See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/343878767

STABILITY RESULT FOR AN ABSTRACT TIME DELAYED EVOLUTION EQUATION

WITH ARBITRARY DECAY OF VISCOELASTICITY

Article · June 2020

CITATIONS READS

0 14

2 authors:

Houria Chellaoua Yamna Boukhatem

Université Amar Telidji Laghouat Université Amar Telidji Laghouat
2 PUBLICATIONS   0 CITATIONS    14 PUBLICATIONS   77 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

BLOW UP OF SOLUTIONS FOR A SEMILINEAR HYPERBOLIC EQUATION View project

All content following this page was uploaded by Houria Chellaoua on 26 August 2020.

The user has requested enhancement of the downloaded file.

Proceedings of International Mathematical Sciences
ISSN:2717-6355, URL:https://dergipark.org.tr/tr/pub/pims
Volume II Issue 1 (2020), Pages 7-25.

STABILITY RESULT FOR AN ABSTRACT TIME DELAYED

EVOLUTION EQUATION WITH ARBITRARY DECAY OF
VISCOELASTICITY

HOURIA CHELLAOUA*, YAMNA BOUKHATEM**

* LABORATORY OF PURE AND APPLIED MATHEMATICS, UNIVERSITY OF
LAGHOUAT, P.O. BOX 37G, LAGHOUAT (03000), ALGERIA.
** LABORATORY OF PURE AND APPLIED MATHEMATICS, UNIVERSITY OF
LAGHOUAT, P.O. BOX 37G, LAGHOUAT (03000), ALGERIA.

Abstract. The paper is concerned with a second-order abstract semilinear

evolution equation with infinite memory and time delay. With the help of the
semigroup arguments and under suitable conditions on initial data and the
kernel memory function, we state and prove the global existence of solution.
Then, we establish the decay rates of the energy using the multiplier method
by defining a suitable Lyapunov functional. This work extends previous works
with time delay for a much wider class of kernels. We give also some applica-
tions to illustrate our results.

1. Introduction
Let H be a real Hilbert space with inner product and related norm denoted
by h., .i and k . k, respectively. Let A : D(A) −→ H and B : D(B) −→ H be a
self-adjoint linear positive operator with domains D(A) ⊂ D(B) ⊂ H such that
the embeddings are dense and compact. Let C : H −→ H is a self-adjoint linear
operator and h : R+ −→ R+ is the kernel of the memory term. τ > 0 represents
1
a time delay and F : D(A 2 ) → H is function satisfying some conditions to be
specified later. We consider the following second-order abstract semilinear evolution
equation with infinite memory and time delay
 R +∞
 utt (t) + Au(t) − 0 h(s)Bu(t − s)ds + Cut (t − τ ) = F (u(t)), t ∈ (0, +∞),
u (t − τ ) = f0 (t − τ ) t ∈ (0, τ ),
 t
u(−t) = u0 (t), ut (0) = u1 , t ∈ R+ ,
(1.1)
where the initial datum (u0 , u1 , f0 ) belongs to a suitable spaces.

2020 Mathematics Subject Classification. 35L90; 35B40; 93D20; 74D05.

Key words and phrases. Abstract evolution equation; Decay rates; Infinite memory; Time
delay.
2019
c Proceedings of International Mathematical Sciences.
Submitted on August 31, 2019. Published on June 30th, 2020.
Area Editors: Valéria Neves DOMINGOS CAVALCANTİ and Idris DAG.
7
8 HOURIA CHELLAOUA, YAMNA BOUKHATEM

In absence of time delay term, a large number of works are available, where
various decay estimates were obtained, see [7, 14, 21]. For the particular case of
the wave equation with finite memory, see [2, 24].
In many cases, delay is a source of instability and even an arbitrarily small delay
may destabilize a system which is uniformly asymptotically stable in the absence
of delay. Nicaise and Pignotti in [15] considered a wave equation with a linear
damping and delay term and they proved that the energy is exponentially stable
and some instability results are also given by constructing some sequences of delays
for which the energy of some solutions does not tend to zero, see also [3, 17].
When the memory term is replaced by a frictional damping But (t):

utt (t) + Au(t) + But (t) + µut (t − τ ) = 0, t > 0,

where µ, τ are fixed constants and B is a given operator, there exist in the literature
different stability results. These results show that the damping But (t) is strong
enough to stabilize the system in presence of a time delay provided that |µ| is small
enough, see [10, 16, 17].
Guesmia in [11] considered the following second-order abstract linear problem
with infinite memory and time delay terms
 R +∞
 utt (t) + Au(t) − 0 h(s)Au(t − s)ds + µut (t − τ ) = 0, t > 0,
u(−t) = u0 (t), t ∈ R+
ut (0) = u1 , ut (t − τ ) = f0 (t − τ ), t ∈ (0, τ ),


He proved that the unique dissipation given by the memory term is strong enough
to stabilize exponentially the system in presence of delay. In this work and others,
the condition h0 (s) ≤ −δh(s) for all s ≥ 0 and some δ > 0 is assumed to prove
exponential decay of the energy, see [1, 4]. In [13], the previous condition is replaced
by
h0 (s) ≤ −ζ(t)h(s), ∀s ≥ 0, (1.2)

where ζ is a positive nonincreasing differentiable function. The authors established

the existence and the general decay results of the energy. Dai and Yang in [8] con-
sidered the same problem in [13] and solved the open problem proposed by Kirane
and Said-Houari. Recently, Boukhatem and Benabderrahmane in [5] considered a
variable coefficient viscoelastic equation with a time-varying delay in the bound-
ary feedback and acoustic boundary conditions and nonlinear source term. They
established a general decay results of the energy via suitable Lyapunov functionals
and some properties of the convex functions where the kernel memory satisfies the
equation (1.2). In [6], the same results have obtained in the case of constant delay.
Tatar in [23] introduced a new class of admissible kernels which lead to a wide
range of possible decay rates. More precisely, He consider kernels satisfying
Z +∞
h(t − s) ≥ ξ(t) h(π − s)dπ, 0 ≤ s ≤ t,
t

for some ξ(t) > 0. This class contains the polynomial type functions and the expo-
nential type. He proved that the last assumption on the relaxation in a viscoelastic
problem ensuring uniform stability in an arbitrary rate.
 STABILITY RESULT 9

For the case of distributed time delay, Guesmia and Tatar in [12] considered the
following class of second-order linear hyperbolic equations
 R +∞ R +∞
 utt (t) + Au(t) − 0 h(s)Bu(t − s)ds + 0 f (s)ut (t − s)ds = 0, t > 0,
u(−t) = u0 (t), t ∈ R+
ut (0) = u1 , t ∈ R+ ,


where the function f is of class C 1 (R+ , R) and satisfies, for some positive constant
α,
|f (s)| ≤ αh(s), and |f 0 (s)| ≤ αh(s), ∀s ∈ R+ .
They given well-posedness and stability of the system and they proved that the
infinite memory alone guarantees the asymptotic stability of the system and the
decay rate of solutions is found explicitly in terms of the growth at infinity of the
infinite memory and the distributed time delay convolution kernels.
Nicaise and Pignotti in [18] considered the following system

