Polynomial stability of Timoshenko system with

strong damping and strong distributed delay terms

Imene Laribi
Université Djilali Bounaama Khemis Miliana
Ali Krelifa
Université Djilali Bounaama Khemis Miliana
Salah Zitouni
University Mohamed-Cherif Messaadia of Souk Ahras
Djamel Ouchenane (  [email protected] )
Laghouat University

Research Article

Keywords: Strong damping, Polynomial stability, Timoshenko system, delay term

Posted Date: November 15th, 2023

DOI: https://doi.org/10.21203/rs.3.rs-3597010/v1

License:   This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License

Additional Declarations: No competing interests reported.

Polynomial stability of Timoshenko system with
strong damping and strong distributed delay
terms
Imene Laribi∗,1 , Ali Krelifa†1 , Salah Zitouni‡2 , and Djamel
Ouchenane§,3
1
Department of Mathematics, University of Khemis Miliana, Road of Theniet
El-Had, Khemis Miliana 44225,
,1
Laboratory of Mathematics and Computer Science Laghouat, Algeria
2
Department of Mathematics and Informatics, University Mohamed-Cherif
Messaadia of Souk Ahras, P.O. Box 1553, Souk Ahras 41000, Algeria
3
Laboratory of Pure and Applied Mathematics, Laghouat University, Laghouat,
Algeria

November 11, 2023

Abstract
In this research work, we study the linear Timoshenko problem with
a strong damping and a strong distributed delay when it acts on the first
equation. We first prove the existence and uniqueness solution of the
system by using the semi-group theory. Then, we illustrate the lack of
exponential stability even in the condition of equal wave propagation is
satisfied. Furthermore, we show the polynomial decay of solution with
1
rate t− 2 . From this perspective, we demonstrate the optimality of the
rate.

Keywords: Strong damping, Polynomial stability, Timoshenko system,

delay term.

1 Introduction
In this article, we are interested in Timoshenko beam model, which repre-
sents shear deformations and rotational bending effects on the beam [3].
∗ [email protected],
† [email protected],
‡ [email protected],
§ [email protected].

1
It is given by the following system (see[15, 8])

ρAϕtt − Px = 0, (0, 1) × (0, ∞) ,
(1.1)
ρIψtt − Wx + P = 0, (0, 1) × (0, ∞) ,

where ϕ is the transverse displacement of the beam, ψ is the rotational

angle. The functions P and W are defined by

P = KGA (ϕx + ψ) , W = EIψx

The coefficient ρ is the mass per unit length, A is the cross-section area,
I is the moment of inertia of a cross section, E is Youngs modulus of
elasticity, K the shear modulus and G is modulus of rigidity. We denote
ρ1 = ρA, ρ2 = ρI, κ = KGA, b = EI. Then (1.1) becomes

ρ1 ϕtt − κ (ϕxx + ψx ) = 0, (0, 1) × (0, ∞) ,
(1.2)
ρ2 ψtt − bψxx + κ (ϕx + ψ) = 0, (0, 1) × (0, ∞) ,

Many authors studied this kind of problems adding delay or damping

terms. Under similar conditions, they proved the exponential and polyno-
mial stability of their problems. More precisely, the writters in [11] have
shown the well-posedness and the exponential stability of system (1.2).
In [1], Aparala considered a damped Timoshenko system with second
sound and internal distributed delay. Also, he proved the existence and
uniquness of the solution and showed the exponential stability.
In [10], Messaoudi, Abdelfeteh and Doudi studied the well-posedness
of the following wave equation in bounded and regular domain of Rn .
They took damping and delay terms

wtt − ∆w − γ1 ∆wt − γ2 ∆wt (x, t − ξ) = 0, Ω × (0, ∞) , (1.3)

Under similar assumptions, they demonstrated the exponential stability.

On the other hand, the authors in [2] examined a coupled wave system
with past history, then they have shown the polynomial decay of the
1
solution with an optimal rate equals to t− 2 . In [9], the authors studied
the Timoshenko system below. They added strong damping and delay

 ρ1 ϕtt − κ (ϕx + ψ)x − µ1 ϕxxt (0, 1) × (0, ∞) ,
−µ2 ϕxxt (x, t − τ ) = 0, (1.4)

ρ2 ψtt − bψxx + κ (ϕx + ψ) = 0, (0, 1) × (0, ∞) ,

They studied the well posedness, and have shown the lack of exponential
stability in both conditions ρκ1 = ρb2 and ρκ1 6= ρb2 . Next, they proved the
1
polynomial stability with rate t− 2 . Moreover, this rate is optimal. In [4],
Fatori and Munoz Rivera studied the thermoelastic Bresse system, and
proved the equivalence between the exponential stablilty of their problem
and the condition ρκ1 = ρb2 . Next, they gave a counter-example that have
shown the explosion of solution when ρκ1 6= ρb2 . Then they demonstrated
the plynomial decay and optimality. For comprehensive insights, refer to
references [5, 7, 12, 13, 16].
Inspiriting by previous work, in our article, we study the Timoshenko

2
system below with a strong damping and a strong distributed delay when
it acts on the first equation:



 Rρτ1 ϕtt − κ (ϕx + ψ)x − µ1 ϕxxt (0, 1) × (0, ∞) ,


 − τ12 |µ2 (p) |ϕxxt (x, t − p) dp = 0,


 ρ2 ψtt − bψxx + κ (ϕx + ψ) = 0, (0, 1) × (0, ∞) ,
ϕ (0, t) = ϕ (1, t) = ψx (0, t) = ψx (1, t) = 0 (0, ∞) , (1.5)



 ϕ (x, 0) = ϕ0 (x) , ϕt (x, 0) = ϕ1 (x) , (0, 1) ,



 ψ (x, 0) = ψ0 (x) , ψt (x, 0) = ψ1 (x) , (0, 1) ,

ϕ (x, −t) = f0 (x, −t) , t ∈ (0, τ2 ) .

where ρ1 , κ, ρ2 , b, τ1 , τ2 , µ1 are positive constants, and µ2 is a bounded

function on [τ1 , τ2 ] with values in R. It verifies the following assumption
Z τ2
|µ2 (s)| ds < µ1 . (1.6)
τ1

