
_
2
2
_
max
a,b[r
0
,r
0
]
_
D
1
h
1
(a, b)
_
2
_
r
2
0
+2
_
max
b[r
0
,r
0
]
_
D
2
h
1
(0, b)
_
2
_
r
2
0
(5.19)
for all t R
+
, n =1, 2, . . . . Again using (1.8), (1.9), (5.15), and the mean value theorem,
we can get similar results, as (5.19), for f
1
, f
2
, g
1
, g
2
, and h
2
. Thus from(5.18), it is obvious
that there exists a constant C
1
=C
1
(r
0
, q
1
, q
2
, f
1
, f
2
, g
1
, g
2
, h
1
, h
2
) such that
_
_
`
F
_

n
(t)
__
_
2
`
E
=
_
_
F
_

n
(t)
__
_
2
E
C
1
, t R
+
, n =1, 2, . . . . (5.20)
Now, by using (5.12), we obtain
_
_

n
(t)
_
_
2
E
2
_
_
C
_

n
(t)
__
_
2
E
+2
_
_
F
_

n
(t)
__
_
2
E
. (5.21)
Hence, from (5.16) and (5.20), it follows that there exists a constant C
2
= C
2
(r
0
, C
0
, C
1
)
such that
_
_

n
(t)
_
_
E
C
2
, t R
+
, n =1, 2, . . . . (5.22)
286 Attractor for LDSs of PDRDSs
Let J
k
(k = 1, 2, . . .) be a sequence of compact intervals of R
+
such that J
k
J
k+1
and

k
J
k
= R
+
. Then there exists a subsequence of
n
(t), still denoted by
n
(t), and
(t) C(R
+
, E) such that

n
(t) (t) in C(J, E) as n for any compact set J R
+
, (5.23)

n
(t) (t) weak star in L

_
R
+
, E
_
as n +. (5.24)
From (5.15) and (5.23), we obtain that there exists a constant C
3
= C
3
(r
0
) such that for
(t) =(u(t), v(t))
T
=((u
i
(t)), (v
i
(t)))
T
iZ
r E,
_
_
(t)
_
_
E
=
_
|u|
2
+|v|
2
_
1/2
C
3
, t R
+
. (5.25)
Let i Z
r
and n |i|
m
. Since
n
(t) =(u
n
(t), v
n
(t))
T
=((u
n,i
(t)), (v
n,i
(t)))
T
|i|
m
n

`
E is the
solution of (5.12), it follows that for all t R
+
and i Z
r
n1
,
u
n,i
+
_
Au
n
_
i
+ f
1
_
u
n,i
_
+g
1
_
u
n,i
_
+h
1
_
u
n,i
, v
n,i
_
=q
1,i
,
v
n,i
+ f
2
_
v
n,i
_
+g
2
_
v
n,i
_
h
2
_
u
n,i
, v
n,i
_
=q
2,i
.
(5.26)
Therefore, for all C

0
(J), we obtain
_
J
_
u
n,i
+
_
Au
n
_
i
+ f
1
_
u
n,i
_
+g
1
_
u
n,i
_
+h
1
_
u
n,i
, v
n,i
__
(t)dt =
_
J
q
1,i
(t)dt,
_
J
_
v
n,i
+ f
2
_
v
n,i
_
+g
2
_
v
n,i
_
h
2
_
u
n,i
, v
n,i
__
(t)dt =
_
J
q
2,i
(t)dt.
(5.27)
From (1.8), and by using the mean value theorem, there exists
8i
(0, 1) for each i
such that

_
J
f
1
_
u
n,i
_
(t)dt
_
J
f
1
_
u
i
_
(t)dt

sup
tJ

f
1
_
u
n,i
_
f
1
_
u
i
_

_
J

(t)

dt
sup
tJ
_

1
_
u
n,i
+
8i
_
u
i
u
n,i
__

u
n,i
u
i

_
_
J

(t)

dt.
(5.28)
By using (5.15) and (5.25), it is clear that there exists a constant C
4
=C
4
(r
0
) such that

u
n,i
+
i
_
u
i
u
n,i
_

u
n,i

u
i

_
_
u
n
_
_
+|u| C
4
(5.29)
for all t R
+
and n =1, 2, . . . . In such a case, we obtain
sup
tJ

f

1
_
u
n,i
+
8i
_
u
i
u
n,i
__

sup
a[C
5
,C
5
]

1
(a)

< +. (5.30)
From (5.23), we know that as n , we have
sup
tJ

u
n,i
u
i

0. (5.31)
Ahmed Y. Abdallah 287
Therefore, from (5.28), (5.30), and (5.31), it is clear that