Ut (t) = AU (t) + F (U (t)) + kBU (t − τ ), t ∈ (0, +∞),
U (0) = u0 , BU (t − τ ) = f (t), t ∈ (0, τ ),
where A generates a C0 -semigroup (S(t))t≥0 that is exponentially stable, i.e., there
exist two positive constants M and w such that
kS(t)kL(H) ≤ M e−wt , ∀t ≥ 0,
and L(H) denotes the space of bounded linear operators from H into itself. For
a fixed delay parameter τ , a fixed bounded operator B from H into itself and
for a real parameter k and F : H −→ H satisfies some Lipschitz conditions, the
initial datum U0 belongs to H and f ∈ C([0, τ ]; H). They showed that, if the
C0 -semigroup describing the linear part of the model is exponentially stable, then
the whole system retains this good property when a suitable smallness condition
on the time-delay feedback is satisfied, see also [19].
Motivated by previous works, we study the well-posedness and the stability result
of a semilinear abstract viscoelastic equation with infinite memory in presence of
a time delayed damping and a nonlinear source term. Our results extend the
decay results in previous works to kernels h which do not necessarily converge
exponentially to zero at infinity. Moreover, our problem generalizes the linear
problems to those with a nonlinear source term and to problems with more general
time delayed damping term.
The paper is organized as follows. In Sect. 2, we prove the well-posedness
by using the semigroup arguments under some assumptions on A, B, C, h and F .
Then, we state and prove the stability result of solution by using the energy method
to produce a suitable Lyapunov functional with arbitrary decay on h. Section 4 is
devoted to some concrete examples in the aim to illustrate our abstract result.

2. Well-posedness
In this section, we state some assumptions on A, B, C and h and prove the
well-posedness result by using semigroup theory.
For studying the problem (1.1), we introduce a new variable z as in [15]
z(ρ, t) = ut (t − ρτ ), ρ ∈ (0, 1), t > 0.
10 HOURIA CHELLAOUA, YAMNA BOUKHATEM

Thus, we have
τ zt (ρ, t) + zρ (ρ, t) = 0, ρ ∈ (0, 1), t > 0.
Moreover, as in [9], we define
η t (s) = u(t) − u(t − s), t, s > 0.
Therefore, problem (1.1) takes the form
 R +∞ t
 utt (t) + Au(t) − h0 Bu(t) + 0 h(s)Bη (s)ds

+Cz(1, t) = F (u(t)), t ∈ (0, +∞),




τ z (ρ, t) + z (ρ, t) = 0, ρ ∈ (0, 1), t > 0,



 t ρ
ηtt (s) = ut (t) − ηst (s), t, s > 0,

(2.1)
 z(ρ, 0) = f0 (−ρτ ),
 ρ ∈ (0, 1),
 z(0, t) = ut (t), t > 0,



 u(−t) = u0 (t), ut (0) = u1 , t ≥ 0,



η 0 (s) = u0 (0) − u0 (s), s ≥ 0.


We will need the following assumptions:

(A1) There exist positive constants a and b satisfying
1 2 1 2 1
bkuk2 ≤ B 2 u ≤ a A 2 u , ∀u ∈ D(A 2 ). (2.2)

(A2) The kernel function h : R+ −→ R+ is of class C 1 nonincreasing function

satisfying Z +∞
1
h0 = h(s)ds < . (2.3)
0 a
(A3) There exists µ ∈ R∗ such that
2
kCuk ≤ |µ|kuk2 , ∀u ∈ H. (2.4)
1
(A4) F : D(A ) → H is globally Lipschitz continuous, namely
2
1
∃γ > 0 such that kF (u) − F (v)k ≤ γ A 2 (u − v) , ∀u, v ∈ H

Let us denote U = (u, ut , η t , z)T , the problem (2.1) can be rewritten:

Ut (t) = AU (t) + F(U (t)), ∀t > 0,


T (2.5)
U (0) = U0 = u0 , u1 , η 0 , f0 (−τ.) ,
where the operator A is defined by
 
  φ2
φ1  −(A − h0 B)φ1 − +∞ h(s)Bφ3 (s)ds − Cφ4 (1)
R

 φ2   0 
A ∂φ3
 φ3  = 
  
φ 2− 
 ∂s 
φ 4
 −1 ∂φ4 
τ ∂ρ
and
F(φ1 , φ2 , φ3 , φ4 )T = (0, F (φ1 ), 0, 0)T
The domain D(A) is given by
 R +∞ 

 (φ1 , φ2 , φ3 , φ4 )T ∈ H, (A − h0 B)φ1 + 0 h(s)Bφ3 (s)ds ∈ H,

 1 ∂φ3 2 1


D(A) = φ 2 ∈ D(A 2 ), ∈ Lh (R + , D(B 2 )),
∂s
 ∂φ4 
∈ L2 (0, 1; H), φ3 (0) = 0, φ4 (0) = φ2

 

 
∂ρ
 STABILITY RESULT 11

where
1 1
H = D(A 2 ) × H × L2h (R+ , D(B 2 )) × L2 (0, 1; H).
1
The sets L2h (R+ , D(B 2 )) and L2 (0, 1; H) are respectively defined by
 Z +∞ 1 2 
1 1
2
Lh (R+ , D(B )) = φ : R+ → D(B ), h(s) B φ(s) ds < +∞ ,
2 2
2
0
equipped with the inner product
Z +∞ D 1 E
1
hφ1 , φ2 i 1 = h(s) B 2 φ1 (s), B 2 φ2 (s) ds.
L2h (R+ ,D(B 2 ))
0
And  Z 1 
L2 (0, 1; H) = φ : (0, 1) → H, kφ(ρ)k2 dρ < +∞ ,
0
equipped with the inner product
Z 1
hφ1 , φ2 iL2 (0,1;H) = hφ1 (ρ), φ2 (ρ)i dρ.
0
The Hilbert space H equipped with the following inner product. For all Φ =
(φ1 , φ2 , φ3 , φ4 )T and W = (w1 , w2 , w3 , w4 )T in H, we have
hΦ, W iH = hφ1 , w1 i 1 − h0 hφ1 , w1 i 1 + hφ2 , w2 i
D(A 2 ) D(B 2 )
+hφ3 , w3 i 1 + τ µhφ4 , w4 iL2 (0,1;H) .
L2h (R+ ,D(B 2 ))

The well-posedness of problem (2.5) is ensured by the following theorem:

