a wave equation associated with mixed nonhomogeneous conditions: global existence and asymptotic...

21
Nonlinear Analysis 66 (2007) 1526–1546 www.elsevier.com/locate/na A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions Nguyen Thanh Long , Vo Giang Giai Department of Mathematics and Computer Science, University of Natural Science, Vietnam National University HoChiMinh City, 227 Nguyen Van Cu Str., Dist.5, HoChiMinh City, Viet Nam Received 13 January 2006; accepted 6 February 2006 Abstract The paper deals with the initial–boundary value problem for the linear wave equation u tt u xx + Ku + λu t = F (x , t ), 0 < x < 1, 0 < t < T , u x (0, t ) = P (t ), u(1, t ) = 0, u(x , 0) = u 0 (x ), u t (x , 0) = u 1 (x ), (1) where K , λ are given constants and u 0 , u 1 , F are given functions, and the unknown function u(x , t ) and the unknown boundary value P (t ) satisfy the following nonlinear integral equation: P (t ) = g(t ) + K 1 |u(0, t )| α2 u(0, t ) + λ 1 |u t (0, t )| β2 u t (0, t ) t 0 k (t s )u(0, s )ds, (2) where K 1 , λ 1 , α, β are given constants and g, k are given functions. In this paper, we consider three main parts. In Part 1 we prove a theorem of global existence and uniqueness of a weak solution (u, P ) of problem (1.1)–(1.5). The proof is based on a Galerkin method associated with a priori estimates, weak convergence and compactness techniques. For the case of α = β = 2, Part 2 is devoted the study of the asymptotic behavior of the solution (u, P ) as λ 1 0 + . Finally, in Part 3 we obtain an asymptotic expansion of the solution (u, P ) of the problem (1.1)–(1.5) up to order N + 1 2 in four small parameters K , λ, K 1 , λ 1 . c 2006 Elsevier Ltd. All rights reserved. MSC: 35L20; 35L70 Corresponding author. E-mail addresses: [email protected] , [email protected] (N.T. Long). 0362-546X/$ - see front matter c 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.na.2006.02.007

Upload: nguyen-thanh-long

Post on 21-Jun-2016

215 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

Nonlinear Analysis 66 (2007) 1526–1546www.elsevier.com/locate/na

A wave equation associated with mixednonhomogeneous conditions: Global existence and

asymptotic expansion of solutions

Nguyen Thanh Long∗, Vo Giang Giai

Department of Mathematics and Computer Science, University of Natural Science, Vietnam National UniversityHoChiMinh City, 227 Nguyen Van Cu Str., Dist.5, HoChiMinh City, Viet Nam

Received 13 January 2006; accepted 6 February 2006

Abstract

The paper deals with the initial–boundary value problem for the linear wave equation⎧⎨⎩utt − ux x + K u + λut = F(x, t), 0 < x < 1, 0 < t < T,

ux (0, t) = P(t), u(1, t) = 0,

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

(1)

where K , λ are given constants and u0, u1, F are given functions, and the unknown function u(x, t) andthe unknown boundary value P(t) satisfy the following nonlinear integral equation:

P(t) = g(t) + K1 |u(0, t)|α−2 u(0, t) + λ1 |ut (0, t)|β−2 ut (0, t) −∫ t

0k(t − s)u(0, s)ds, (2)

where K1, λ1, α, β are given constants and g, k are given functions.In this paper, we consider three main parts. In Part 1 we prove a theorem of global existence and

uniqueness of a weak solution (u, P) of problem (1.1)–(1.5). The proof is based on a Galerkin methodassociated with a priori estimates, weak convergence and compactness techniques. For the case of α = β =2, Part 2 is devoted the study of the asymptotic behavior of the solution (u, P) as λ1 → 0+. Finally, in Part3 we obtain an asymptotic expansion of the solution (u, P) of the problem (1.1)–(1.5) up to order N + 1

2 infour small parameters K , λ, K1, λ1.c© 2006 Elsevier Ltd. All rights reserved.

MSC: 35L20; 35L70

∗ Corresponding author.E-mail addresses: [email protected], [email protected] (N.T. Long).

0362-546X/$ - see front matter c© 2006 Elsevier Ltd. All rights reserved.doi:10.1016/j.na.2006.02.007

Page 2: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1527

Keywords: Galerkin method; Global existence and uniqueness of a weak solution; Energy-type estimates; Compactness;Asymptotic expansion

1. Introduction

In this paper, we consider the following problem: Find a pair (u, P) of functions satis-fying

utt − ux x + f (u, ut ) = F(x, t), 0 < x < 1, 0 < t < T, (1.1)

ux (0, t) = P(t), (1.2)

u(1, t) = 0, (1.3)

u(x, 0) = u0(x), ut (x, 0) = u1(x), (1.4)

where f (u, ut ) = K u + λut , where K , λ are given constants and u0, u1, F are givenfunctions satisfying conditions specified later, and the unknown function u(x, t) and the unknownboundary value P(t) satisfy the following integral equation

P(t) = g(t) + K1 |u(0, t)|α−2 u(0, t) + λ1 |ut (0, t)|β−2 ut (0, t)

−∫ t

0k(t − s)u(0, s)ds, (1.5)

where K1, λ1, α, β are given constants and g, k are given functions.In the case of α = 2, λ1 ≡ 0, K1 = h ≥ 0, the problem (1.1)–(1.5) is formed from the problem

(1.1)–(1.4) wherein the unknown function u(x, t) and the unknown boundary value P(t) satisfythe following Cauchy problem for ordinary differential equation{

P ′′(t) + ω2 P(t) = hutt (0, t), 0 < t < T,

P(0) = P0, P ′(0) = P1,(1.6)

where h ≥ 0, ω > 0, P0, P1 are given constants [6].In [2], An and Trieu have studied a special case of problem (1.1)–(1.4) and (1.6) with

u0 = u1 = P0 = 0 and f (u, ut ) = K u + λut , with K ≥ 0, λ ≥ 0 given constants. In thelatter case the problem (1.1)–(1.4) and (1.6) is a mathematical model describing the shock of arigid body and a linear viscoelastic bar resting on a rigid base [2].