As we have Neumann boundary conditions for ψ, we can’t apply Poincare’s

inequality for it. Thus we shall introduce a new variable ψ such that

int10 ψ(x, t)dx = 0. (1.7)

For this, let us use (1.5)2 and (1.5)3 to obtain

Z 1 Z 1
d2 κ
ψ(x, t)dx + ψ(x, t)dx = 0, (1.8)
dt2 0 ρ2 0

The solution of (1.8) with (1.5)5 gives

Z 1 Z 1  r r Z 1  r
κ ρ2 κ
ψ(x, t)dx = ψ0 (x)dx cos t+ ψ1 (x)dx sin t,
0 0 ρ 2 κ 0 ρ 2

So, we let Z 1
ψ(x, t) = ψ(x, t) − ψ(x, t)dx
0
to get (1.7).
Thus, we justify the application of Poincare’s inequality for ψ and (ϕ, ψ)
verifies (1.5) with the following initial conditions
 R1
ψ 0 (x) = ψ0 (x) − 0 ψ0 dx,
R1
ψ 1 (x) = ψ1 (x) − 0 ψ1 dx.

For the rest of this paper, we use ψ rather than ψ, however we set ψ for
simplification.
Now, we start by showing the existence and uniquness of the solution
of system (1.5). Then, we show that is not exponentially stable. Next, we
demonstrate the decay of the energy associated with the system, then we
construct a suitable Lyapunov functional, and establish the polynomial
1
stability with rate equals to t− 2 . we finish by showing the optimality of
that rate.

3
2 Existence and uniquness of solution
This section consists to show the well posedness of the problem (1.5) using
the semigroup theory.
To simplify our calculations, we start by giving another variable z such
that [1, 10]

z(x, ̺, τ, t) = ϕt (x, t − ̺τ ),
Then, z satisfies

τ zt (x, ̺, τ, t) + z̺ (x, ̺, τ, t) = 0,

and
z(x, 0, τ, t) = ϕt (x, t). (2.1)
Thus, we can rewrite the system (1.5) as

 ρ1 ϕtt − κ (ϕx + ψ)x − µ1 ϕxxt
 − R τ2 |µ (τ ) |z (x, 1,

τ, t)dτ = 0,
τ1 2 xx
(2.2)

 ρ 2 ψ tt − bψ xx + κ (ϕ x + ψ) = 0,

τ zt + z̺ = 0,

with initial and boundary conditions



 ϕ (0, t) = ϕ (1, t) = ψx (0, t) = ψx (1, t) = 0,

ϕ (x, 0) = ϕ0 (x) , ϕt (x, 0) = ϕ1 (x) ,
(2.3)

 ψ (x, 0) = ψ0 (x) , ψt (x, 0) = ψ1 (x) ,

z(x, ̺, τ, 0) = f0 (x, −τ ̺) ,

where (x, ̺, τ, t) ∈ (0, 1) × (0, 1) × (τ1 , τ2 ) × (0, ∞).

Now, let us consider the vector Ü

Ü = (ϕ, ϕt , ψ, ψt , z),

for simplification, we set u rather than ϕt , and v rather than ψt .

Then, we give these spaces below
n R1 o
L2⋄ (0, 1) = g ∈ L2 (0, 1), 0 g(x)dx = 0 ,
H⋄1(0, 1) = H 1 (0, 1) ∩ L2⋄ (0, 1),
H⋄2 (0, 1) = g ∈ H 2 (0, 1), gx (0) = gx (1) = 0 ,

Now, we define the following Hilbert space

H = H01 (0, 1)×H01 (0, 1)×H⋄1 (0, 1)×L2⋄ (0, 1)×L2 (0, 1) × (τ1 , τ2 ), H01 (0, 1) ,

endowed with this product

R1 R1 R1 R1
(Ü , U ) = ρ1 0 uudx + ρ2 0 vvdx + b 0 ψx ψ x dx + k 0 (ϕx + ψ)(ϕx + ψ)dx
R 1 R 1 R τ2
+ 0 0 τ1 τ |µ2 (τ )|zx z x dτ d̺dx

Consequently, the system (2.2) is equivalent to


Ü ′ (t) = AÜ ,
(2.4)
Ü (0) = Ü0 = (ϕ0 , ϕ1 , ψ0 , ψ1 , f0 )T ,

4
where A : D(A) ⊂ H → H is a linear operator given by
 
h u i
 1 κ(ϕ + ψ ) + µ u + R τ2 |µ (τ )|z (x, 1, τ, t)dτ 
 ρ1 xx x 1 xx τ1 2 xx 
 
AÜ =  v 
 1


ρ2
[bψxx − κ(ϕx + ψ)] 
1
− τ z̺

And the domaine of A is

 h Rτ i 
 Ü ∈ H| κϕxx + µ1 uxx + τ12 |µ2 (τ )|zxx (x, 1, τ, t)dτ ∈ H01 (0, 1),
 

D(A) = ψ ∈ H⋄2 (0, 1) ∩ H⋄1 (0, 1), u ∈ H01 (0, 1), v ∈
H⋄1 (0, 1),
 
z, z̺ ∈ L2 (0, 1) × (τ1 , τ2 ), H01 (0, 1) .
 

Then , we have the following lemmas.

Lemma 2.1. Under assumption (1.6), A is a dissipative operator in H.

Proof. The goal is to demonstrate (AÜ , Ü )H < 0. Then, by the inner

product (., .)H and using integration by part, we find
R1 R1 Rτ
(AÜ , Ü )H = −µ1 0 u2x (x, t)dx − 0 ux (x, t) τ12 |µ2 (τ )|zx (x, 1, τ, t)dτ dx
R R  R R
1 τ τ 1
− 21 0 τ12 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx + 12 τ12 |µ2 (τ )|dτ 0 u2x (x, t)dx,
(2.5)
Furthermore, by using Young’s inequality, we obtain
R1 Rτ R1Rτ
− 0 ux (x, t) τ12 |µ2 (τ )|zx (x, 1, τ, t)dτ dx ≤ 21 0 τ12 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx
 R R
τ 1
+ 12 τ12 |µ2 (τ )|dτ 0 u2x (x, t)dx,
(2.6)
Now, replacing (2.6) into (2.5) and recalling (1.6), we deduce
 Z τ2 Z 1
(AÜ , Ü )H = − µ1 − |µ2 (τ )|dτ u2x (x, t)dx < 0.
τ1 0

Lemma 2.2. Assuming that (1.6) holds. Then, The surjectivity of I − A

is established.