Theorem 2.1. Under the assumptions (A1)-(A4), for an initial datum U0 ∈ H,
the system (2.5) has a unique mild solution U ∈ C(R+ , H) satisfies the following
formula, Z t
U (t) = S(t)U0 + S(t − s)F(U (s))ds.
0
Moreover, if U0 ∈ D(A) and F ∈ C 1 (H), then the solution of (2.5) satisfies (clas-
sical solution)
U ∈ C(R+ , D(A)) ∩ C 1 (R+ , H).
Proof. To prove Theorem 2.1, we use the semigroup theory. The problem (2.5)
can be seen as an inhomogeneous evolution problem. It’s clear that F is globally
lipschitz continuous, let show that the operator A generate a linear C0 -semigroup
(S(t))t≥0 on H. Indeed,
• First, we prove that the linear operator A is dissipative.
Take Φ = (φ1 , φ2 , φ3 , φ4 )T ∈ D(A), then
Z +∞

 D ∂φ3 E
AΦ, Φ H = hφ2 , φ1 i 1 + h(s) φ 2 − , φ3 1 ds
D(A 2 )
0 ∂s D(B 2 )
Z 1D
−1 ∂φ4 E
−h0 hφ2 , φ1 i 1 + τ |µ| , φ 4 dρ
D(B 2 )
0 τ ∂ρ
D Z +∞ E
− (A − h0 B)φ1 + h(s)Bφ3 (s)ds + Cφ4 (1), φ2 .
0
1 1
Using the definition of A and B and the fact that H is a real Hilbert space, we
2 2

conclude
1 1 1 1
hA − h0 Bφ1 , φ2 i = hA 2 φ2 , A 2 φ1 i − h0 hB 2 φ2 , B 2 φ1 i (2.6)
12 HOURIA CHELLAOUA, YAMNA BOUKHATEM

using the Cauchy-Schwarz and Young’s inequalities and by (2.4), we have

|µ|
kφ4 (1)k2 + kφ2 k2 .


− hCφ4 (1), φ2 i ≤ (2.7)
2
Z +∞  Z +∞
h(s)Bφ3 (s)ds, φ2 = h(s)hφ2 , φ3 i 1 ds.
2 D(B )
0 0
Integrating by parts and using the definition of D(A) (φ3 (0) = 0), we obtain
Z +∞
1 +∞ 0
D ∂φ E Z
3 1
h(s) − , φ3 1 ds ≤ h (s)kB 2 φ3 (s)k2 ds. (2.8)
0 ∂s D(B 2 ) 2 0
Also using the fact that φ4 (0) = φ2 , we obtain
Z 1D
−1 ∂φ4 E |µ|   |µ|  
τ |µ| , φ4 dρ = kφ4 (0)k2 − kφ4 (1)k2 = kφ2 k2 − kφ4 (1)k2 .
0 τ ∂ρ 2 2
(2.9)
Consequently, inserting (2.6), (2.7), (2.8) and (2.9) in (2.6) and using the fact that
h is nonincreasing, we find
1 +∞ 0
Z 2
1
hAΦ, ΦiH ≤ h (s) B 2 φ3 (s) ds + |µ|kut k2 ≤ |µ|kΦk2 , (2.10)

2 0
which means that the operator A − |µ|I is dissipative.
• Let us now prove that λI − A is surjective. Indeed, let (f1 , f2 , f3 , f4 )T ∈ H,
we show that there exists Φ = (φ1 , φ2 , φ3 , φ4 )T ∈ D(A) satisfying
   
φ1 f1
 φ2   f2 
(λI − A)   φ3  =  f3  ,
  

φ4 f4
which is equivalent to

 λφ1 − φ2 = f1
 R +∞
λφ 2 + (A − h0 B)φ1 + 0 h(s)Bφ3 (s)ds + Cφ4 (1) = f2




∂φ3 (2.11)
λφ3 − φ2 + = f3

 ∂s

 1 ∂φ4
 λφ4 +
 = f4 .
τ ∂ρ
Suppose that we have found φ1 with the appropriate regularity. Then, we have
φ2 = λφ1 − f1 . (2.12)
We note that the third equation in (2.11) with φ3 (0) = 0 has a unique solution
Z s
φ3 (s) = e−λs eλy (f3 (y) − f1 + λφ1 ) dy. (2.13)
0
On the other hand, the fourth equation in (2.11) with φ4 (0) = φ2 = λφ1 − f1 has
a unique solution
 Z ρ 
φ4 (ρ) = λφ1 − f1 + τ f4 (y)eλτ y dy e−λτ ρ , ρ ∈ (0, 1). (2.14)
0
In particular,  Z 1 
φ4 (1) = λφ1 − f1 + τ f4 (y)eλτ y dy e−λτ .
0
It remains only to determine φ1 .
 STABILITY RESULT 13

Next, plugging (2.12) and (2.13) into the second equation in (2.11), we get

A − αB + λe−λτ C + λ2 I φ1 = f˜,


(2.15)
where
Z ∞ Z s  Z ∞
α = h0 − λ h(s)e−λs eλy dy ds = h(s)e−λs ds,
0 0 0
and
 Z 1 
f˜ = f2 + λf1 + e−λτ C f1 − τ f4 (y)eτ y dy
0
Z ∞ Z s
− e−λs h(s) e−λy B (f3 (y) − f1 ) dyds.
0 0
1
We have just to prove that (2.15) has a solution φ1 ∈ D(A 2 ) and replace in (2.12),
(2.13) and (2.14) to obtain Φ ∈ D(A) satisfying (2.11).
We have α < h0 , by (2.3) and (2.2), we deduce that A − αB is a positive definite
1
operator. Then, we take the duality brackets h., .i 1
2 0
1 with w ∈ D(A 2 ) :
2 D(A ) ×D(A )
D E
A − αB + λe−λτ C + λ2 I φ1 , w = f˜, w



1 1 1 1 . (2.16)
D(A 2 )0 ×D(A 2 ) D(A 2 )0 ×D(A 2 )

Consequently, the left-hand side of (2.16) is bilinear, continuous and coercive on

1
D(A 2 ). Since, applying the Lax-Milgram theorem and classical regularity argu-
1
ments, we conclude that (2.11) has a unique solution φ1 ∈ D(A 2 ) satisfying. Using
(2.13),
 Z +∞ 
(A − h0 B)φ1 + h(s)Bφ3 (s)ds ∈ H.
0

In conclusion, we have found Φ = (φ1 , φ2 , φ3 , φ4 )T ∈ D(A), which verifies (2.11),

and thus λI − A is surjective for all λ > 0 and the same holds for the operator
λI − (A − |µ|I).
Then, the Lumer-Phillips theorem implies that |µ|I − A is a maximal monotone
operator, A−|µ|I is an infinitesimal generator of a strongly continuous semigroup of
contraction in H. Hence, the operator A generates a strongly continuous semigroup
(S(t))t≥0 in H. Consequently, by using Theorem 1.2, Ch. 6 of [22], the problem
(2.5) has a unique solution U ∈ C([0, +∞), H). 

