Stabilization of the wave equation with boundary

time-varying delay

Hao Li
School of Ocean Sciences
China University of Geosciences
Beijing, 100083, China

Changsong Lin
School of Ocean Sciences
China University of Geosciences
Beijing, 100083, China

Shupeng Wang
Institute of Information Engineering
Chinese Academy of Sciences
Beijing, 100093, China

Yanmei Zhang
School of Information Engineering
China University of Geosciences
Beijing, 100083, China

Abstract We study the stabilization of the wave equation with variable coefficients

in a bounded domain and a time-varying delay term in the time-varying, weakly

nonlinear boundary feedbacks. By the Riemannian geometry methods and a suitable

assumption of nonlinearity, we obtain the uniform decay of the energy of the closed

Corresponding author, E-mail address:comin3000@163.com, TEL: +8615169080900.

1
 loop system.

Keywords wave equation, time-varying delay, weakly nonlinear boundary feedback

MR Subject Classification 93C20, 93D20.

1 Introduction

Let be a bounded domain in IRn with smooth boundary . It is assumed that consists
of two parts 1 and 2 ( = 1 2 ) with 2 6= , 1 2 = . Define

Au = div A(x)u for u H 1 () (1.1)

where div is the divergence operator of the standard metric of IRn . A(x) = (aij (x)) is
symmetric, positively definite matrices for each x IRn and aij (x) are smooth functions
on IRn .
We consider the stabilization of the wave equations with variable coefficients and time-
varying delay in the dissipative boundary feedback:

utt + Au = 0 (x, t) (0, +),

u(x, t)|2 = 0 t (0, +),


u(x,t)
A + (t)(g (u (x, t)) + g (u (x, t (t)))) = 0
1 t 2 t (x, t) 1 (0, +),



u(x, 0) = u0 (x), ut (x, 0) = u1 (x) x ,







ut (x, t (0)) = f0 (x, t (0)) (x, t) 1 (0, (0))

(1.2)
where (t) satisfies

0 (t) 0 and d0 0 (t) d < 1 t 0, (1.3)

where 0 > 0, d0 and d are constants. g1 C(IR) and there exist positive constants
c1 , p 1 such that

g1 (0) = 0, sg1 (s) |s|2 f or s IR. (1.4)

1
|g1 (s)| c1 |s| f or |s| > 1 and s2 + (g1 (s))2 c1 (sg1 (s)) p f or |s| 1. (1.5)

2
g2 (s) C(IR) satisfies
(g2 (s))2 sg1 (s) f or s IR, (1.6)

and (t) C([ (0), +)) satisfies

0 < (t) 0 t (0), (1.7)

F (t)
c2 (t) (t (t)) c3 (t) t 0 and lim = 0, (1.8)
t> t
1
where 0 , c2 and c3 are are positive constants and F (t) = inf{()|0t} .
u
A is the co-normal derivative

u
= hA(x)u, i,
A

where h, i denotes the standard metric of the Euclidean space IRn and (x) is the outside
unit normal vector for each x . Moreover, > 0, IR, 6= 0 and the initial data
(u0 , u1 , f0 , w0 , w1 , h0 ) belongs to a suitable space.
There is a specific example for (t). Let 1 > 0 be a constant. If (t) satisfies

1 (t) 0 t 0. (1.9)

Then
1 0 F (t) 0
(t) (t (t)) (t) t 0, lim = 0. (1.10)
0 1 0> t t
Condition 1.7 and 1.8 hold.
In absence of delay ( = 0), the problem (1.2) was studied by [1], [2], [3], [4], [5] [6],
[7], [8] and many others. The decay rate of the energy (when t goes to infinity) depends
on the function and the growth of g1 .
The system (1.2) with constant coefficient (the case:A(x) is a constant matrix on )
was studied by [9],[10], [11] and many other authors . For the system (1.2) with variable
coefficients, the main tools to cope with the system (1.2) are the differential geometrical
methods which were introduced by [12] and have been applied in many papers. See [13],
[14], [15], [16], [18], [19], [20], [21],[22],[23] and references cited there. For a survey on the
differential geometric methods, see [17], [24].
The authors in [11] considered the system (1.2) with constant coefficients operator
and dissipative boundary conditions of time dependent delay and proved the exponential