_
J
f
1
_
u
n,i
_
(t)dt
_
J
f
1
_
u
i
_
(t)dt

0, as n . (5.32)
Similarly, one can get the same result, as (5.32), for f
2
, g
1
, g
2
, h
1
, and h
2
. In such a case
from (5.24), (5.26), and (5.27), it follows that
u
i
+(Au)
i
+ f
1
_
u
i
_
+g
1
_
u
i
_
+h
1
_
u
i
, v
i
_
=q
1,i
,
v
i
+ f
2
_
v
i
_
+g
2
_
v
i
_
h
2
_
u
i
, v
i
_
=q
2,i
.
(5.33)
But J is arbitrary, thus (5.33) holds for all t R
+
, which means that (t) is a solution of
(2.9). From (5.25), it follows that (t) is bounded for t R
+
, that is, (t) , therefore

n
(0) (0) , and the proof is completed.
Now we are ready to represent the main result of this section. In fact, the following
theorem shows that the global attractor of the lattice dynamical system (2.9) is upper
semicontinuous with respect to the (cuto) approximate nite-dimensional dynamical
system (5.12).
Theorem 5.3. If (1.5), (1.7), (1.8), (1.9), (1.10), (1.11), and (1.12) are satised, then
lim
n
d
E
_

n
, ) =0, (5.34)
where d
E
(
n
, ) =sup
a
n
inf
b
|a b|
E
.
Proof. We argue by contradiction. If the conclusion is not true, then there exists a se-
quence
n
k

n
k
and a constant K > 0 such that
d
E
_

n
k
, ) K > 0. (5.35)
However, by Lemma 5.2, we know that there exists a subsequence
n
km
of
n
k
such that
d
E
_

n
km
, ) 0, (5.36)
which contradicts (5.35). The proof is completed.
References
[1] V. S. Afraimovich, S.-N. Chow, and J. K. Hale, Synchronization in lattices of coupled oscillators,
Phys. D 103 (1997), no. 1-4, 442451.
[2] V. S. Afraimovich and V. I. Nekorkin, Chaos of traveling waves in a discrete chain of diusively
coupled maps, Internat. J. Bifur. Chaos Appl. Sci. Engrg. 4 (1994), no. 3, 631637.
[3] P. W. Bates and A. Chmaj, Adiscrete convolution model for phase transitions, Arch. Ration. Mech.
Anal. 150 (1999), no. 4, 281305.
[4] P. W. Bates, K. Lu, and B. Wang, Attractors for lattice dynamical systems, Internat. J. Bifur. Chaos
Appl. Sci. Engrg. 11 (2001), no. 1, 143153.
[5] J. Bell, Some threshold results for models of myelinated nerves, Math. Biosci. 54 (1981), no. 3-4,
181190.
288 Attractor for LDSs of PDRDSs
[6] J. Bell and C. Cosner, Threshold behavior and propagation for nonlinear dierential-dierence
systems motivated by modeling myelinated axons, Quart. Appl. Math. 42 (1984), no. 1, 1
14.
[7] S.-N. Chow and J. Mallet-Paret, Pattern formation and spatial chaos in lattice dynamical systems.
I, IEEE Trans. Circuits Systems I Fund. Theory Appl. 42 (1995), no. 10, 746751.
[8] , Pattern formation and spatial chaos in lattice dynamical systems. II, IEEE Trans. Circuits
Systems I Fund. Theory Appl. 42 (1995), no. 10, 752756.
[9] S.-N. Chow, J. Mallet-Paret, and W. Shen, Traveling waves in lattice dynamical systems, J. Dier-
ential Equations 149 (1998), no. 2, 248291.
[10] S.-N. Chow, J. Mallet-Paret, and E. S. Van Vleck, Pattern formation and spatial chaos in spatially
discrete evolution equations, Random Comput. Dynam. 4 (1996), no. 2-3, 109178.
[11] R. Dogaru and L. O. Chua, Edge of chaos and local activity domain of FitzHugh-Nagumo equa-
tion, Internat. J. Bifur. Chaos Appl. Sci. Engrg. 8 (1998), no. 2, 211257.
[12] R. FitzHugh, Impulses and physiological states in theoretical models of nerve membrane, Biophys.
J. 1 (1961), 445466.
[13] J. P. Keener, Propagation and its failure in coupled systems of discrete excitable cells, SIAM J. Appl.
Math. 47 (1987), no. 3, 556572.
[14] M. Marion, Finite-dimensional attractors associated with partly dissipative reaction-diusion sys-
tems, SIAM J. Math. Anal. 20 (1989), no. 4, 816844.
[15] J. Nagumo, S. Arimoto, and S. Yosimzawa, An active pulse transmission line simulating nerve
axon, Proc. I. R. E. 50 (1964), 20612070.
[16] A. Rodriguez-Bernal and B. Wang, Attractors for partly dissipative reaction diusion systems in
R
n
, J. Math. Anal. Appl. 252 (2000), no. 2, 790803.
[17] W. Shen, Lifted lattices, hyperbolic structures, and topological disorders in coupled map lattices,
SIAM J. Appl. Math. 56 (1996), no. 5, 13791399.
[18] R. Temam, Innite-Dimensional Dynamical Systems in Mechanics and Physics, Applied Mathe-
matical Sciences, vol. 68, Springer-Verlag, New York, 1988.
[19] S. Zhou, Attractors for second order lattice dynamical systems, J. Dierential Equations 179
(2002), no. 2, 605624.
[20] , Attractors for rst order dissipative lattice dynamical systems, Phys. D 178 (2003), no. 1-
2, 5161.
[21] , Attractors and approximations for lattice dynamical systems, J. Dierential Equations
200 (2004), no. 2, 342368.
[22] B. Zinner, Existence of traveling wavefront solutions for the discrete Nagumo equation, J. Dier-
ential Equations 96 (1992), no. 1, 127.
Ahmed Y. Abdallah: Department of Mathematics, University of Jordan, Amman 11942, Jordan
E-mail address: farah@ju.edu.jo