Proof. It suffices to demostrate, ∀ F = (f1 , f2 , f3 , f4 , f5 ) ∈ H, ∃! Ü ∈

D(A) such that
Ü − AÜ = F,
In another way, we write


 h ϕ − u = f1 , i

 1
R τ2
 u − ρ1 κ(ϕx + ψ)x + µ1 uxx + τ1 |µ2 (τ )|zxx (x, 1, τ, t)dτ = f2 ,


 ψ − v = f3 ,
 1


 v− ρ2
[bψxx − κ(ϕx + ψ)] = f4 ,

z + τ1 z̺ = f5 ,
(2.7)

5
The equations (2.7)1 , (2.7)3 give

u = ϕ − f1 ,
(2.8)
v = ψ − f3 ,

Also, the equation (2.7)5 with z(x, 0, τ ) = u and (2.8)1 gives

Z ̺
z(x, ̺, τ ) = (ϕ − f1 )e−τ ̺ + τ e−τ ̺ eτ s f5 (x, s, τ )ds (2.9)
0

Now, we replace (2.8)1 , (2.9) in (2.7)2 , and (2.8)2 in (2.7)4 , we obtain

 h R τ2
1 −τ
 ϕ − f1 − ρ1 κ(ϕx + ψ)x + µ1 (ϕxx − f1,xx ) + τ1 |µ2 (τ )|i (ϕxx − f1,xx ) e dτ


R τ2 R 1
 + τ1 |µ2 (τ )|τ e−τ 0 eτ s f5,xx (x, s, τ )dsdτ = f2 ,

 ψ − f3 − ρ12 [bψxx − κ(ϕx + ψ)] = f4
(2.10)
Then, (2.10) becomes
  Rτ 


 ρ1 ϕ − κ(ϕx + ψ)x − µ1 + τ12 e−τ |µ2 (τ )|dτ ϕxx = ρ1 (f1 + f2 )
 Rτ  Rτ R1
 − µ1 + τ12 e−τ |µ2 (τ )|dτ f1,xx + τ12 |µ2 (τ )|τ e−τ 0 eτ s f5,xx (x, s, τ )dsdτ,


ρ2 ψ − bψxx + κ(ϕx + ψ) = ρ2 (f3 + f4 ) ,
(2.11)
Next, we solve the problem (2.11). Thus, we shall obtain the weak for-
mulation of the problem. For this work, let us multiply (2.11)1 by ϕ
and (2.11)2 by ψ, then integrate over (0, 1), we get
 R1  R τ2 −τ  
 0 ρ1 ϕϕ + κ(ϕx + ψ)ϕx+ µ1 + τ1 e |µ2 (τ )|dτ

 R
ϕx ϕx dx =

 R1 Rτ 1
ρ (f1 + f2 ) ϕdx + µ1 + τ12 e−τ |µ2 (τ )|dτ 0 f1,x ϕx dx
0 1
R R R
1 1 τ



 − 0 0 τ12 |µ2 (τ )|τ eτ (s−1) f5,x (x, s, τ )ϕx dτ dsdx,
 R1
R1
0
ρ2 ψψ + bψx ψ x + κ(ϕx + ψ)ψ dx = 0 ρ2 (f3 + f4 ) ψdx,
Adding the two equations, then we rewrite the system as



B Φ, Φ = L Φ , (2.12)

where
Φ = (ϕ, ψ), Φ = (ϕ, ψ)
and

R1

B Φ, Φ = 0 ρ1 ϕϕ + ρ2 ψψ + κ(ϕx +ψ)(ϕx + ψ) + bψx ψ x dx
Rτ R1
+ µ1 + τ12 e−τ |µ2 (τ )|dτ 0 ϕx ϕx dx,

and

R1 R1  Rτ R
1
L Φ = 0 ρ1 (f1 + f2 ) ϕdx + 0 ρ2 (f3 + f4 ) ψdx + µ1 + τ12 e−τ |µ2 (τ )|dτ 0 f1,x ϕx dx
R 1 R 1 R τ2
− 0 0 τ1 |µ2 (τ )|τ eτ (s−1) f5,x (x, s, τ )ϕx dτ dsdx.

Now, we consider G = H01 (0, 1)×H⋄1 (0, 1) endowed with this inner product
Z 1 Z 1


Φ, Φ G = b ψx ψ x dx + κ (ϕx + ψ)(ϕx + ψ)dx
0 0

6
Clearly B is a bilinear form on G×G, and L is linear form on G. Moreover,
B is continuous and coercive on G × G, and L is continuous on G. In-
deed, ∀ Φ ∈ G, ∀ Φ ∈ G, and by using Cauchy-Schwartz’s and Poincare’s
inequalities, we have


B Φ, Φ ≤ − max(M, ρ2 ) kϕx + ψk ϕx + ψ + κ kϕx + ψk ϕx + ψ
+b kψx k ψ x ,

where M = ρ2 c2 + µ1 + Akµ2 kL∞ (τ1 ,τ2 ) , A = e−τ1 − e−τ2 > 0. Thus, by

inserting H = κ − max(M, ρ2 ), then we deduce



B Φ, Φ ≤ max(H, b) (kϕx + ψk + kψx k) ϕx + ψ + ψ x
≤ max(H, b) kΦkG Φ G .