3
decay of the energy by combining the multiplier method with the use of a suitable integral
inequalities. Different from this paper, g1 is assumed to be linearly bounded and is
assumed to be a constant function in the paper [11].
Based on [11], the purpose of this paper is to solve the stability of the system (1.2) with
variable coefficients and time-varying, weakly nonlinear terms. To obtain our stabilization
result, we assume that
1 1 ||
( + ) < , (1.11)
2c2 2 1 d
where c2 is defined in (1.8).
Define the energy of the system (1.2) by

n Z t
1
Z X Z
E(t) = u2 + aij uxi uxj dx + ()u (x, )g1 (u (x, ))dd. (1.12)
t
2 i,j=1 t (t) 1

where is a positive constant satisfying

|| ||
< < . (1.13)
2c2 1 d 2 1d

We define
g = A1 (x) for x IRn (1.14)

as a Riemannian metric on IRn and consider the couple (IRn , g) as a Riemannian manifold
with an inner product

hX, Y ig = hA1 (x)X, Y i, |X|2g = hX, Xig X, Y IRxn .

Let Dg denote the Levi-Civita connection of the metric g. For the variable coefficients,
the main assumptions are:
Assumption (A) There exists a vector field H on and a constant 0 > 0 such
that
Dg H(X, X) 0 |X|2g for X IRxn x . (1.15)

Moreover we assume that

sup div H < inf div H + 20 (1.16)
x x

and
H 0 x 2 and H x 1 , (1.17)

4
where is a positive constant.
Assumption (1.15) was introduced by [12] as a checkable assumption for the exact
controllability of the wave equation with variable coefficients. The assumption (A) is also
useful for the controllability and the stabilization of the quasilinear wave equation [15].
For the examples of the condition, see [12], [17].
Based on Assumption(1.15), Assumption (A) was given by [23] to study the stabiliza-
tion of the wave equation with variable coefficients and boundary condition of memory
type. The authors in [23] also constructed some examples of the condition based on the
assumption A(x) = a(x)I or A(x) is a perturbation of a symmetric positive definite matrix
A.
Define

H12 () = {u H 1 ()| u|2 = 0}. (1.18)

To obtain the stabilization of the system (1.2), we assume the system (1.2) is well-posed
such that
u C 1 ([0, +), L2 ()) C([0, +), H12 ()) (1.19)

and
ut C([0, +), L2 (1 (t, t (t)))). (1.20)

The main result of this paper is the following:

Theorem 1.1 Let the assumption (A) holds true. Then there exists a constant C,
such that
1
E p (0) F (t + 0 )
E(t) C( 1 + E(0)) t > 0. (1.21)
(t + 0 ) p t + 0
Remark If p = 1 and (t) satisfies

1 (t) 0 t 0. (1.22)

where 0 and 1 are positive constants. Then it follows from (1.12) that there exit con-
stants T0 > 0 and 0 < C0 < 1 such that

E(T0 ) C0 E(0). (1.23)

Then the decay of the energy E(t) is exponential. Methods in [22],[23] are useful for
Theorem 1.1.

5
2 Basic inequality of the System

In this section we work on with two metrics at the same time, the standard dot metric
h, i and the Riemannian metric g = h, ig given by (1.14).
If f C 1 (IRn ), we define the gradient g f of f in the Riemannian metric g, via the
Riesz representation theorem, by

X(f ) = hg f, Xig , (2.1)

where X is any vector field on (IRn , g). The following lemma provides further relations
between the two metrics, see [12], Lemma 2.1.

Lemma 2.1 Let x = (x1 , , xn ) be the natural coordinate system in IRn . Let f , h be
functions and let H, X be vector fields. Then
(a)
hH(x), A(x)X(x)ig = hH(x), X(x)i, x IRn ; (2.2)

(b)
n X
X n 
g f = aij (x)fxj = A(x)f, x IRn , (2.3)
i=1 j=1
xi
where f is the gradient of f in the standard metric;
(c)
g f (h) = hg f, g hig = hf, A(x)hi, x IRn , (2.4)

where the matrix A(x) is given in the formula (1.1);

To prove Theorem 1.1, we still need several lemmas further. Define

1
Z  
E0 (t) = u2t + |g u|2g dx. (2.5)
2

Then, we have
Z t Z
E(t) = E0 (t) + u2t (x, )dd. (2.6)
t (t) 1

Lemma 2.2 Suppose that the condition (1.13) holds true. Let (u, w) be the solution
of system (1.2). Then there exits constants C1 , C2 > 0 such that
Z TZ
E(0) E(T ) C1 (t)[ut (x, t)g1 (ut (x, t)) + ut (x, t (t))g1 (ut (x, t (t))]ddt
(2.7)
0 1

where T 0. The assertion (2.7) implies that E(t) is decreasing.