3. Stability result
The stability result of the solution of (2.1) holds under the following additional
assumptions:
(A5) There exist a positive constant d satisfying
1 2 1 2 1
A u ≤ d B 2 u , ∀u ∈ D(A 2 ). (3.1)
2

(A6) Moreover, we assume that F (0) = 0 and there exists a continuous and
1
differentiable mapping ψ : D(A 2 ) → R satisfying
1
Dψ = F and hF (u), ui ≥ 2ψ(u), ∀u ∈ D(A 2 ). (3.2)
14 HOURIA CHELLAOUA, YAMNA BOUKHATEM

(A7) The function h satisfies (A2) and there exists a positive function ξ ∈
C(R+ , R∗+ ) satisfying lims→+∞ ξ(s) exists such that

+∞
 Z
h(t − s) ≥ ξ(t) h(π − s)dπ, ∀t ∈ R+ , ∀s ∈ [0, t],

t
(3.3)
0
h (s) < 0, ∀s ∈ R+ .


The first inequality in (3.3), introduced in [25] and [23], implies that h converges
to zero at least exponentially but it does not involve the derivative of h. This class
contains the polynomial (or power) type (h(t) = (1 + t)−a , a > 1) functions and
the exponential type (h(t) = e−at , a > 0) functions.
Let establish some several Lemmas needed of our main result. We define the
modified energy functional E associated to problem (2.1) by
 Z +∞
21 2
1 1 2 1 2
2
E(t) = A u − h0 B u + kut k + h(s) B 2 η t (s) ds
2
2 0
Z 1 
−2ψ(u) + τ |µ| kz(ρ, t)k2 dρ . (3.4)
0

Lemma 3.1. Assume that (A1)-(A4) hold and let U0 ∈ D(A). Then, the energy
functional defined by (3.4) satisfies
Z +∞ 2
0 1 1
E (t) ≤ h0 (s) B 2 η t (s) ds + |µ|kut k2 . (3.5)

2 0

Proof. Multiplying the first equation of (2.1) by ut . Using (A6) and repeating
exactly the same arguments to obtain (2.10). 

Remark. Note that, from (3.5), the energy of solutions to problem (2.1) is not
decreasing in general. Indeed, the second term in the right-hand side of (3.5),
coming from the delay term, is nonnegative.

Now, as in [20], for n ∈ N∗ , let consider the set

An = {s ∈ R+ , h(s) + nh0 (s) ≤ 0},

and put hn = Ac h(s)ds. We have hn > 0, otherwise, Acn = ∅. Furthermore, by

R
n
the second inequality in (3.3), we have

lim Acn = ∩n∈N∗ Acn = ∅, and then lim hn = 0.

n→+∞ n→+∞

In order to state our results, we need the following four lemmas.

Lemma 3.2. Let U be solution of (2.1). Then the functional

 Z +∞ 
t
I1 (t) = − ut (t), h(s)η (s)ds , (3.6)
0
 STABILITY RESULT 15

satisfies, for ε1 , ε2 > 0,

√
12 2 h20 12 2
 
dhn 
I10 (t) ≤ −(h0 − ε1 )kut k + ε2 + 2
A u − B u
2 2
 √  Z +∞
h0 dhn 2

1
+ 2hn − + h(s) B 2 η t (s) ds
2 2 0
h0 +∞
Z 1 2
+ h(s) B 2 u(t − s) ds

2 0
  Z +∞ 2
dnh0 h(0) 1
− 2nh0 + + h0 (s) B 2 η t (s) ds

4ε2 4bε1 0
 Z +∞ 
+ Cz(1, t) − F (u), h(s)η t (s)ds , (3.7)
0

Proof. Differentiating (3.6) with respect to t and using the third equation of 2.1,
we find
 Z +∞   Z +∞ 
0 t
I1 (t) = − utt (t), h(s)η (s)ds + ut (t), h(s)ηs (s)ds − h0 kut k2 .
t
0 0

Integrating by parts with respect to s the second term in the right hand side of the
previous equality and using the fact that lims→+∞ h(s) = 0, η t (0) = 0, we obtain
 Z +∞   Z +∞ 
0 t 0
I1 (t) = − utt (t), h(s)η (s)ds − ut (t), h (s)η (s)ds − h0 kut k2 .
t
0 0

On the other hand, by the first equation of (2.1), we have

 Z +∞   Z +∞ 
utt (t), h(s)η t (s)ds + Au(t), h(s)η t (s)ds
0 0
 Z +∞   Z +∞ Z +∞ 
− h0 Bu(t), h(s)η t (s)ds + h(s)Bη t (s)ds, h(s)η t (s)ds
0 0 0
 Z +∞ 
t
+ Cz(1, t) − F (u), h(s)η (s)ds = 0,
0
1 1
using the definitions of A and B 2 , we get
2

 Z +∞ 
I10 (t) = −h0 kut k2 + Cz(1, t) − F (u), h(s)η t (s)ds
0
 Z +∞   Z +∞ 
1 1
− ut (t), h0 (s)η t (s)ds + A 2 u(t), h(s)A 2 η t (s)ds
0 0
Z +∞ 2  Z +∞ 
1 1 1
t t


h(s)B η (s)ds − h0 B u(t),
2 2 h(s)B η (s)ds .(3.8)
2

0 0

Let estimate the last three terms in the right hand by using Cauchy-Schwarz and
Young’s inequalities and the definition of An . Then, using (2.2), (3.1) and (2.3),
we get
Z +∞
h(0) +∞ 0
  Z 
21 t 2

0 t 2
− ut (t), h (s)η (s)ds ≤ ε1 kut k − h (s) B η (s) ds ,
0 4bε1 0
16 HOURIA CHELLAOUA, YAMNA BOUKHATEM

 Z +∞ 
1 1
A 2 u(t), h(s)A 2 η t (s)ds
0
 Z   Z 
1 1 1 1
= A u(t),
2 h(s)A 2 η t (s)ds + A 2 u(t), h(s)A 2 η t (s)ds .
An Ac
1 2 dh Z 2 √n
dhn 12 2
1
0
≤ ε2 A 2 u + h(s) B 2 η t (s) ds + A u

4ε2 An 2
√ Z
dhn 2
1
+ h(s) B 2 η t (s) ds

2 Anc

1 2 dnh Z +∞ √
2 t 2 21 2
dhn
1
0
≤ ε2 A u − h(s) B η (s) ds + A u
2
4ε2 0 2
√
dhn +∞
Z 1 2
+ h(s) B 2 η t (s) ds,

2 0
Z +∞ 2
1
t


h(s)B 2 η (s)ds
0
Z Z 2
1 1
t t

=
h(s)B η (s)ds +
2 h(s)B η (s)ds
2

An Acn
Z 2 Z 2
1 1
2 η t (s)ds + 2 2 η t (s)ds

≤ 2
h(s)B c h(s)B
A An
Zn 1 2 Z 1 2
≤ 2h0 h(s) B 2 η t (s) ds + 2hn h(s) B 2 η t (s) ds

An Acn
Z +∞ 1 2 Z +∞ 1 2
≤ −2nh0 h0 (s) B 2 η t (s) ds + 2hn h(s) B 2 η t (s) ds.