From (1.6) we represent P(t) in terms of P0, P1, ω, h, utt(0, t) and then by integrating byparts, we have

P(t) = g(t) + hu(0, t) −∫ t

0k(t − s)u(0, s)ds, (1.7)

where

g(t) = (P0 − hu0(0)) cosωt + 1

ω(P1 − hu1(0)) sin ωt, (1.8)

k(t) = hω sin ωt . (1.9)

In [3] Bergounioux, Long and Dinh studied problem (1.1) and (1.4) with the mixed boundaryconditions (1.2) and (1.3) standing for

ux (0, t) = g(t) + hu(0, t) −∫ t

0k(t − s)u(0, s)ds, (1.10)

Page 3: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1528 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546

ux(1, t) + K1u(1, t) + λ1ut (1, t) = 0, (1.11)

where h ≥ 0, K , λ, K1, λ1 are given constants and g, k are given functions.In [8], Long, Dinh and Diem obtained the unique existence, regularity and asymptotic

expansion of the problem (1.1) and (1.4) in the case of −ux(1, t) = K1u(1, t) + λut (1, t),ux(0, t) = P(t) where P(t) satisfies (1.7).

In [10], Long, Ut and Truc gave the unique existence, stability, regularity in time variableand asymptotic expansion for the solution of problem (1.1)–(1.5) when α = β = 2, f (u, ut ) =K u + λut , u0 ∈ H 2 and u1 ∈ H 1. In this case, the problem (1.1)–(1.5) is the mathematicalmodel describing a shock problem involving a linear viscoelastic bar.

In this paper, we consider three main parts. In Part 1, under conditions (u0, u1) ∈ H 1 × L2,F ∈ L2(QT ), g, k ∈ H 1(0, T ), K ∈ R, λ, K1 ≥ 0, λ1 > 0, α, β ≥ 2, we prove atheorem of global existence and uniqueness of a weak solution (u, P) of problem (1.1)–(1.5).The proof is based on a Galerkin method associated with a priori estimates, weak convergenceand compactness techniques. We remark that the linearization method in the papers [4,8] cannotbe used in [3,6,7]. For the case of α = β = 2, Part 2 is devoted the study of the asymptoticbehavior of the solution (u, P) as λ1 → 0+. Finally, in Part 3 we obtain an asymptotic expansionof the solution (u, P) of the problem (1.1)–(1.5) up to order N + 1

2 in four small parameters K ,λ, K1, λ1. The results obtained here may be considered as the generalizations of those in An andTrieu [2] and in [3,4,6–10].

2. The existence and uniqueness theorem

Put Ω = (0, 1), QT = Ω × (0, T ), T > 0. We omit the definitions of usual function spaces:Cm

(Ω), L p (Ω), W m,p (Ω).

We define W m,p = W m,p (Ω), L p = W 0,p (Ω), H m = W m,2 (Ω), 1 ≤ p ≤ ∞,m = 0, 1, . . ..

The norm in L2 is denoted by ‖·‖. We also denote by 〈·, ·〉 the scalar product in L2 or pair ofdual scalar products of a continuous linear functional with an element of a function space. Wedenote by ‖·‖X the norm of a Banach space X and by X ′ the dual space of X . We denote byL p(0, T ; X), 1 ≤ p ≤ ∞, the Banach space of the real functions u : (0, T ) → X , measurable,such that

‖u‖L p(0,T ;X) =(∫ T

0‖u(t)‖p

X dt

)1/p

< ∞ for 1 ≤ p < ∞,

and

‖u‖L∞(0,T ;X) = ess sup0<t<T

‖u(t)‖X for p = ∞.

Let u(t), u′(t) = ut (t), u′′(t) = utt (t), ux(t), ux x(t) denote u(x, t), ∂u∂t (x, t), ∂2u

∂t2 (x, t), ∂u∂x (x, t),

∂2u∂x2 (x, t), respectively.

We put

V = {v ∈ H 1(0, 1) : v(1) = 0}, (2.1)

a(u, v) =∫ 1

0

∂u

∂x

∂v

∂xdx . (2.2)

Page 4: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1529

V is a closed subspace of H 1 and on V , ‖v‖H1 and ‖v‖V = √a(v, v) = ‖vx‖ are two equivalent

norms. We then have the following lemma.

Lemma 1. The imbedding V ↪→ C0(0, 1) is compact and

‖v‖C0([0,1]) ≤ ‖v‖V for all v ∈ V . (2.3)

The proof is straightforward and we omit the details.We make the following assumptions:(H1) u0 ∈ V and u1 ∈ L2,(H2) F ∈ L2(QT ),(H3) g, k ∈ H 1(0, T ),(H4) K ∈ R, λ, K1 ≥ 0, λ1 > 0,(H5) α, β ≥ 2.Then, we have the following theorem.

Theorem 1. Let (H1)–(H5) hold. Let T > 0. Then the problem (1.1)–(1.5) has at least one weaksolution (u, P) such that⎧⎨⎩u ∈ L∞(0, T ; V ), ut ∈ L∞(0, T ; L2),

u(0, ·) ∈ W 1,β (0, T ) , P ∈ Lβ ′(0, T ) , β ′ = β

β − 1.