Similary, by using the same steps above, we deduce that for all Φ =
(ϕ, ψ) ∈ G, we obtain

L Φ ≤C Φ G.
Finally, ∀ Φ ∈ G, we have

R1
R1 R1 2
B Φ, Φ = 0 ρ1 ϕ2 + ρ2 ψ 2 dx + κ 0 (ϕ 2
 Rx + ψ) dx + b 0 ψx dx
R τ2 −τ 1
+ µ1 + τ1 e |µ2 (τ )|dτ 0 ϕ2x dx,

Now, note that ρ1 ϕ2 + ρ2 ψ 2 and e−τ are positive terms, then using (1.6),
we find
 
R1 R1  Rτ R1
B (Φ, Φ) > κ 0
(ϕx + ψ)2 dx + bψx2 dx + 2 τ12 |µ2 (τ )| dτ  0 ϕ2x dx
0
| {z } |{z}
≥0 ≥0
R1 R 1
> κ 0 (ϕx + ψ)2 dx + b 0 ψx2 dx = k(ϕ, ψ)k2G .

Consequently, by applying Lax-Milgram theorem, we deduce that the sys-

tem (2.12) have a unique solution

Φ = (ϕ, ψ) ∈ G

Now, recalling F ∈ H and (2.8), we find

ϕt ∈ H01 (0, 1), ψt ∈ H⋄1 (0, 1),

Similary, by (2.9) and (2.7)5 , we get

z, z̺ ∈ L2 (0, 1) × (τ1 , τ2 ), H01 (0, 1) ,

Now, if ϕ ≡ 0 ∈ H01 (0, 1), then (2.12) becomes

Z 1 Z 1
b ψx ψ̄x dx = [ρ2 (f3 + f4 ) − ρ2 ψ − κ (ϕx + ψ)] ψ̄dx, ∀ψ̄ ∈ H⋄1 (0, 1) .
0 0
(2.13)
Here, we can’t use the regularity theorem, because ψ̄ ∈ H⋄1 (0, 1) . So, we
take Ψ̄ ∈ H01 (0, 1) and set
Z 1
ψ̄ = Ψ̄ (x) − Ψ̄ (x) dx (2.14)
0

7
It is clear that ψ̄ ∈ H⋄1 (0, 1) . Then a substitution in (2.13) leads to
Z 1 Z 1
b ψx Ψ̄x dx = rΨ̄dx, ∀Ψ̄ ∈ H01 (0, 1) , (2.15)
0 0

where
r = −(κ + ρ2 )ψ − κϕx + ρ2 (f3 + f4 ) ∈ L2⋄ (0, 1) (2.16)
2
So ψ ∈ H (0, 1) and
Z 1 Z 1
b ψxx φdx = rφdx, ∀φ ∈ H 1 (0, 1) , (2.17)
0 0

by using integration by part and reminder (2.13), we find

ψx (1)ψ̄(1) − ψx (0)ψ̄(0) = 0

Then, ψx (1) = ψx (0) = 0. Finally, we conclude that

ψ ∈ H⋄2 (0, 1) ∩ H⋄1 (0, 1).

With the same manner, if we let ψ ≡ 0 ∈ H⋄1 (0, 1), then (2.12) becomes
R1  Rτ R
1
0
(ρ1 ϕϕ + κ(ϕx + ψ)ϕx ) dx + µ1 + τ12 e−τ |µ2 (τ )|dτ 0 ϕx ϕx dx
R1  Rτ  R1
= 0 ρ1 (f1 + f2 ) ϕdx + µ1 + τ12 e−τ |µ2 (τ )|dτ 0 f1,x ϕx dx
R 1 R 1 R τ2
− 0 0 τ1 |µ2 (τ )|τ eτ (s−1) f5,x (x, s, τ )ϕx dτ dsdx, ∀ϕ ∈ H01 (0, 1)

by using integration par part, we get

R1
 0
[ρ1 (ϕ − f1 ) − ρ1 f2 − κϕxx − κψx − µ1 (ϕxxi − f1,xx )
Rτ R1
− τ12 |µ2 (τ )| e−τ (ϕxx − f1,xx ) + τ e−τ 0 eτ s f5,xx (x, s, τ )dsdτ ϕdx = 0, ∀ϕ ∈ H01 (0, 1)
(2.18)
Now, recalling (2.8)1 and (2.9). Then, (2.18) is equivalent to
Z τ2
κϕxx + µ1 uxx + |µ2 (τ )|zxx (x, 1, τ, t)dτ = ρ1 u − ρ1 f2 − κψx ∈ H01 (0, 1)
τ1

Consequently
Z τ2
κϕxx + µ1 uxx + |µ2 (τ )|zxx (x, 1, τ, t)dτ ∈ H01 (0, 1).
τ1

According to above, we conclude that Ü ∈ D(A) . Therfore, we establish

the surjectivity of I − A.

Theorem 2.3. Let Ü0 ∈ H, and assume that (1.6) is satisfied. Then the
problem (2.4) admits a unique solution Ü ∈ C(R+ , H). Furthermore, if
Ü0 ∈ D(A), then
Ü ∈ C(R+ , D(A)) ∩ C 1 (R+ , H).

Proof. The proof is done by Lemmas 1,2 and applying Lumer-Phillips

theorem.

8
3 Lack of exponential stability
in this part, we will show that the solution of our problem is not exponen-
tially stable even if the condition of equal wave propagation is satisfied.
As in [9, 4, 2], we use the theorem 1.1 in [4] (see [14, 6]).

Now, by following the same steps used in [9, 4], we establish our goal.
More precisely, we show that there exists a subsequence (λυ )υ∈N ⊂ R such
that
k(iλυ I − A)−1 kL(H) −→ ∞,
υ→∞

that is to say we demonstrate the existence of a subsequence (λυ )υ∈N ⊂ R

and (m athrmFυ )υ∈N ⊂ H, with kFυ k ≤ 1, such that

k(iλυ I − A)−1 Fυ kH = kÜυ kH −→ ∞,

υ→∞

Thus, we have the following spectral equation

(iλυ I − A)Üυ = Fυ , (3.1)

Therefore, we demonstrate the lemma below.

Lemma 3.1. The solution of our problem is not exponentially stable.