0 0
And for the last one, we have
 Z +∞ 
1 1
−h0 B 2 u(t), h(s)B 2 η t (s)ds
0
2  Z +∞ 
2 12 1 1
= −h0 B u + h0 B u(t), 2 h(s)B u(t − s)ds
2

0
+∞
h2
Z
1 2 h0
1 2
= − 0 B 2 u + h(s) B 2 u(t − s) ds

2 2 0
h0 +∞
Z 1 2
− h(s) B 2 η t (s) ds. (3.9)

2 0
Inserting these four inequalities in (3.8), we get (3.7). 
Lemma 3.3. Let U be solution of (2.1). Then the functional
I2 (t) = hut (t), u(t)i, (3.10)
satisfies,
1 2 h 1 2 1 Z +∞ 1 2
0
I20 (t) = kut k2 − A 2 u + B 2 u + h(s) B 2 u(t − s) ds

2 2 0
1 +∞
Z  
2 t 2
1
− h(s) B η (s) ds − Cz(1, t) + F (u), u . (3.11)
2 0
 STABILITY RESULT 17

Proof. Differentiating (3.10) with respect to t, we find

I20 (t) = kut k2 + hutt (t), u(t)i.
On the other hand, multiplying the first equation of (2.1) by u(t), we have
 Z +∞ 
t
hutt (t), u(t)i + h(A − h0 B)u(t), u(t)i + h(s)Bη (s)ds, u(t)
0
+hCz(1, t), u(t)i = 0,
1 1
By the definitions of A 2 and B 2 , we have
1 2 1 2  Z +∞ 
t
hutt (t), u(t)i + A u − h0 B 2 u + h(s)Bη (s)ds, u(t)
2
0
+hCz(1, t), u(t)i = 0.
Consequently,
1 2 1 2  Z +∞ 
I20 (t) = kut k2 − A 2 u +h0 B 2 u − h(s)Bη t (s)ds, u(t) −hCz(1, t), u(t)i,

0

By using the inequality (3.9), we get (3.11). 

Similarly to [15], we introduce the following functional.

Lemma 3.4. Let U be solution of (2.1). Then the functional
Z 1
I3 (t) = τ e2τ e−2τ ρ kz(ρ, t)k2 ds, (3.12)
0
satisfies,
Z 1
I30 (t) ≤ −2τ kz(ρ, t)k2 ds + e2τ kut k2 − kz(1, t)k2 . (3.13)
0

Proof. By using the second equation of (2.1), we get

Z 1
I30 (t) = 2τ e2τ e−2τ ρ hzt (ρ, t), z(ρ, t)idρ
0
Z 1
= −2e2τ e−2τ ρ hzρ (ρ, t), z(ρ, t)idρ
0
Z 1
∂
= −2e 2τ
e−2τ ρ kz(ρ, t)k2 dρ.
0 ∂ρ
Then, by integrating by parts and z(0, t) = ut (t), we get
Z 1
I30 (t) = −2τ e2τ e−2τ ρ kz(ρ, t)k2 ds + e2τ kut k2 − kz(1, t)k2 ,
0

which is (3.13) by using the fact that e−2τ ρ ≥ e−2τ , for any ρ ∈]0, 1[. 

Now, we consider two functionals J1 and J2 and we give their derivatives in the
following lemma.
Lemma 3.5. Let
Z tZ +∞  2
1
J1 (t) = h(π − s)dπ B 2 η t (s) ds, ∀t ∈ R+ , (3.14)

0 t
18 HOURIA CHELLAOUA, YAMNA BOUKHATEM

and Z tZ +∞ 
12 t 2

J2 (t) = h(π − s)dπ A η (s) ds, ∀t ∈ R+ . (3.15)
0 t
Then, for any λ1 ∈]0, 1[,
1 2 Z t 1 2
0
J1 (t) ≤ h0 B u − (1 − λ1 )ξ(t)J1 (t) − λ1 h(s) B 2 u(t − s) ds
2
0
Z +∞ 1 2
+λ1 h(s) B 2 u0 (s − t) ds, ∀t ∈ R+ , (3.16)

t
and
λ1 t
1 2 Z 1 2
J20 (t) ≤ h0 A 2 u − (1 − λ1 )ξ(t)J2 (t) − h(s) B 2 u(t − s) ds

a 0
Z +∞ 1 2
+dλ1 h(s) B 2 u0 (s − t) ds, ∀t ∈ R+ . (3.17)

t

Proof. The functional J1 is well-defined. Indeed, by using the fact that η ∈

1
L2h (R+ , D(B 2 )) and (3.3), we have
Z t 2 Z t 2
1 1 1 1
J1 (t) ≤ h(t − s) B 2 u(s) ds ≤ h(s) B 2 u(t − s) ds < +∞.

ξ(t) 0 ξ(t) 0
By (3.1), we conclude that J2 also is well defined.
Then, differentiating J1 with respect to t and using the definition of u0 and (3.3),
we obtain
 Z +∞  2 Z t 1 2
0 21
J1 (t) = h(π − s)dπ B u(t) − h(t − s) B 2 u(s) ds

t 0
1 2 Z t 1 2
= h0 B u − (1 − λ1 ) h(t − s) B u(s) ds
2 2
0
Z t 1 2 Z 0 1 2
−λ1 h(t − s) B 2 u(s) ds + λ1 h(t − s) B 2 u(s) ds

−∞ −∞
1 2 Z +∞ 1 2
≤ h0 B 2 u − (1 − λ1 )ξ(t)J1 (t) − λ1 h(s) B 2 u(t − s) ds

0
Z +∞ 1 2
+aλ1 h(s) B u0 (s − t) ds,
2
t
which is exactly (3.16). A similar argument yields the relation (3.17). 
In this case, the Lyapunov functional L we will work with is


L(t) = E(t) +  N1 I1 (t) + N2 I2 (t) + I3 (t) + M1 J1 (t) + aM1 J2 (t), (3.18)
where , N1 , N2 , M1 > 0 are positive constants to be chosen later.
Now we are in position to state and prove the decay result of solution of problem
(2.1).
Theorem 3.6. Assume that (A1)-(A7) hold. For any initial datum U0 ∈ H.
Assume that h satisfies
Z +∞
γ2
h(s)ds < , (3.19)
0 b
 STABILITY RESULT 19

and there exists a positive constant δ0 independent of µ such that, if

|µ| < δ0 , (3.20)
then, for any U0 ∈ H, there exist positive constants δ1 and δ2 such that
Z t Z +∞ !
1 2
−δ1 t δ1 s
E(t) ≤ δ2 e 1+ e h(π) B u0 (π − s) dπds , ∀t ∈ R+ , (3.21)
2
0 s

if limt→+∞ ξ(t) > 0, and

!
Z t Z +∞ 1 2
E(t) ≤ δ2 e−δ1 ξ̂(t) 1+ eδ1 ξ̂(s) h(π) B u0 (π − s) dπds , ∀t ∈ R+ ,
2
0 s
(3.22)
if limt→+∞ ξ(t) = 0, where
Z s
ˆ =
ξ(s) ξ(π)dπ, ∀t ∈ R+ . (3.23)
0