(2.4)

Furthermore, if α = 2 or α ≥ 3, the solution is unique.

Remark 1. In the special case α = β = 2, u0 ∈ V ∩ H 2 and u1 ∈ H 1 we have obtained someresults in the paper [10].

Proof of Theorem 1. The proof consists of Steps 1–4.

Step 1. The Galerkin approximation. Let {w j } be a denumerable base of V . We find theapproximate solution of problem (1.1)–(1.5) in the form

um(t) =m∑

j=1

cmj (t)w j , (2.5)

where the coefficient functions cmj satisfy the system of ordinary differential equation⟨u′′

m(t),w j⟩+ ⟨

umx (t),w j x⟩+ Pm(t)w j (0) + ⟨

K um(t) + λu′m(t),w j

⟩= ⟨

F (t),w j⟩, 1 ≤ j ≤ m, (2.6)

Pm(t) = g(t) + K1 |um(0, t)|α−2 um(0, t) + λ1∣∣u′

m(0, t)∣∣β−2

u′m(0, t)

−∫ t

0k(t − s)um(0, s)ds, (2.7)⎧⎪⎪⎪⎪⎨⎪⎪⎪⎪⎩

um(0) = u0m =m∑

j=1

αmj w j → u0 strongly in H 1,

u′m(0) = u1m =

m∑j=1

βmjw j → u1 strongly in L2.

(2.8)

From the assumptions of Theorem 1, system (2.6)–(2.8) has the solution (um, Pm ), on someinterval [0, Tm]. The following estimates allow one to take Tm = T for all m.

Page 5: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1530 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546

Step 2. A priori estimates. Substituting (2.7) into (2.6), then multiplying the j th equation of (2.6)by c′

mj (t) and summing up with respect to j , afterwards, integrating by parts with respect to thetime variable from 0 to t , we get after some rearrangements

Sm(t) = Sm(0) + 2g(0)u0m(0) − 2g(t)um(0, t) + 2∫ t

0

⟨F(s), u′

m(s)⟩ds

− 2K∫ t

0

⟨um(s), u′

m(s)⟩ds − 2λ

∫ t

0

∥∥u′m(s)

∥∥2 ds

+ 2∫ t

0g′(s)um(0, s)ds + 2

∫ t

0u′

m(0, r)

(∫ r

0k(r − s)um(0, s)ds

)dr

= Sm(0) + 2g(0)u0m(0) − 2g(t)um(0, t) + I1 + I2 + I3 + I4 + I5, (2.9)

where

Sm(t) = ∥∥u′m(t)

∥∥2 + ‖umx (t)‖2 + 2

αK1 |um(0, t)|α + 2λ1

∫ t

0

∣∣u′m(0, s)

∣∣β ds. (2.10)

Using the inequality

2ab ≤ εa2 + 1

εb2, ∀a, b ∈ R, ∀ε > 0, (2.11)

and the following inequalities

|um(0, t)| ≤ ‖um(t)‖C0(Ω)≤ ‖umx (t)‖ ≤ √

Sm(t), (2.12)

‖um(t)‖2 ≤ 2 ‖u0m‖2 + 2∫ t

0

∥∥u′m(s)

∥∥2 ds ≤ 2 ‖u0m‖2 + 2∫ t

0Sm(s)ds, (2.13)

we shall estimate respectively the following terms on the right-hand side of (2.9) as follows

−2g(t)um(0, t) ≤ 1

εg2(t) + εSm(t), for all ε > 0, (2.14)

I1 = 2∫ t

0

⟨F(s), u′

m(s)⟩ds ≤

∫ t

0‖F(s)‖2 ds +

∫ 0

tSm(s)ds, (2.15)

I2 = −2K∫ t

0

⟨um(s), u′

m(s)⟩ds ≤ |K |

∫ t

0Sm(s)ds, (2.16)

I3 = −2λ

∫ t

0

∥∥u′m(s)

∥∥2ds ≤ 2 |λ|

∫ t

0Sm(s)ds, (2.17)

I4 = 2∫ t

0g′(s)um(0, s)ds ≤ ∥∥g′∥∥2

L2(0,T )+∫ t

0Sm(s)ds, (2.18)

I5 = 2∫ t

0u′

m(0, r)

(∫ r

0k(r − s)um(0, s)ds

)dr

= 2um(0, t)

(∫ t

0k(t − s)um(0, s)ds

)− 2

∫ t

0um(0, r)

(∫ r

0k ′(r − s)um(0, s)ds

)dr − 2k(0)

∫ t

0u2

m(0, r)dr

≤ εSm(t) + 1

ε

∫ t

0k2(θ)dθ

∫ t

0Sm(s)ds

Page 6: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1531

+ 2

√t∫ t

0|k ′(θ)|2dθ

∫ t

0Sm(s)ds + 2 |k(0)|

∫ t

0Sm(r)dr

≤ εSm(t) +(

1

ε‖k‖2

L2(0,T )+ 2

√T∥∥k ′∥∥

L2(0,T )+ 2 |k(0)|

)∫ t

0Sm(s)ds, (2.19)

for all ε > 0. On the other hand, using (2.8) and (2.10), (H1), (H3)–(H5) and Lemma 1, we have

Sm(0) + 2g(0)u0m(0)

= ‖u1m‖2 + ‖u0mx‖2 + 2

αK1 |u0m(0)|α + 2g(0)u0m(0) ≤ C1 for all m, (2.20)

where C1 is a constant depending only on u0, u1, g, K1 and α.Combining (2.9), (2.10), (2.14)–(2.20), we obtain

Sm(t) ≤ C1 + 1

ε‖g‖2

L∞(0,T ) + ∥∥g′∥∥2L2(0,T )

+ ‖F‖2L2(QT )