Proof. The goal is to show that the solution Üυ corresponding to the
problem (3.1) is not bounded when Fυ is bounded in H. Indeed, (3.1) is
equivalent to


 h iλϕ − u = f1 , i

 1
R τ2
 iλu − ρ1 κ(ϕx + ψ)x + µ1 uxx + τ1 |µ2 (τ )|zxx (x, 1, τ )dτ = f2 ,


 iλψ − v = f3 ,
 1


 iλv − ρ2
[bψxx − κ(ϕx + ψ)] = f4 ,

iλτ z + z̺ = τ f5 ,
(3.2)
where λ ∈ R and F = (f1 , f2 , f3 , f4 , f5 ) ∈ H.
As in [9], we choose f1 = f2 = f3 = f5 = 0. Then we obtain

u = iλϕ,
(3.3)
v = iλψ,

Remplacing (3.3)1 in (3.2)2 and (3.3)2 in (3.2)4 , we obtain

 2 κ+iλµ κ 1
R τ2
 −λ ϕ − ρ1 1 ϕxx − ρ1 ψx − ρ1 τ1 |µ2 (τ )|zxx (x, 1, τ )dτ = 0,
−λ2 ψ − ρb2 ψxx + ρκ2 ϕx + ρκ2 ψ = f4 ,

iλτ z + z̺ = 0,
(3.4)
Now, we choose f4 (x) = cos(υπx), and the boundary conditions lead us
to choose

ϕ(x) = A sin(υπx), ψ(x) = B cos(υπx), z(x, ̺, τ ) = δ(̺, τ ) sin(υπx),

9
where A, B are two real numbers depending of λ, and δ(̺, τ ) depends of
̺, τ and λ. Therfore, the system (3.4) becomes
 h i
2 κ+iλµ1 2 κ(υπ) (υπ)2 R τ2
 −λ + ρ1 h(υπ) A + ρ1 B i+ ρ1 τ1 |µ2 (τ )|δ(1, τ )dτ = 0,


2

 −λ2 + b(υπ)
ρ2
+ ρκ2 B + κ(υπ) ρ2
A = 1,


iλτ δ(̺, τ ) + δ̺ (̺, τ ) = 0,
(3.5)
By solving the differential equation (3.5)3 and remember that z(x, 0, τ ) =
u, we get
δ(̺, τ ) = iλAe−iλτ ̺ .
Thus, (3.5) becomes
 h 2 R
i
 −λ2 + κ+iλµ1 (υπ)2 + iλ(υπ) τ2 |µ2 (τ )|e−iλτ dτ A + κ(υπ) B = 0,
ρ1 ρ1 τ 1 ρ1
h 2
i
 −λ2 + b(υπ)
ρ2
+ κ
ρ2
B + κ(υπ)
ρ2
A = 1,
(3.6)
Next, we opt λ = λυ such that
s r
b(υπ)2 κ b κ
−λ2υ + =− =⇒ |λυ | = υπ + −→ ∞
ρ2 ρ2 ρ2 ρ2 υ→∞
Then, (3.6)2 leads to
ρ2
A= −→ 0,
κ(υπ) υ→∞
Next, by substituting the expression of A in (3.6)1 , we find
  Z
ρ1 ρ2 b κ ρ1 iρ2 µ1 λυ iλυ ρ2 τ2
B= 2 − + − − |µ2 (τ )|e−iλυ τ dτ,
κ ρ2 ρ1 κ(υπ)2 κ2 κ2 τ1
b κ b κ
It should be noted that |B| −→ ∞ in the two cases ρ2
= ρ1
and ρ2
6= ρ1
.
υ→∞
Then, we have
Z 1 Z 1
ρ2 2
kÜυ k2H ≥ ρ2 |vυ |2 dx = ρ2 λ2υ |B|2 | cos(υπx)|2 dx = λυ |B|2 −→ ∞
0 0 2 υ→∞

In conclusion, the solution is not exponentially stable.

4 Polynomial stability
In this section, we start by showing the decay of the energy, then we
introduce a Lyapunov functional. Next, we show the polynomial stability
1
with rate equals to t− 2 by using the energy functional, and we finish by
the optimality of this rate.
Lemma 4.1. The energy of the system (2.2)- (2.3) is defined as
R1 
E(t) = 21 0 ρ1 ϕ2t + ρ2 ψt2 + bψx2 + κ(ϕx + ψ)2 dx
R 1 R 1 R τ (4.1)
+ 12 0 0 τ12 τ |µ2 (τ )|zx2 (x, ̺, τ, t)dτ d̺dx,
which satisfies Z 1
E ′ (t) ≤ −η0 ϕ2xt < 0, (4.2)
0
R τ2
where η0 = µ1 − τ1
|µ2 (p)|dp > 0.

10
Proof. Multiplying (2.2)1 by ϕt and (2.2)2 by ψt , then integrating by
part over (0, 1), and using the boundary conditions, we find
R1  R1
1 d
2 dt 0
ρ1 ϕ2t + ρ2 ψt2 + bψx2 + κ(ϕx + ψ)2 dx + µ1 0 ϕ2xt dx
R1 R τ2 (4.3)
+ 0 ϕxt τ1 |µ2 (τ )|zx (x, 1, τ, t)dτ dx = 0.

Next multiplying (2.2)3 by −zxx |µ2 (τ )| and integrating over (0, 1)×(0, 1)×
(τ1 , τ2 ), and recalling (2.1), we get
R 1 R 1 R τ2 R1R1Rτ
1 d
2 dt 0 0 τ1  2
τ |µ (τ )|zx2 (x, ̺, τ, t)dτ d̺dx = − 0 0 τ12 |µ2 (τ )|zx z̺x (x, ̺, τ, t)dτ d̺dx
Rτ  R1 R1Rτ
= 21 τ12 |µ2 (p)|dp 0 ϕ2xt dx − 12 0 τ12 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx.
(4.4)
By adding (4.4) to (4.3), then applying Young’s inequality, we obtain
 Z τ2 Z 1
E ′ (t) ≤ − µ1 − |µ2 (p)|dp ϕ2xt dx, (4.5)
τ1 0
R τ2
we establish (4.2) by inserting η0 = µ1 − τ1
|µ2 (p)|dp and recalling (1.6).