Proof. In order to proof the decay estimates, we start by the derivative of the
function L. On the other hand, by using (A6) and (2.2), we have
 Z +∞  Z +∞ 2
t 1 2 b t

− F (u), h(s)η (s)ds ≤ kF (u)k + h(s)η (s)ds
0 b 4 0
+∞
γ 2 1 2 h0
Z 2
1
≤ A 2 u + h(s) B 2 η t (s) ds,

b 4 0
Combining (3.5), (3.7), (3.11), (3.13), (3.16) and (3.17), we obtain
 Z 1
|µ| 1 2 1 2
L0 (t) ≤ − (C1 − )kut k2 + C2 A 2 u + C3 h0 B 2 u − 2τ kz(ρ, t)k2 dρ

 0
Z +∞  2 1 2  
1
+ h(s) C4 B 2 η t (s) + C5 B 2 u(t − s) ds − 2N2 ψ(u)

0
√ 1 2  √  Z +∞ 2
dhn dhn 1
+ N1 A u + 2hn + N1 h(s) B 2 η t (s) ds
2
2 2 0
  Z +∞ 2
1 1
+ − C6 h0 (s) B 2 η t (s) ds − C7 ξ(t)(J1 (t) + J2 (t))

2 0
Z +∞ 1 2
+C8 h(s) B 2 u0 (s − t) ds − kz(1, t)k2

t
 Z +∞ 
+ Cz(1, t), N1 h(s)η t (s)ds − N2 u , (3.24)
0
where
γ2 ah0
C1 = (h0 − ε1 )N1 − N2 − e2τ , C2 = N2 − (ε2 + )N1 − M1 ,
b 
h0 N2 M1 h0 N2
C3 = N1 − − , C4 = N1 + ,
2 2  4 2
2λ1 h0 N2

dnh0 h(0)
 (3.25)
C5 = M1 − N1 − , C6 = 2nh0 + + N1 ,
 2 2 4ε2 4bε1
C7 = (1 − λ1 )M1 min{1, a}, C8 = M1 λ1 (1 + ad).
20 HOURIA CHELLAOUA, YAMNA BOUKHATEM

At this point, we choose the different constants to obtain some results. First, we
select N2 = (1 + ah0 )e2τ and we choose M1 , N1 such that

N2 e2τ
< M1 < .
2(1 + ah0 ) 2
(   )  
b M1 1 2τ 1 2M1
max 2(1 + ah0 ) − N2 , (N2 + e ) < N1 < N2 + .
bh0 − 2γ 2  h0 h0 

Note that M1 exists as a result of the selection of N2 for certain value of  to be

choose later and the choice of M1 and N2 guarantees the existence of N1 . Now, let
pick ε1 , ε2 and λ1 such that

N2 + e2τ
0 < ε1 < h0 − ,
N1
 
h0 γ2 1 M1
ε2 = − + N2 − 2(1 + ah0 ) ,
2 b 2N1 
and

(N2 + h0 N1 ) ≤ λ1 < 1,
4M1
ε2 and λ1 exist by the previous selection of N1 and N2 . Consequently, it result that
C1 > 0, C2 = −C3 , C3 < 0 and C5 ≥ 0. Moreover, it’s clear that C4 > 0, so, we
have
  1 2 
21 2

2
− C1 kut k + C2 A u − h0 B 2 u − 2N2 ψ(u)

Z +∞  2  
12 t 2
1
+ h(s) C4 B η (s) + C5 B u(t − s) ds
2
0
 1 2 1 2 Z +∞ 
2 t 2
1
2
≤ −C9 kut k + A u − h0 B u − 2ψ(u) + h(s) B η (s) ds
2 2
0

where
1 
C9 = min C1 , C2 , C4 .
N2
Observe that C9 is positive and independent on µ. Next, using Cauchy-Schwarz’s
and Young’s inequalities for estimate the last term in the right hand in (3.24).
Then, by (2.2) and (2.4), we get
 Z +∞ 
 Cz(1, t), N1 h(s)η t (s)ds − N2 u
0
 Z +∞ 
12 2 2 t 2
1
2
≤ kz(1, t)k + |µ|C10 A u + h(s) B η (s) ds ,
0

where
1
C10 = max{aN22 , h0 N12 }.
2b
 STABILITY RESULT 21

Inserting the above inequality and (3.26) in (3.24), we obtain

 √ 
4hn + dhn
L0 (t) ≤ −C11 E(t) + N1 E(t)
2
  Z +∞ 2
1 1
+ − C6 h0 (s) B 2 η t (s) ds − C7 ξ(t)(J1 (t) + J2 (t))

2 0
Z +∞ 1 2
+C8 h(s) B 2 u0 (s − t) ds, (3.26)

t

where  
|µ| 2
C11 = 2 min C9 − , , C9 − |µ|C10 .
 |µ|
Finally, we assume that |µ| satisfies (3.20) under the following choice of δ0
 √ 
C9 C9 2
δ0 = min ,√ . (3.27)
C6 C10
Then, we can choose n big enough and we fix  such that
|µ| 1 1
<≤ < , (3.28)
2C9 2C6 M
where
2e2τ
   
h0 a
M = N1 max 1, + N2 max 1, + .
b b |µ|
which imply that E is equivalent to E + (N1 I1 + N2 I2 + I3 ). Indeed, by using
Cauchy-Schwarz’s and Young’s inequalities, we have
h0 +∞
 Z 2 
1 1
|I1 (t)| ≤ kut k2 + h(s) B 2 η t (s) ds (3.29)

2 b 0
 
h0
≤ max 1, E(t) (3.30)
b
and
   
1 2 a 12 2 a
|I2 (t)| ≤ kut k + A u ≤ max 1, E(t). (3.31)
2 b b
From (3.12), it follows
Z 1 1
2e2τ
Z
|I3 (t)| = τ e2τ e−2τ ρ kz(ρ, t)k2 ds ≤ τ e2τ e−2τ ρ kz(ρ, t)k2 ds ≤
E(t)
0 0 |µ|
(3.32)
Combining (3.4), (3.29), (3.31) and (3.32) and by using (3.28), we have
E ∼ E + (N1 I1 + N2 I2 + I3 ).
Moreover, the third term in the right hand of (3.26) is non-positive. Note that
δ0 is a positive constant independent of µ. Under the condition (3.20), we conclude
that C11 is a positive constant and by using the fact that limn→+∞ hn = 0, we get
 √ 
4hn + dhn
C12 = C11 + N1 > 0.
2
22 HOURIA CHELLAOUA, YAMNA BOUKHATEM

Consequently, we obtain, for all t ∈ R+ ,

 
L0 (t) ≤ −C12 E(t) − C7 ξ(t) J1 (t) + J2 (t)
Z +∞ 1 2
+C8 h(s) B 2 u0 (s − t) ds. (3.33)

t

Let distinguish two cases corresponding to the limit of ξ at infinity.