+ 2εSm(t)

+(

2 + |K | + 2 |λ| + 1

ε‖k‖2

L2(0,T )+ 2

√T∥∥k ′∥∥

L2(0,T )+ 2 |k(0)|

)∫ t

0Sm(s)ds,

(2.21)

for all ε > 0. On choosing ε = 14 , it follows from (2.21) that

Sm(t) ≤ M(1)T + N (1)

T

∫ t

0Sm(s)ds, (2.22)

where⎧⎨⎩M(1)T = 2

(C1 + 4 ‖g‖2

L∞(0,T ) + ∥∥g′∥∥2L2(0,T )

+ ‖F‖2L2(QT )

),

N (1)T = 2

(2 + |K | + 2 |λ| + 4 ‖k‖2

L2(0,T )+ 2

√T∥∥k ′∥∥

L2(0,T )+ 2 |k(0)|

).

(2.23)

By Gronwall’s lemma, we deduce from (2.22), (2.23) that

Sm(t) ≤ M(1)T exp

(t N (1)

T

)≤ M(1)

T exp(

T N (1)T

)= CT , for all t ∈ [0, T ]. (2.24)

On the other hand, from the assumptions (H3), (H4), we deduce from (2.7), (2.10), (2.24) that

|Pm(t)| ≤ ‖g‖L∞(0,T ) + K1

(√CT

)α−1 +√

CT

∫ T

0|k(s)| ds + λ1

∣∣u′m(0, t)

∣∣β−1

= C(1)T + λ1

∣∣u′m(0, t)

∣∣β−1, (2.25)∥∥∥∣∣u′

m(0, ·)∣∣β−2u′

m(0, ·)∥∥∥β ′

Lβ′(0,T )

=∫ T

0

∣∣u′m(0, s)

∣∣β ds ≤ CT

2λ1. (2.26)

Hence, we deduce from (2.25), (2.26) that

‖Pm‖Lβ′(0,T )

≤ T 1/β ′C(1)

T + λ1

[∫ T

0

∣∣u′m(0, s)

∣∣β ds

]1/β ′

≤ T 1/β ′C(1)

T + λ1−1/β ′1

(CT

2

)1/β ′

= CT for all m, (2.27)

where CT is a positive constant depending only on T .

Page 7: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1532 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546

Step 3. Limiting process. From (2.10), (2.24), (2.26) and (2.27), we deduce the existence of asubsequence of {(um, Pm)} still also so denoted, such that⎧⎪⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎪⎩

um → u in L∞(0, T ; V ) weak*,u′

m → u′ in L∞(0, T ; L2) weak*,um(0, ·) → u(0, ·) in L∞(0, T ) weakly*,u′

m(0, ·) → u′(0, ·) in Lβ(0, T ) and L2(0, T ) weakly,∣∣u′m(0, ·)∣∣β−2

u′m(0, ·) → χ in Lβ ′

(0, T ) weakly,Pm → P in Lβ ′

(0, T ) weakly.

(2.28)

By the compactness lemma of Lions [5, p. 57] and the imbedding H 1(0, T ) ↪→ C0 ([0, T ]),we can deduce from (2.28)1,2,3,4 the existence of a subsequence still denoted by {um}, such that{

um → u strongly in L2(QT ) and a.e. in QT ,

um(0, ·) → u(0, ·) strongly in C0 ([0, T ]) .(2.29)

From (2.29)2 we have

Pm(t) = g(t) + K1 |um(0, t)|α−2 um(0, t) −∫ t

0k(t − s)um(0, s)ds → P(t), (2.30)

strongly in C0 ([0, T ]), where

P(t) = g(t) + K1 |u(0, t)|α−2 u(0, t) −∫ t

0k(t − s)u(0, s)ds. (2.31)

Hence

Pm(t) → P(t) + λ1χ(t) = P(t), (2.32)

in Lβ ′(0, T ) weakly.

Passing to the limit in (2.6)–(2.8) by (2.28)1,2,5, (2.32), we have (u, P) satisfying the equation

d

dt

⟨u′(t), v

⟩+ 〈ux(t), vx 〉 + (P(t) + λ1χ(t)

)v(0) + 〈K u + λut , v〉

= 〈F (t), v〉 ,∀ v ∈ V . (2.33)

We can prove in a manner similar to that of [6] that

u(0) = u0, u′(0) = u1. (2.34)

Then, in order to prove the existence of the solution of the problem (1.1)–(1.5), we only haveto prove that χ(t) = ∣∣u′(0, t)

∣∣β−2u′(0, t). We shall now require the following lemma.

Lemma 2. Let u be the weak solution of the following problem⎧⎪⎪⎨⎪⎪⎩utt − ux x + Φ1 = 0, 0 < x < 1, 0 < t < T,

ux (0, t) = P1(t), u(1, t) = 0,

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

u ∈ L∞(0, T ; V ), ut ∈ L∞(0, T ; L2).

(2.35)

Then we have1

2

∥∥u′(t)∥∥2 + 1

2‖ux (t)‖2 +

∫ t

0P1(s)u

′(0, s)ds +∫ t

0

⟨Φ1(s), u′(s)

⟩ds

≥ 1

2‖u1‖2 + 1

2‖u0x‖2 a.e. t ∈ [0, T ]. (2.36)

Page 8: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1533

Furthermore, if u0 = u1 = 0 there is equality in (2.36).