6
 Proof. Differentiating (1.12), we obtain
Z
E 0 (t) = (ut utt + g u ut ) dx
Z

+ (t)ut (x, t)g1 (ut (x, t))d

1
Z
(t (t))(1 0 (t))ut (x, t (t))g1 (ut (x, t (t)))d. (2.8)
1

Applying Greens formula, and by integrating by parts, with (1.3), (1.8) we arrive at
Z
0
E (t) = (t)[ut g1 (ut ) ut g2 (ut (x, t (t)))]d
1
Z
+ (t)ut (x, t)g1 (ut (x, t))d
1
Z
(t (t))(1 0 (t))ut (x, t (t))g1 (ut (x, t (t)))d
1
||
Z  
(t) ut g1 (ut ) + u2t + ut g1 (ut ) d
1 2 1d
Z
+ (t)[c2 (1 d)ut (x, t (t))g1 (ut (x, t (t)))
1

1 d|| 2
+ g2 (ut (x, t (t)))]d. (2.9)
2
It follows from (1.3), (1.4), (1.11) and (1.13) that
Z
E 0 (t) C(t) [ut (x, t)g1 (ut (x, t)) + ut (x, t (t))g1 (ut (x, t (t)))]d.(2.10)
1

where C > 0 satisfies

( )
|| 1 d||
C = min , c2 (1 d) .
2 1d 2
Then the inequality (2.7) follows directly from (2.10) integrating from 0 to T .2

3 Proofs of Theorem 1.2

From Proposition 2.1 in [12], we have the following identities.

Lemma 3.1 Suppose that u(x, t) solves equation utt + Au = 0, (x, t) (0, +)
and that H is a vector field defined on . Then for T 0.
Z TZ Z TZ 
u 1 
H(u)ddt + u2t |g u|2g H ddt
0 A 2 0
Z TZ
= (ut , H(u))|T0 + Dg H(g u, g u)dxdt
0
Z TZ 
1 
+ u2t |g u|2g divHdxdt. (4.1)
2 0

7
 Moreover, assume that P C 1 (). Then
Z TZ  
u2t |g u|2g P dxdt
0
Z TZ Z TZ
1 u
= (ut , uP )|T0 + g P (u2 )dxdt Pu ddt. (4.2)
2 0 0 A

Lemma 3.2 Suppose that all assumptions in Theorem 1.1 hold true. Let u be the
solution of the system (1.2). Then there exists a positive constant C for which
Z T Z T +0 Z
u 2
E(t)dt CE0 (T + 0 ) + C (u2t + ( ) )ddt
0 0 1 A
Z T +0 Z
+C (t)ut (x, t (t))g1 (ut (x, t (t)))ddt, (4.3)
0

where T 0.
Proof. Let be a positive constant satisfying

1 1
sup div H < < inf div H + 0 . (4.4)
2 x 2 x

Set
H = H, P = 0

Substituting the identity (4.2) into the identity (4.1),we have

Z TZ  
= (ut , H(u) + P u) |T0 + Dg H(g u, g u) 0 |g u|2g dxdt
0
Z TZ 
1 1

+ ( div H + 0 )u2t + ( div H)|g u|2g dxdt, (4.5)
0 2 2

where
Z TZ Z TZ 
u 1 
= (H(u) + uP ) ddt + u2t |g u|2g H ddt. (4.6)
0 A 2 0

We decompose as
= 1 + 2 .

Since u|2 = 0, we obtain u|2 = 0, that is,

u A
g u = for x 2 . (4.7)
A |A |2g

Similarly, we have

u H
H(u) = hH, g uig = for x 2 . (4.8)
A |A |2g

8
Using the formulas (4.7) and (4.8) in the formula (4.6) on the portion 2 , with (1.17) we
obtain
Z TZ 2
1 u H

2 = ddt 0. (4.9)
2 0 2 A |A |2g
From (1.17), we have
Z TZ
u 1 T  Z Z
1 = (H(u) + uP ) ddt + u2t |g u|2g H ddt
0 1 A 2 0 1
Z TZ Z TZ Z TZ 
u 2 2 2

C ( ) ddt + (u + |g u|g )ddt + Cu2t |g u|2g ddt
0 1 A 0 1 0 1
Z TZ Z TZ
u 2
C ( ) ddt + E0 (t) + C u2t ddt (4.10)
0 1 A 0 1