Now, we introduce a suitable functionals to construct a lyapunov func-

tional.
Lemma 4.2. For any c, ǫ1 > 0, the functional G1 defined by
Z 1 Z 1 Z
µ1 1 2
G1 (t) = ρ1 ϕt ϕdx + ρ2 ψt ψdx + ϕx dx, (4.6)
0 0 2 0
satisfies

R1 R1
G′1 (t) < − κ2 + µ1 ǫ1 0 (ϕx + ψ)2 dx + (cµ1 ǫ1 − b) 0 ψx2 dx
R1 2 R1 2 R R
1 1 τ2 2 (4.7)
+ρ1 c 0 ϕxt dx + ρ2 0 ψt dx + 2 0 τ1 |µ2 (τ )|zx (x, 1, τ, t)dτ dx.

Proof. By differentiating (4.6) and using (2.2)1 , (2.2)2 then using integra-
tion by part, we find
R1 R1 R1 R1
G′1 (t) = −κ 0 (ϕx + ψ)2 dx − b 0 ψx2 dx + ρ1 0 ϕ2t dx + ρ2 0 ψt2 dx
R1 R τ2
− 0 ϕx τ1 |µ2 (τ )|zx (x, 1, τ, t)dτ dx,

By applying Young’s and Poincare’s inequalities and (1.6), we obtain

R1 Rτ R1 R1Rτ
− 0 ϕx τ12 |µ2 (τ )|zx (x, 1, τ, t)dτ dx < µ12ǫ1 0 ϕ2x dx + 12 0 τ12 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx,
R1 R1
ρ1 0 ϕ2t dx ≤ ρ1 c 0 ϕ2xt ,
Thus
R1 R1 R1 R1 R1
G′1 (t) < − κ2 0
(ϕx + ψ)2 dx − b 0 ψx2 dx + cρ1 0 ϕ2xt dx + ρ2 0 ψt2 dx + µ1 ǫ1
2 0
ϕ2x dx
R R
1 τ
+ 12 0 τ12 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx.

Finally, by using the following inequality,

Z 1 Z 1 Z 1 Z 1
ϕ2x dx = (ϕx + ψ − ψ)2 dx ≤ 2 (ϕx + ψ)2 dx + 2c ψx2 dx,
0 0 0 0

we obtain (4.7).

11
Lemma 4.3. Let the functional G2 defined by
Z 1
G2 (t) = −ρ2 ψt ψdx. (4.8)
0

Then, for any ǫ2 > 0 and c1 > 0, G2 satisfies the following estimate
Z Z 1 Z 1
κǫ2 1
G′2 (t) < (ϕx + ψ)2 dx + c1 ψx2 dx − ρ2 ψt2 dx. (4.9)
2 0 0 0

Proof. By differentiating (4.8) and using (2.2)2 , we find

Z 1 Z 1 Z 1
G′2 (t) = −ρ2 ψt2 dx + b ψx2 dx + κ ψ(ϕx + ψ)dx,
0 0 0

Now, applying Young’s and Poincare’s inequalities to find

Z 1 Z 1 Z Z 1
κǫ2 1 κc
G′2 (t) < −ρ2 ψt2 dx + b ψx2 dx + (ϕx + ψ)2 dx + ψx2 dx,
0 0 2 0 2ǫ2 0
κc
by inserting c1 = b + 2ǫ2
> 0, we establish (4.9).

Lemma 4.4. The functional G3 defined by

Z 1 Z 1 Z τ2
G3 (t) = τ e−τ ̺ |µ2 (τ )|zx2 (x, ̺, τ, t)dτ d̺dx (4.10)
0 0 τ1

establishes for η1 > 0,

R1R1Rτ R1
G′3 (t) < −η1 0 0 τ12 τ |µ2 (τ )|zx2 (x, ̺, τ, t)dτ d̺dx + µ1 0 ϕ2xt dx
R 1 R τ2
−η1 0 τ1 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx.
(4.11)

Proof. we differentiate (4.10) to arrive at

R1R1Rτ
G′3 (t) = −2 0 0 τ12 e−τ ̺ |µ2 (τ )|zx zx̺ (x, ̺, τ, t)dτ d̺dx
R 1 R 1 R τ2 d −τ ̺

= − 0 0 τ1 d̺ e |µ2 (τ )|zx2 (x, ̺, τ, t) dτ d̺dx
R 1 R 1 R τ2 −τ ̺
− 0 0 τ1 τ e |µ2 (τ )|zx2 (x, ̺, τ, t)dτ d̺dx
R 1 R τ2

= − 0 τ1 |µ2 (τ )| e−τ zx2 (x, 1, τ, t) − ϕ2tx dτ dx
R 1 R 1 R τ2 −τ ̺
− 0 0 τ1 τ e |µ2 (τ )|zx2 (x, ̺, τ, t)dτ d̺dx.

Recalling (2.1), then using the fact that e−τ ≤ e−τ ̺ ≤ 1, ∀̺ ∈ (0, 1), we
obtain
Rτ R1 R1Rτ
G′3 (t) < τ12 |µ2 (p)|dp 0 ϕ2xt dx − 0 τ12 e−τ |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx
R1R1Rτ
− 0 0 τ12 τ e−τ |µ2 (τ )|zx2 (x, ̺, τ, t)dτ d̺dx.

Furthermore, we have −e−τ ≤ −e−τ2 , ∀τ ∈ [τ1 , τ2 ]. Then, by inserting

η1 = e−τ2 and using (1.6), we complete the proof.

12
Lemma 4.5. we give the second-order energy as follows
R1 
E1 (t) = 12 0 ρ1 ϕ2xt + ρ2 ψxt
2
+ bψxx2
+ κ(ϕxx + ψx )2 dx
R R R
1 1 τ 2 (4.12)
+ 12 0 0 τ12 τ |µ2 (τ )|zxx (x, ̺, τ, t)dτ d̺dx,

which satisfies the following estimate

Z 1
E1′ (t) ≤ −η0 ϕ2xxt < 0, (4.13)
0
R τ2
where η0 = µ1 − τ1
|µ2 (p)|dp > 0.