The proof of Lemma 2 can be found in [1].We now return to the proof of the existence of a solution of the problem (1.1)–(1.5).It follows from (2.6)–(2.8) that

λ1

∫ t

0

∣∣u′m(0, s)

∣∣β ds =∫ t

0

⟨F(s), u′

m(s)⟩ds − 1

2

∥∥u′m(t)

∥∥2 + 1

2‖u1m‖2

− 1

2‖umx (t)‖2 + 1

2‖u0mx‖2 −

∫ t

0Pm(s)u′

m(0, s)ds

− K∫ t

0

⟨um(s), u′

m(s)⟩ds − λ

∫ t

0

∥∥u′m(s)

∥∥2ds. (2.37)

By Lemma 2 we have

λ1 lim supm→∞

∫ t

0

∣∣u′m(0, s)

∣∣β ds ≤∫ t

0

⟨F(s), u′(s)

⟩ds − 1

2lim infm→∞

∥∥u′m(t)

∥∥2 + 1

2‖u1‖2

− 1

2lim infm→∞ ‖umx(t)‖2 + 1

2‖u0x‖2 −

∫ t

0P(s)u′(0, s)ds

−K∫ t

0

⟨u(s), u′(s)

⟩ds − λ lim inf

m→∞

∫ t

0

∥∥u′m(s)

∥∥2ds

≤ 1

2‖u1‖2 + 1

2‖u0x‖2 − 1

2

∥∥u′(t)∥∥2 − 1

2‖ux (t)‖2

−∫ t

0P(s)u′(0, s)ds

−∫ t

0

⟨K u(s) + λu′(s) − F(s), u′(s)

⟩ds

+ λ1

∫ t

0χ(s)u′(0, s)ds ≤ λ1

∫ t

0χ(s)u′(0, s)ds. (2.38)

Next, consider

Xm(t) =∫ t

0

(Ψ(u′

m(0, s))− Ψ (w(s))

) (u′

m(0, s) − w(s))

ds ≥ 0, ∀w ∈ Lβ(0, T ),

(2.39)

where Ψ (z) = |z|β−2 z.It follows from (2.28)4,5 and (2.38) that

0 ≤ lim supm→∞

Xm(t) ≤∫ t

0(χ(s) − Ψ (w(s)))

(u′(0, s) − w(s)

)ds, ∀w ∈ Lβ(0, T ).

(2.40)

In (2.40), we choose w(s) = u′(0, s) − δφ(s) with δ > 0 and φ ∈ Lβ(0, T ) and use theargument of Minty and Browder (cf. Lions [5], p. 172), we deduce that χ(t) = Ψ

(u′(0, t)

) =∣∣u′(0, t)∣∣β−2

u′(0, t).The proof of existence is completed.

Page 9: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1534 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546

Step 4. Uniqueness of the solution. Let (u, P),(u, P

)be two weak solutions of problem

(1.1)–(1.5), such that⎧⎨⎩u, u ∈ L∞(0, T ; V ), ut , ut ∈ L∞(0, T ; L2),

u(0, ·), u(0, ·) ∈ W 1,β (0, T ) ,

P, P ∈ Lβ ′(0, T ) .

(2.41)

Then (v, Q) with v = u − u and Q = P − P is the weak solution of the following problem⎧⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎩

vt t − vx x + Kv + λvt = 0, 0 < x < 1, 0 < t < T,

vx (0, t) = Q(t), v(1, t) = 0,

v(x, 0) = vt (x, 0) = 0,

Q(t) = P(t) − P(t) = K1 (H (u(0, t)) − H (u(0, t)))

+ λ1 (Ψ (ut (0, t)) − Ψ (ut (0, t))) −∫ t

0k(t − s)v(0, s)ds,

(2.42)

where H (z) = |z|α−2 z.By using Lemma 2 with u0 = u1 = 0, Φ1 = Kv + λvt , P1(t) = Q(t), we have

Z(t) + 2K1

∫ t

0H1(s)v

′(0, s)ds + 2λ1

∫ t

0(Ψ (ut (0, s)) − Ψ (ut (0, s))) v′(0, s)ds

= 2∫ t

0v′(0, s)ds

∫ s

0k(s − r)v(0, r)dr − 2K

∫ t

0

⟨v(s), v′(s)

⟩ds

− 2λ

∫ t

0

∥∥v′(s)∥∥2 ds, a.e. t ∈ [0, T ]. (2.43)

where{Z(t) = ∥∥v′(t)

∥∥2 + ‖vx (t)‖2 ,

H1(t) = H (u(0, t)) − H (u(0, t)) = |u(0, t)|α−2 u(0, t) − |u(0, t)|α−2 u(0, t).(2.44)

Using the inequality

∀p ≥ 2, ∃Cp > 0 :(|x |p−2 x − |y|p−2 y

)(x − y) ≥ Cp |x − y|p , ∀x, y ∈ R, (2.45)

we obtain

2λ1

∫ t

0[Ψ (ut (0, s)) − Ψ (ut (0, s))] v′(0, s)ds ≥ 2λ1Cβ

∫ t

0

∣∣v′(0, s)∣∣β ds, (2.46)

−2K∫ t

0

⟨v(s), v′(s)

⟩ds − 2λ

∫ t

0

∥∥v′(s)∥∥2 ds

≤ 2 |K |∫ t

0‖v(s)‖ ∥∥v′(s)

∥∥ ds + 2 |λ|∫ t

0

∥∥v′(s)∥∥2

ds

≤ (|K | + 2 |λ|)∫ t

0Z(s)ds. (2.47)

On the other hand, we have

Page 10: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1535

2∫ t

0v′(0, s)ds

∫ s

0k(s − r)v(0, r)dr

= 2v(0, t)∫ t

0k(t − r)v(0, r)dr − 2k(0)

∫ t

0v2(0, s)ds

− 2∫ t

0v(0, s)ds

∫ t

0k ′(s − r)v(0, r)dr

≤ 1

2Z(t) +

(2 ‖k‖2

L2(0,T )+ T

∥∥k ′∥∥2L2(0,T )

+ 2 |k(0)| + 1) ∫ t

0Z(r)dr. (2.48)

It follows from (2.43), (2.44), (2.46)–(2.48) that

Z(t) + 4K1

∫ t

0H1(s)v

′(0, s)ds + 4λ1Cβ

∫ t

0

∣∣v′(0, s)∣∣β ds

≤ 2[2 ‖k‖2

L2(0,T )+ T

∥∥k ′∥∥2L2(0,T )

+ 2 |k(0)| + 1 + |K | + 2 |λ|] ∫ t

0Z(s)ds. (2.49)

Now, we consider two cases for α.