Substituting the formulas (4.9) and (4.10) into the formula (4.5), with (4.4) we obtain
Z T Z TZ
u 2
E0 (t)dt C(E0 (0) + E0 (T )) + C (u2t + ( ) )ddt. (4.11)
0 0 1 A

Let % = (), from (1.3), (1.7), (1.8) and (2.6) we have

Z T +0 Z T +0 Z
E0 (t)dt + (t)ut (x, t (t))g1 (ut (x, t (t)))ddt
0 0
Z T +0
1 T t+0
Z Z Z
E0 (t)dt + ()u (x, ())g1 (ut (x, ())dddt
0 0 0 t 1
1 T t+T1 (t+0 )
Z T +0
d%
Z Z Z
= E0 (t)dt + ()ur ho(x, %)g1 (ut (x, %))d dt
0 0 0 t (t) 1 1 0 ()
Z T +0 Z TZ t
c2
Z
E0 (t)dt + (%)u% (x, %)g1 (ut (x, %))dd%dt
0 (1 d)0 0 t (t) 1
Z T
c2
min{1, } E(t)dt. (4.12)
(1 d)0 0

Since
Z T +0 Z
u
E0 (0) = E0 (T + 0 ) ddt ut
0 1 A
1 T +0 u 2
Z Z
E0 (T + 0 ) + (u2t + ) ddt, (4.13)
2 0 1 A

substituting the formula (4.12) into the formula (4.11), we obtain

Z T Z T +0 Z
u 2
E(t)dt CE0 (T + 0 ) + C (u2t + ( ) )ddt
0 0 1 A
Z T +0 Z
+ (t)ut (x, t (t))g1 (ut (x, t (t)))ddt. (4.14)
0

The inequality (4.3) holds.2

9
 Proof of Theorem 1.2 Since E(t) is decreasing, from (4.3), for sufficiently large T
we have
Z T +0 Z Z T +0 Z
u 2
T E(T ) C (u2t +( ) )ddt + C (t)ut (x, t (t))g1 (ut (x, t (t)))ddt
0 1 A 0
Z T +0 Z Z T +0 Z
C ((t))2 [(g1 (ut ))2 + (g2 (ut ))2 ]ddt + C u2t (x, t)ddt
0 1 0 1
Z T +0 Z
+C (t)ut (x, t (t))g1 (ut (x, t (t)))ddt
0
Z T +0 Z Z T +0 Z
C ((t))2 [(g1 (ut ))2 + (g2 (ut ))2 ]ddt + CF (T + 0 ) (t)u2t (x, t)ddt
0 1 0 1
Z T +0 Z
+C (t)ut (x, t (t))g1 (ut (x, t (t)))ddt. (4.15)
0

where F (t) is defined in (1.8). With (1.4) (1.8) and (2.7), we deduce that
Z T +0 Z
T E(T ) C (t)(s2 + (g1 (s))2 )ddt + CF (T + 0 )(E(0) E(T + 0 ))
0 1
Z T +0 Z
1
C (t)(ut g1 (ut )) p ddt + CF (T + 0 )(E(0) E(T + 0 ))
0 {x1 ,|ut (x,t)|1}
Z T +0 Z
1
C (t)(ut g1 (ut )) p ddt + CF (T + 0 )(E(0) E(T + 0 ))
0 1
Z T +0 Z T +0 Z
1 1
Cmeas(1 ) (t)dt ( R T +0 (t)ut g(ut )ddt) p
meas(1 )
0 0 (t)dt 0 1
+CF (T + 0 )(E(0) E(T + 0 ))
1 p1 1
C(T + 0 ) (E(0) E(T + 0 )) p

+CF (T + 0 )(E(0) E(T + 0 )). (4.16)

Therefore
1
E p (0) F (T + 0 )
E(t) C( 1 + E(0)). (4.17)
(T + 0 ) p T + 0

Note that E(t) is decreasing, the estimate (1.21) holds.2

Conflict of Interests

On behalf of my co-authors, the authors declare that there is no conflict of interests

regarding the publication of this article.

10
References

[1] A. Benaissa and A. Guesmia, Energy decay for wave equations of -Laplacian type with
weakly nonlinear dissipation, Electron. J. Differ. Equations 2008(109), 1-22 (2008).

[2] G. Chen, Control and stabilization for the wave equation in a bounded domain, Part
I, SIAM J. Control Optim. 17, 66-81 (1979).