I If limt→+∞ ξ(t) > 0, there exist t0 ≥ 0 and ξ0 > 0 such that ξ(t) ≥ ξ0 , for all
t ≥ t0 . Therefore, using (3.18), we find
Z +∞ 1 2
0
L (t) ≤ −δ1 L(t) + C8 h(s) B 2 u0 (s − t) ds, ∀t ∈ R+ , (3.34)

t

where  
C12 C7 ξ0 C7 ξ0
δ1 = min , , .
1 + M M1 aM1
Then, integrating the differential inequality (3.34) over [t0 , t], we obtain
 Z t Z +∞ 1 2 
−δ1 t δ1 t0 δ1 s
L(t) ≤ e e L(t0 ) + C7 e h(π) B u0 (π − s) dπds , ∀t ∈ R+ .
2
0 s

So, using (3.18) and (3.34), we get, for all t ≥ t0 ,

1
E(t) ≤ L(t)
1 − M
 
1
≤ max C7 , eδ1 t0 L(t0 ) ×
1 − M
 Z t Z +∞ 1 2 
× 1+ eδ1 s h(π) B 2 u0 (π − s) dπds . (3.35)

0 s

For t ∈ [0, t0 ], we have

1 1
E(t) ≤ L(t)eδ1 t e−δ1 t ≤ max L(s)eδ1 t0 e−δ1 t . (3.36)
1 − M 1 − M s∈[0,t0 ]
Inequalities (3.35) and (3.36) gives (3.21) with
 
1
δ2 = C7 , eδ1 t0 max L(s) .
1 − M s∈[0,t0 ]

I If limt→+∞ ξ(t) = 0, there exist t0 ≥ 0 such that ξ(t) ≤ C12 , for all t ≥ t0 .
Therefore, using (3.18), we obtain, for
 
1 C7 C7
δ1 = min , , ,
1 + M M1 aM1
Z +∞ 1 2
0
L (t) ≤ −δ1 ξ(t)L(t) + C8 h(s) B 2 u0 (s − t) ds, ∀t ∈ R+ , (3.37)

t
By integrating the above differential inequality over [t0 , t], we get, for all t ∈ R+ ,
 Z t Z +∞ 1 2 
−δ1 ξ̂(t) δ1 ξ̂(t0 ) δ1 ξ̂(s)
L(t) ≤ e e L(t0 ) + C7 e h(π) B u0 (π − s) dπds .
2
0 s
 STABILITY RESULT 23

Then, using (3.18) and (3.37), we get, for all t ≥ t0 ,

 
1
E(t) ≤ max C7 , eδ1 ξ̂(t0 ) L(t0 ) ×
1 − M
 Z t Z +∞ 1 2 
× 1+ eδ1 s h(π) B 2 u0 (π − s) dπds . (3.38)

0 s

For t ∈ [0, t0 ], we have

 
1 δ1 ξ̂(t) −δ1 ξ̂(t) 1 ˆ
E(t) ≤ L(t)e e ≤ max L(s)e δ1 ξ(s)
e−δ1 ξ̂(t) . (3.39)
1 − M 1 − M s∈[0,t0 ]
Inequalities (3.38) and (3.39) gives (3.21) with
  
1 ˆ
δ2 = C7 , max L(s)eδ1 ξ(s) .
1 − M s∈[0,t0 ]

Thus the proof of Theorem 3.6 is completed. 

4. Applications
We can seek our results in some problems. In this section, we consider only three
illustrative problems. In the whole section, Ω is a bounded and regular domain of
Rn , with n ≥ 1.
1-: Abstract linear problem
 R +∞
 utt (t) + Au(t) − 0 h(s)Au(t − s)ds + Cut (t − τ ) = 0, t ∈ (0, +∞),
u (t − τ ) = f0 (t − τ ), t ∈ (0, τ ),
 t
u(−t) = u0 (t), ut (0) = u1 , t ≥ 0,
(4.1)
where the operators A and C are a self-adjoint linear positive operators satisfy the
assumptions (A1) and (A3), respectively. The memory kernel h satisfying (A2) and
(A7).
2-: Let us consider the semilinear problem
 R +∞

 utt (t) + Au(t) + 0 h(s)∆u(t − s)ds + b(x)ut (t − τ )
= F (u(t)), t ∈ (0, +∞),



u(x, t) = 0, x ∈ ∂Ω, (4.2)
u(x, −t) = u0 (x, t), ut (x, 0) = u1 (x), x ∈ Ω, t ≥ 0,




ut (t − τ ) = f0 (t − τ ) t ∈ (0, τ ),


with initial data (u0 , u1 , f0 ) ∈ [H 2 (Ω) ∈ ∩H01 (Ω)] × H01 (Ω) × H 1 (0, τ ; L2 (Ω)). The
constant β > 0 satisfies a suitable restriction to be specified below. The memory
kernel h satisfying (A2) and (A7) and b ∈ L∞ (Ω) is a function such that
b(x) ≥ 0 a. e. in Ω.
The source term F be globally Lipschitz continuous functional such that F (0) = 0
and satisfies (3.2).Our results hold with H = L2 (Ω) and the operators A, B are
given by
n  
X ∂ ∂u
A : D(A) −→ H : u 7→ − aij (x) , B : D(B) −→ H : u 7→ −∆u,
i,j=1
∂xi ∂xj
24 HOURIA CHELLAOUA, YAMNA BOUKHATEM

where D(A) = D(B)) = H 2 (Ω) ∩ H01 (Ω). aij ∈ C 1 (Ω), is symmetric and
n
X
∃a0 > 0, aij (x)ζj ζi ≥ a0 |ζ|2 , x ∈ Ω, ζ = (ζ1 , · · · , ζn ) ∈ Rn .
i,j=1

The operators A and B are a linear,

self-adjoint and positive
operators in H such
1 12 21
1 1/2
that D(A ) = H0 (Ω) with A u = (a(u, u))
2 and B u = k∇uk2 , where
n Z
X ∂u ∂u
a(u, u) = aij (x) dx.
i,j=1 Ω ∂xi ∂xj

Moreover, by using Poincare’s inequality and the Sobolev’s embedding theorem, we

get (A1) and (A5). Then, the assumption (A3) holds with Cu(x, t) = b(x)u(x, t).
3-: Coupled systems
 Z +∞
 wtt (t) − α∆w(t) + h(s)div(a1 (x)∇w(t − s))ds