Case α = 2. H1(t) = H (u(0, t)) − H (u(0, t)) = u(0, t) − u(0, t) = v(0, t).We then have

4K1

∫ t

0H1(s)v

′(0, s)ds = 2 K1v2(0, t) ≥ 0. (2.50)

By Gronwall’s lemma, we obtain from (2.49) and (2.50) that Z ≡ 0, i.e., u ≡ u.Case α ≥ 3. By using integration by parts, it follows that

2K1

∫ t

0H1(s)v

′(0, s)ds = 2K1

∫ t

0

∫ 1

0

[d

dθH (u(0, s) + θv(0, s)) dθ

]v′(0, s)ds

= K1

[∫ 1

0H ′ (u(0, t) + θv(0, t)) dθ

]v2(0, t)

− K1

∫ t

0v2(0, s)ds

[∫ 1

0H ′′ (u(0, s) + θv(0, s)) (u ′(0, s)

+ θv′(0, s))dθ

]

≥ −K1(α − 1)(α − 2)Rα−3∫ t

0ω1(s)Z(s)ds, (2.51)

where

R = max{‖u‖L∞(0,T ;V ) , ‖u‖L∞(0,T ;V )}, ω1(s) = ∣∣u′(0, s)∣∣+ ∣∣u ′(0, s)

∣∣ . (2.52)

Using Gronwall’s lemma, it follows from (2.49)–(2.51) that Z ≡ 0, i.e., u ≡ u. Theorem 1 isproved completely. �

3. Asymptotic behavior of the solutions as λ1 → 0+

In this part, we assume that α = β = 2 and (u0, u1, F , g, k, K , λ, K1) satisfy the assumptions(H1)–(H4). Let λ1 > 0. By Theorem 1, the problem (1.1)–(1.5) has a unique weak solution

Page 11: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1536 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546

(u, P) depending on λ1:

u = uλ1, P = Pλ1 . (3.1)

We consider the following perturbed problem, where λ1 is a small parameter:⎧⎪⎪⎪⎨⎪⎪⎪⎩Au ≡ utt − ux x + K u + λut = F(x, t), 0 < x < 1, 0 < t < T,

ux (0, t) = P(t), u(1, t) = 0,

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

P(t) = g(t) + K1u(0, t) + λ1ut (0, t) −∫ t

0k(t − s)u(0, s)ds.

(Pλ1)

We shall study the asymptotic expansion of the solution of problem (Pλ1) with respect to λ1.

Theorem 2. Let T > 0. Let (H1)–(H4) hold. Then(i) The problem (P0) corresponding to λ1 = 0 has only the solution

(u0, P0

)satisfying{

u0 ∈ L∞(0, T ; V ), u ′0 ∈ L∞(0, T ; L2),

u0(0, ·), P0 ∈ H 1 (0, T ) .(3.2)

(ii) The solution(uλ1, Pλ1

)converges strongly in W (QT ) × L2 (0, T ) to

(u0, P0

), as λ1 →

0+,where

W (QT ) = {v ∈ L∞(0, T ; V ) : vt ∈ L∞(0, T ; L2)}.Furthermore, we have the estimates⎧⎪⎪⎨⎪⎪⎩

∥∥u′λ1

− u ′0

∥∥L∞(0,T ;L2)

+ ∥∥uλ1 − u0∥∥

L∞(0,T ;V )

+ √λ1

∥∥∥u′λ1

(0, ·) − u ′0(0, ·)

∥∥∥L2(0,T )

≤ CT√

λ1,∥∥Pλ1 − P0∥∥

L2(0,T )≤ CT

√λ1,

(3.3)

where CT is a positive constant depending only on T .

Proof. (i) Case λ1 = 0. We prove in the same manner as in [7] (Theorem 1, p. 1263) that thefollowing estimations hold a priori∥∥u′

m(t)∥∥2 + ‖umx (t)‖2 + K1u2

m(0, t) ≤ CT , ∀t ∈ [0, T ],∀T > 0, (3.4)∫ T

0

∣∣u′m(0, t)

∣∣2 dt ≤ CT ,

∫ T

0

∣∣P ′m(t)

∣∣2 dt ≤ CT , ∀T > 0, (3.5)

and that the limit(u0, P0

)of the sequence (um, Pm) defined by (2.6)–(2.8) satisfies (3.2) and the

problem (P0) corresponding to λ1 = 0. Furthermore, this solution(u0 , P0

)is unique.

(i) Consider λ0 > 0 fixed and the parameter λ1 ∈ (0, λ0). Proving in the same manner as inthe case of Theorem 1 with 0 < λ1 < λ0, we have the following results:∥∥u′

λ1(t)∥∥2 + ∥∥uλ1x (t)

∥∥2 + K1∣∣uλ1(0, t)

∣∣2+ 2λ1

∫ t

0

∣∣u′λ1

(0, s)∣∣2 ds ≤ CT , ∀t ∈ [0, T ],∀T > 0, (3.6)∥∥Pλ1

∥∥L2(0,T )

≤ CT , (3.7)√λ1∥∥u′

λ1(0, ·)∥∥

L2(0,T )≤ CT , (3.8)

where CT is a constant independent of λ1.