[3] G. Chen, Control and stabilization for the wave equation in a bounded domain, Part
II, SIAM J. Control Optim. 19, 114-122 (1981).

[4] A. Haraux, Two Remarks on Dissipative Hyperbolic Problems, Research Notes in

Mathematics, Vol. 122. (Pitman, Boston, MA, 1985), pp. 161-179.

[5] I. Lasiecka and D. Tataru, Uniform boundary stabilization of semilinear wave equations
with nonlinear boundary dampin, Diff. Integral Eq. 6, 507-533 (1993).

[6] K. Liu, Locally distributed control and damping for the conservative systems, SIAM
J. Contr. Optim., 35 (1997), pp. 1574-1590.

[7] E. Zuazua, Exponential decay for the semi-linear wave equation with locally distributed
damping, Commun. Part. Diff. Eq, 15 (1990), pp. 205-235.

[8] S. Feng and D. X. Feng, Locally distributed control of wave equation with variable
coefficients, Sci. China ser. A 44 (3) (2001), pp. 345-350.

[9] S. Nicaise, C. Pignotti, Interior feedback stabilization of wave equations with time
dependent delay, Electron. J. Differ. Equations. Vol. 2011 (2011), No. 41, pp. 1-20.

[10] S. Nicaise and C. Pignotti, Stability and instability results of the wave equation with
a delat term in the boundary or internal feedbacks, SIAM J. Contr. Optim. Vol. 45,
No. 5(2006), pp. 1561-1585.

[11] S. Nicaise, C. Pignotti and J. Valein. Exponential stability of the wave equation with
boundary time-varying delay. Discrete Contin. Dyn. Syst. Ser. S, 4:693 C 722, 2011.

[12] P.F. Yao, On the observability inequalities for the exact controllability of the wave
equation with variable coefficients, SIAM J. Contr. Optim, 37 (6) (1999) 1568-1599.

[13] , Observability inequalities for the shallow shell, SIAM J. Control Optim., 38 (6)
(2000) 1729-1756.

11
[14] , Global smooth solutions for the quasilinear wave equation with boundary dissi-
pation, J. Differential Equations 241 (2007) 62-93.

[15] , Boundary controllability for the quasilinear wave equation, Appl. Math. Optim.,
61 (2010), no. 2, 191-233.

[16] , Energy decay for the Cauchy problem of the linear wave equation of variable coef-
ficients with disspation, Chinese Annals of Mathematics, Series B, Vol. 31, no.1,(2010),
59-70.

[17] , Modeling and Control in Vibrational and Structural Dynamics. A Differential

Geometric Approach. Chapman and Hall/CRC Applied Mathematics and Nonlinear
Science Series. CRC Press, Boca Raton, FL, 2011. xiv+405 pp. ISBN: 978-1-4398-
3455-8.

[18] Z. F. Zhang and P. F. Yao, global smooth solutions of the quasilinear wave equation
with internal velocity feedbacks, SIAM J. Control Optim., Vol 47(2008), No. 4, pp
2044-2077.

[19] I. Lasiecka, R. Triggiani, P.F. Yao, Inverse/observability estimates for second-order

hyperbolic equations with variable systems, J. Math. Anal. Appl. 235 (1999) 13-57.

[20] Z. H. Ning and Q. X. Yan, Stabilization of the wave equation with variable coefficients
and a delay in dissipative boundary feedback, J.Math.Annal.Appl.367(2010)167-173

[21] Z.H. Ning, C.X. Shen, X.P.Zhao, Stabilization of the wave equation with variable
coefficients and a delay in dissipative internal feedback, J.Math.Annal.Appl.405(2013),
148-155.

[22] Z.H. Ning, C.X. Shen, X.P.Zhao, H.Li, C.S.Lin, Y.M. Zhang, Nonlinear Boundary
Stabilization of the Wave Equations with Variable coefficients and time dependent
delay, Applied Mathematics and Computation, in press.

[23] S.Nicaise and C.Pignotti, Stabilization of the wave equation with variable coefficients
and boundary condition of memory type, Asymptotic Analysis 50 (2006) 31-67

[24] R. Gulliver, I. Lasiecka, W. Littman, and R. Triggiani,: The case for differential
geometry in the control of single and coupled PDEs: the structural acoustic chamber.
In: Geometric Methods in Inverse Problems and PDE Control. IMA Vol. Math. Appl.,
vol. 137, pp. 73-181. Springer, New York (2004).

12