 0
+µwt (t − τ ) + dv(t) = f1 (w(t)), t ∈ (0, +∞),



 Z +∞


 vtt (t) − β∆v(t) + h(s)div(a2 (x)∇v(t − s))ds


0 (4.3)

 +µvt (t − τ ) + dw(t) = f2 (v(t)), t ∈ (0, +∞),
w(x, t) = v(x, t) = 0, x ∈ ∂Ω,




w(x, −t) = w (x, t), v(x, −t) = v (x, t), x ∈ Ω, t ≥ 0,



 0 0
w (x, 0) = w (x), v (x, 0) = v (x), x ∈ Ω, t ≥ 0,



 t 1 t 1
wt (t − τ ) = f0 (t − τ ), vt (t − τ ) = f0 (t − τ ), t ∈ (0, τ ),


where α and β are positive constants, a1 , a2 ∈ C 1 (Ω), a1 (x), a2 (x) > 0 with
The memory kernel h satisfying (A2) and (A7). The above system is equivalent to
(1.1) where u = (w, v), f0 = (l0 , m0 ) and H = (L2 (Ω))2 with
Z
h(w1 , v1 ), (w2 , v2 )i = w1 w2 + v1 v2 dx.
Ω

We take D(A) = D(B)) = (H 2 (Ω) ∩ H01 (Ω))2 and the operators A, B are given
by
Au = −(α∆w, β∆v) + d(v, w),
Bu = −(div(a1 (x)∇w), div(a2 (x)∇w)).
The function F2 (u(t)) = (f1 (w(t)), f2 (v(t))) satisfies (A6).

Acknowledgements
The authors are highly grateful to the anonymous referee for his/her valuable
comments and suggestions for the improvement of the paper.

References
[1] F. Alabau-Boussouira, S. Nicaise and C. Pignotti, Exponential stability of the wave equation
with memory and time delay, New Prospects in Direct, Inverse and Control Problems for
Evolution Equations. Springer, Cham, (2014), 1-22.
[2] M. Aassila, M. M. Cavalcanti and V. N. Domingos Cavalcanti, Existence and uniform decay
of the wave equation with nonlinear boundary damping and boundary memory source term,
Calc. Var. Partial Dif., 15 (2002), 155-180.
[3] A. Bàtkai, S. Piazzera, Semigroups for delay equations, Research Notes in Mathematics, 10.
AK Peters, Ltd., Wellesley, MA, (2005).
 STABILITY RESULT 25

[4] A. Benaissa, A. K. Benaissa and S. A. Messaoudi, Global existence and energy decay of
solutions for the wave equation with a time varying delay term in the weakly nonlinear
internal feedbacks, J. Math. Phys., 53 (2012), 1-19.
[5] Y. Boukhatem, B. Benabderrahmane, General decay for a viscoelastic equation of variable
coefficients with a time-varying delay in the boundary feedback and acoustic boundary
conditions, Acta Math. Sci., 37(5) (2017), 1453-1471.
[6] Y. Boukhatem, B. Benabderrahmane, General Decay for a Viscoelastic Equation of Variable
Coefficients in the Presence of Past History with Delay Term in the Boundary Feedback and
Acoustic Boundary Conditions, Acta Appl. Math., 154(1) (2018), 131-152.
[7] M. M. Cavalcanti, V. N. Domingos and J. A. Soriano, Exponential decay for the solution
of semilinear viscoelastic wave equations with localized damping, Elect. J. Diff. Equa., 44
(2002), 1-14.
[8] Q. Dai and Y. Zhifeng, Global existence and exponential decay of the solution for a vis-
coelastic wave equation with a delay.” Z. Angew. Math. Phys., 65(5) (2014), 885-903.
[9] CM. Dafermos, Asymptotic stability in viscoelasticity, Arch. ration. mech. anal., 37 (1970),
297-308.
[10] R. Datko, Two questions concerning the boundary control of certain elastic systems, J. Diff.
Equa., 1 (1991), 27-44.
[11] A. Guesmia, Well-posedness and exponential stability of an abstract evolution equation with
infinite memory and time delay, IMA J. Math. Control I., 30 (2013), 507-526.
[12] A. Guesmia, N. E.Tatar, Some well-posedness and stability results for abstract hyperbolic
equations with infinite memory and distributed time delay, Commun. Pur. Appl. Anal., 14
(2015), 457-491.
[13] M. Kirane and B. Said-Houari, Existence and asymptotic stability of a viscoelastic wave
equation with a delay, Z. Angew. Math. Phys., 62 (2011), 1065-1082.
[14] S. A. Messaoudi, General decay of solutions of a viscoelastic equation, J. Math. Anal. Appl.,
341 (2008), 1457-1467.
[15] S. Nicaise, C. Pignotti, Stability and instability results of the wave equation with a delay
term in the boundary or internal feedbacks, SIAM J. Control Optim., 45 (2006), 1561-1585.
[16] S. Nicaise and J. Valein, Stabilization of second order evolution equations with unbounded
feedback with delay, ESAIM Control Optim., 2 (2010), 420-456.
[17] S. Nicaise, C. Pignotti, Stabilization of second-order evolution equations with time delay,
Math. Control Signal., 26(4) (2014), 563-588.
[18] S. Nicaise, C. Pignotti, Exponential stability of abstract evolution equations with time delay,
J. Evol. Equ., 15 (2015), 107-129.
[19] S. Nicaise, C. Pignotti, Well-posedness and stability results for nonlinear abstract evolution
equations with time delays, J. Evol. Equ., 18 (2018), 947-971.
[20] V. Pata, Exponential stability in linear viscoelasticity, Quart. Appl. Math., 3 (2006), 499-513
[21] V. Pata, Exponential stability in linear viscoelasticity with almost flat memory kernels,
Commun. Pure. Appl. Anal., 9 (2010), 721-730.
[22] A. Pazy, Semigroups of Linear Operators and Applications to Partial Differential Equations,
277. Springer, New York (1983).
[23] N.E. Tatar, A new class of kernels leading to an arbitrary decay in viscoelasticity, Mediterr.
J. Math., 10(1) (2013), 213-226.
[24] N. E. Tatar, On a large class of kernels yielding exponential stability in viscoelasticity, Appl.
Math. Comp., 215 (2009), 2298-2306.
[25] N. E. Tatar, Uniform decay in viscoelasticity for kernels with small non-decreasingness zones,
Appl. Math. Comp., 218 (2012), 7939-7946.

Houria Chellaoua,
Laboratory of Pure and Applied Mathematics, University of Laghouat, P.O. BOX 37G,
Laghouat (03000), Algeria.
Email address: h.chellaoua@lagh-univ.dz

Yamna Boukhatem,
Laboratory of Pure and Applied Mathematics, University of Laghouat, P.O. BOX 37G,
Laghouat (03000), Algeria.
Email address: y.boukhatem@lagh-univ.dz

View publication stats