Page 12: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1537

Let λ1m be a sequence such that λ1m > 0, λ1m → 0 as m → ∞. From (3.6)–(3.8), we deducethat, there exists a subsequence of the sequence {(uλ1m , Pλ1m )} still denoted by {(uλ1m , Pλ1m )},such that⎧⎪⎪⎪⎪⎨⎪⎪⎪⎪⎩

uλ1m → u∗ in L∞(0, T ; V ) weak*,u′

λ1m→ u′∗ in L∞(0, T ; L2) weak*,

uλ1m (0, ·) → u∗(0, ·) in L∞(0, T ) weakly*,√λ1mu′

λ1m(0, ·) → ζ∗ in L2(0, T ) weakly,

Pλ1m → P∗ in L2(0, T ) weakly.

(3.9)

By the compactness lemma of Lions [4, p. 57] and the imbeddings H 1(0, T ) ↪→ C0 ([0, T ]),we can deduce from (3.9)1,2,3,4 the existence of a subsequence still denoted by {uλ1m }, such that{

uλ1m → u∗ strongly in L2(QT ) and a.e. in QT ,√λ1muλ1m (0, ·) → 0 strongly in C0 ([0, T ]) .

(3.10)

It follows from (3.9)4, (3.10)2 that

ζ∗ = 0. (3.11)

From (3.9)3,4,5, (3.10)2 and (3.11), we have

Pλ1m (t) = g(t) + K1uλ1m (0, t) + λ1mu′λ1m

(0, t) −∫ t

0k(t − s)uλ1m (0, s)ds

→ g(t) + K1u∗(0, t) −∫ t

0k(t − s)u∗(0, s)ds = P∗(t), (3.12)

in L2(0, T ) weakly.By passing to the limit similarly to in the proof of Theorem 1, we conclude that (u∗, P∗) is a

solution of the problem (P0) corresponding to λ1 = 0 satisfying (3.2). From the uniqueness ofthe solution we have

(u∗, P∗) = (u0, P0

). (3.13)

Put u = uλ1 − u0, P = Pλ1 − P0, then (u, P) satisfies⎧⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎩

Au ≡ utt − ux x + K u + λut = 0, 0 < x < 1, 0 < t < T,

ux(0, t) = P(t), u(1, t) = 0,

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

P(t) = λ1u ′0(0, t) + K1u(0, t) + λ1u′(0, t) −

∫ t

0k(t − s)u(0, s)ds,

u ∈ L∞(0, T ; V ), ut ∈ L∞(0, T ; L2), u(0, ·) ∈ H 1 (0, T ) , P ∈ L2 (0, T ) .

(3.14)

Using Lemma 2 again, we prove, in a similar manner to that in the above part, that

Z(t) + 2K1u2(0, t) + 2λ1

∫ t

0

∣∣u′(0, s)∣∣2 ds + 4λ

∫ t

0

∥∥u′(s)∥∥2 ds

≤ 2λ1

∫ t

0

∣∣u′0(0, s)

∣∣2 ds + 2CT

∫ t

0Z(r)dr, (3.15)

where

Z(t) = ∥∥u′(t)∥∥2 + ‖ux (t)‖2 , (3.16)

CT = 1 + |K | + 2 |k(0)| + 2 ‖k‖2L2(0,T )

+ T∥∥k ′∥∥2

L2(0,T ). (3.17)

Page 13: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1538 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546

Using Gronwall’s lemma, it follows from (3.15)–(3.17) that

Z(t) + 2K1u2(0, t) + 2λ1

∫ t

0

∣∣u′(0, s)∣∣2 ds ≤ 2λ1

∥∥u ′0(0, ·)∥∥2

L2(0,T )exp

(2CT t

)≤ 2λ1

∥∥u ′0(0, ·)∥∥2

L2(0,T )exp

(2CT T

) = KT λ1. (3.18)

This implies∥∥u′λ1

− u ′0

∥∥L∞(0,T ;L2)

+ ∥∥uλ1 − u0∥∥

L∞(0,T ;V )

+√λ1∥∥u′

λ1(0, ·) − u ′

0(0, ·)∥∥L2(0,T )

≤ CT

√λ1, (3.19)

where CT is a constant depending only on T .On the other hand, it follows from (3.14)4, (3.15) that

‖P‖L2 ≤ λ1(∥∥u ′

0(0, ·)∥∥L2 + ∥∥u′(0, ·)∥∥L2

)+(

K1 + √T∥∥k ′∥∥

L2

)‖u(0, ·)‖L2 . (3.20)

Hence∥∥Pλ1 − P0∥∥

L2(0,T )≤ CT

√λ1, (3.21)

where CT is a constant depending only on T . Theorem 2 is proved completely. �

The next result gives an asymptotic expansion of the weak solution(uλ1, Pλ1

)of order N + 1

2in λ1, for λ1 sufficiently small.

Let(u0, P0

)be a weak solution of problem (P0) (corresponding to λ1 = 0) as in Theorem 2.⎧⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎩

Au0 ≡ u ′′0 − u0x x + K u0 + λu ′

0 = F(x, t), 0 < x < 1, 0 < t < T,

u0x(0, t) = P0(t), u0(1, t) = 0,

u0(x, 0) = u0(x), u ′0(x, 0) = u1(x),

P0(t) = g(t) + K1u0(0, t) −∫ t

0k(t − s )u0(0, s)ds,

u0 ∈ L∞(0, T ; V ), u ′0 ∈ L∞(0, T ; L2), u0(0, ·), P0 ∈ H 1 (0, T ) .