Proof. Multiplying (2.2)1 by −ϕxxt and (2.2)2 by −ψxxt , then integrating

by part over (0, 1), we find
R1  R1
1 d
2 dt 0
ρ1 ϕ2xt + ρ2 ψxt
2
+ bψxx2
+ κ(ϕx + ψ)2x dx + µ1 0 ϕ2xxt dx
R1 R τ2
+ 0 ϕxxt τ1 |µ2 (τ )|zxx (x, 1, τ, t)dτ dx = 0.
(4.14)
Next, (2.2)3 yields
τ zxxt + zxx̺ = 0, (4.15)
Then, by multiplying (4.15) by zxx |µ2 (τ )| and integrating over (0, 1) ×
(0, 1) × (τ1 , τ2 ), and recalling (2.1), we get
1 d
R 1 R 1 R τ2 2
R1R1Rτ
2 dt 0 0 τ1  2
τ |µ (τ )|zxx (x, ̺, τ, t)dτ d̺dx = − 0 0 τ12 |µ2 (τ )|zxx z̺xx (x, ̺, τ, t)dτ d̺dx
R R R
τ 1 1Rτ
= 21 τ12 |µ2 (p)|dp 0 ϕ2xxt dx − 21 0 τ12 |µ2 (τ )|zxx 2
(x, 1, τ, t)dτ dx.
(4.16)
By adding (4.16) to (4.14) we obtain
R1 R1 Rτ
E1′ (t) = −µ1 0 ϕ2xxt dx − 0 ϕxxt τ12 |µ2 (τ )|zxx (x, 1, τ, t)dτ dx
R R R
τ 1 1Rτ
+ 21 τ12 |µ2 (p)|dp 0 ϕ2xxt dx − 21 0 τ12 |µ2 (τ )|zxx
2
(x, 1, τ, t)dτ dx,
(4.17)
The Young’s inequality leads us to obtain
 Z τ2 Z 1
E1′ (t) ≤ − µ1 − |µ2 (p)|dp ϕ2xxt dx, (4.18)
τ1 0
R τ2
we establish (4.13) by inserting η0 = µ1 − τ1
|µ2 (p)|dp and recalling (1.6).

Theorem 4.6. Asumming that (1.6) is satisfied. Then for any constant
a > 0 the energy (4.1) of our problem decays polynomially, which means
a
E(t) ≤ , ∀t > 0.
t
Proof. We define the Lyapunov functional by

L(t) = N E(t) + N1 G1 (t) + G2 (t) + N2 G3 (t) + N3 E1 (t), (4.19)

13
Where N , N1 , N2 , N3 are positive constants.
Now, we differentiate (4.19) and use (4.2), (4.7), (4.9), (4.11), (4.13), we
obtain
R1 R1
L′ (t) ≤ − [η0 N − cρ1 N1 − µ1 N2 ] 0 ϕ2xt dx − [ρ2 − ρ2 N1 ] 0 ψt2 dx
 κ
 R 1 R1
− 2 − µ1 ǫ1 N1 − κǫ22 0 (ϕx + ψ)2 dx − [(b − cµ1 ǫ1 )N1 − c1 ] 0 ψx2 dx
  R R
1 τ R 1
− η1 N2 − N21 0 τ12 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx − η0 N3 0 ϕ2xxt dx
R 1 R 1 R τ2 2
−η1 N2 0 0 τ1 τ |µ2 (τ )|zx (x, ̺, τ, t)dτ d̺dx

As we have for any constant c2 > 0

Z 1 Z 1
− ϕ2xxt dx ≤ −c2 ϕ2xt dx
0 0

Then
R1 R1
L′ (t) ≤ − [η0 N + c2 η0 N3 − cρ1 N1 − µ1 N2 ] 0 ϕ2xt dx − [ρ2 − ρ2 N1 ] 0 ψt2 dx
 κ
 R 1 R1
− 2 − µ1 ǫ1 N1 − κǫ22 0 (ϕx + ψ)2 dx − [(b − cµ1 ǫ1 )N1 − c1 ] 0 ψx2 dx
  R R
1 τ
− η1 N2 − N21 0 τ12 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx
R 1 R 1 R τ2
−η1 N2 0 0 τ1 τ |µ2 (τ )|zx2 (x, ̺, τ, t)dτ d̺dx

Now, we choose
κ N1
ǫ1 = , ǫ2 = , N3 = N,
4µ1 4
Thus, we obtain
R1 R1
L′ (t) ≤ − [η0 (1 + c2 )N − cρ1 N1 − µ1 N2 ] 0 ϕ2xt dx − [ρ2 − ρ2 N1 ] 0 ψt2 dx
κN1
R1 2
 cκ
R1 2
− 8 0 (ϕx + ψ) dx − (b − 4 )N1 − c1 0 ψx dx
 R1 Rτ
− η1 N2 − N21 0 τ12 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx
R1R1Rτ
−η1 N2 0 0 τ12 τ |µ2 (τ )|zx2 dτ d̺dx

Next, we choose N1 large enough such that

 cκ 
α1 = ρ2 − ρ2 N1 > 0, α2 = b − N1 − c1 > 0.
4
Then, we move to choosing N2 large enough such that
N1
α 3 = η1 N 2 − > 0.
2
Finally, we choose N large enough such that

α4 = η0 (1 + c2 )N − cρ1 N1 − µ1 N2 > 0.

Then, we get
R1 R1 R1 R1
L′ (t) ≤ −α4 0 ϕ2xt dx − α1 0 ψt2 dx − κN 8
1
0
(ϕx + ψ)2 dx − α2 0 ψx2 dx
R 1 R τ2
−α3 0 τ1 |µ2 (τ )|zx2 (x, 1, τ, t)dτ dx
R1R1Rτ
−η1 N2 0 0 τ12 τ |µ2 (τ )|zx2 (x, ̺, τ, t)dτ d̺dx

As we have for any constant c > 0

Z 1 Z 1
− ϕ2xt ≤ −c ϕ2t ,
0 0

14
Then
hR R1 R1 R1
1
L′ (t) ≤ −β 0 ϕ2t dx + 0 ψt2 dx + 0 (ϕx + ψ)2 dx + 0 ψx2 dx
R1R1Rτ i
+ 0 0 τ12 τ |µ2 (τ )|zx2 (x, ̺, τ, t)dτ d̺dx

where β = min(α4 , α1 , α2 , η1 N2 ) > 0. So, by the definition of the energy,

we get
L′ (t) ≤ −βE(t), (4.20)
We integrate (4.20) over (0, t), we find
Z t
m
E(t)dt ≤
0 β
where m = L(0). Then, by using the decay of the energy, we obtain
m
tE(t) ≤ ,
β
which means
m
E(t) ≤
,
βt
m
by inserting a = β , we complete the proof.
1
Now, we move to proving the optimality of the rate Q = t− 2 by using
the theorem 5.3 in [4].
1
Lemma 4.7. the rate Q = t− 2 is optimal.
1 − 1
Proof. Assume that Q is not t− 2 , for example it is t 2−η for any 0 < η <
2. Furthermore, we have
√
a a
kÜ k2H = E(t) ≤ =⇒ kÜ kH ≤ √ =⇒ lim kÜ kH = 0,
t t t→∞

which means that iR ⊂ ρ(A). Now, we can apply the theorem 5.3 in [4]
by showing that there exists δ > 0 such that

|λ|−δ−(2−η) k(iλI − A)−1 kH −→ ∞

|λ|→∞

η
Let us take δ = 2
,
then we demonstrate the existence of a real subsequence
(λυ )υ∈N with limυ→∞ |λυ | = ∞, (Üυ )υ∈N ⊂ D(A) and (Fυ )υ∈N ⊂ H, such
that
(iλI − A)Üυ = Fυ
is bounded in H and
η
lim |λυ |−2+ 2 kÜυ kH = ∞
υ→∞

Now, taking the same choices and the same procedure in the proof of
theorem 5, we get
η ρ2 η
|λυ |−2+ 2 kÜυ kH ≥ |λυ | 2 |B|2 −→ ∞
2 υ→∞

1
Consequently, Q = t− 2 .

15
References
[1] T. A. Apalara, Well-posedness and exponential stability for a linear
damped Timoshenko system with second sound and internal distributed
delay, Electron. J. Differential Equations , vol. 2014, no. 254, p. 1-15,
2014.
[2] S. M. S. Cordeiro, R. F. C. Lobato, C. A. Raposo, Optimal polyno-
mial decay for a coupled system of wave with past history, Open J.
Math. Anal., vol. 4, no. 1, p. 4959, dc. 2020, doi: 10.30538/psrp-
oma2020.0052.
[3] M. Elhindi, T. EL Arwadi, Analysis of the thermoviscoelastic Timo-
shenko system with diffusion effect, P. D.Equations in Applied Math-
ematics, vol. 4, p. 113, 2021, doi: 10.1016/j.padiff.2021.100156.
[4] L. H. Fatori and J. E. Munoz Rivera, Rates of decay to weak thermoe-
lastic Bresse system, IMA Journal of Applied Mathematics, vol. 75,
no. 6, p. 881904, dc. 2010, doi: 10.1093/imamat/hxq038.
[5] L. Foughali, S. Zitouni, L. Bouzet and H. E. Khochemane ,Well-
posedness and general decay for a porous-elastic system with microtem-
peratures effects and time-varying delay term, Z. Angew. Math. Phys.,
vol. 73, 183 (2022). https://doi.org/10.1007/s00033-022-01801-0.
[6] F. Huang, Characteristic condition for exponential stability of linear
dynamical systems in Hilbert space, Ann Diff Eqs, vol. 1,p. 43-56, 1985.
[7] M. Houasni, A. Krelifa, H. Boukabcha, S. Zitouni, On the exponen-
tial decay of a thermoelastic porous system with a forcing term and
constant delay, MESA, Vol. 14,no 2,p.1.16. 2023
[8] J.U. Kim , Y. Renardy,Boundry control of the Timoshenko beam,
SIAM J Control Optim, vol. 25,no. 6,p.1417-1429, nov. 1987.
[9] H. Makheloufi, M. Bahlil, et B. Feng, Optimal polynomial decay
for a Timoshenko system with a strong damping and a strong de-
lay, Math Meth Appl Sci, vol. 44, n8, p. 63016317, mai 2021, doi:
10.1002/mma.7183.
[10] S. A. Messaoudi, A. Fareh, et N. Doudi, Well posedness and exponen-
tial stability in a wave equation with a strong damping and a strong
delay, Journal of Mathematical Physics, vol. 57, no. 11, p. 111501,
nov. 2016, doi: 10.1063/1.4966551.
[11] J. E. Muoz Rivera et M. G. Naso, About the stability to Timoshenko
system with one boundary dissipation, Applied Mathematics Letters,
vol. 86, p. 111118, dc. 2018, doi: 10.1016/j.aml.2018.06.023.
[12] D. Ouchenane, Z. Khalili, Global nonexistence of solution for cou-
pled nonlinear Klein–Gordon with degenerate damping and source
terms, Studia Universitatis Babes-Bolyai Mathematica 67(4), 801–815
(2022).
[13] D. Ouchenane, L. Bouzettouta, F. Hebhoub, F. Yazid, Exponential
stability for porous system of thermoelasticity-type III with past his-
tory and distributed delay term, Asian-European Journal of Mathe-
maticsVol. 16, No. 09, 2350160 (2023).

16
[14] J. Prss, On the spectrum of C0 -semigroups, Trans AMS, vol. 28, p.
847-857, 1984.
[15] C. A. Raposo, J. Ferreira, M. L. Santos, et N. N. O. Cas-
tro,Exponential stability for the Timoshenko system with two weak
dampings, Applied Mathematics Letters, vol. 18, n5, p. 535541, mai
2005, doi: 10.1016/j.aml.2004.03.017.
[16] F. Yazid, D. Ouchenane, K. Zennir, Global nGonexistence of solutions
to system of Klein–Gordon equations with degenerate damping and
strong source terms in viscoelasticity , Studia Universitatis Babes-
Bolyai Mathematica 67(3), 563–578 (2022).

17