(P0)

Let us consider the sequence of weak solutions(ui , Pi

), i = 1, 2, . . . , N , defined by the

following problems:⎧⎪⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎪⎩

Aui ≡ u ′′i − uix x + K ui + λu ′

i = 0, 0 < x < 1, 0 < t < T,

uix (0, t) = Pi (t), ui (1, t) = 0,

ui (x, 0) = u ′i (x, 0) = 0,

Pi (t) = K1ui (0, t) + u ′i−1(0, t) −

∫ t

0k(t − s )ui (0, s)ds,

ui ∈ L∞(0, T ; V ), u ′i ∈ L∞(0, T ; L2), ui (0, ·) ∈ H 1 (0, T ) , Pi ∈ L2 (0, T ) .

(Pi )

Let(uλ1, Pλ1

)be a unique weak solution of problem (Pλ1). Then (v, Q), with

v = uλ1 −N∑

i=0

uiλi1, Q = Pλ1 −

N∑i=0

Piλi1, (3.22)

Page 14: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1539

satisfies the problem⎧⎪⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎪⎩

Av ≡ vt t − vx x + Kv + λvt = 0, 0 < x < 1, 0 < t < T,

vx (0, t) = Q(t), v(1, t) = 0,

v(x, 0) = vt (x, 0) = 0,

Q(t) = λN+11 u ′

N (0, t) + K1v(0, t) + λ1v′(0, t) −

∫ t

0k(t − s)v(0, s)ds,

v ∈ L∞(0, T ; V ), vt ∈ L∞(0, T ; L2), v(0, ·) ∈ H 1 (0, T ) , Q ∈ L2 (0, T ) .

(3.23)

Using again Lemma 2, we prove, in a manner similar to that of Theorem 2, that

Z(t) + 2K1v2(0, t) + 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds + 4λ

∫ t

0

∥∥v′(s)∥∥2 ds

≤ 2λ2N+11

∫ t

0

∣∣u′N (0, s)

∣∣2 ds + 2CT

∫ t

0Z(r)dr, (3.24)

where{Z(t) = ∥∥v′(t)

∥∥2 + ‖vx (t)‖2 ,

CT = 1 + |K | + 2 |k(0)| + 2 ‖k‖2L2(0,T )

+ T∥∥k ′∥∥2

L2(0,T ).

(3.25)

Using Gronwall’s lemma, it follows from (3.24) and (3.25) that

Z(t) + 2K1v2(0, t) + 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds ≤ 2λ2N+1

1

∥∥u ′N (0, ·)∥∥2

L2(0,T )exp

(2CT t

)≤ 2λ2N+1

1

∥∥u ′N (0, ·)∥∥2

L2(0,T )exp

(2CT T

) ≡ KT λ2N+11 . (3.26)

This implies∥∥∥∥∥u′λ1

−N∑

i=0

u ′i λ

i1

∥∥∥∥∥L∞(0,T ;L2)

+∥∥∥∥∥uλ1 −

N∑i=0

uiλi1

∥∥∥∥∥L∞(0,T ;V )

+√λ1

∥∥∥∥∥u′λ1

(0, ·) −N∑

i=0

u ′i (0, ·)λi

1

∥∥∥∥∥L2(0,T )

≤ CT λN+ 1

21 , (3.27)

∥∥∥∥∥Pλ1 −N∑

i=0

Piλi1

∥∥∥∥∥L2(0,T )

≤ CT λN+ 1

21 , (3.28)

where CT is a constant depending only on T .Thus, we have the following theorem.

Theorem 3. Let (H1)–(H4) hold. Then, for every λ1 ∈ (0, λ0), problem (Pλ1) has a uniqueweak solution

(uλ1, Pλ1

)satisfying the asymptotic estimations up to order N + 1

2 as in (3.27)and (3.28), the functions

(ui , Pi

), i = 1, 2, . . . , N, being the weak solutions of problems (P0),(

P1), . . . ,

(PN), respectively. �

4. Asymptotic expansion of the solution with respect to four parameters (K, λ, K1, λ1)

In this part, we assume that α = β = 2 and (u0, u1, F , g, k) satisfy the assumptions(H1)–(H3). Let (K , λ, K1, λ1) ∈ R × R

3+. By Theorem 1, the problem (1.1)–(1.5) has a unique

Page 15: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1540 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546

weak solution (u, P) depending on (K , λ, K1, λ1):

u = u (K , λ, K1, λ1) , P = P (K , λ, K1, λ1) . (4.1)

We consider the following perturbed problem, where K , λ, K1, λ1 are small parameters suchthat, |K | ≤ K∗, 0 ≤ λ ≤ λ∗, 0 ≤ K1 ≤ K1∗, 0 ≤ λ1 ≤ λ1∗:⎧⎪⎪⎪⎨⎪⎪⎪⎩

Lu ≡ utt − ux x = −K u − λut + F(x, t), 0 < x < 1, 0 < t < T,

ux (0, t) = P(t), u(1, t) = 0,

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

P(t) = g(t) + K1u(0, t) + λ1ut (0, t) −∫ t

0k(t − s)u(0, s)ds.

(PK ,λ,K1,λ1 )

We shall study the asymptotic expansion of the solution of problem (PK ,λ,K1,λ1) with respect to(K , λ, K1, λ1).

We use the following notation. For a multi-index γ = (γ1, γ2, γ3, γ4) ∈ Z4+, and

−→K =

(K , λ, K1, λ1) ∈ R × R3+, we put⎧⎪⎨⎪⎩

|γ | = γ1 + γ2 + γ3 + γ4, γ ! = γ1!γ2!γ3!γ4!,∥∥∥−→K ∥∥∥ =√

K 2 + λ2 + K 21 + λ2

1,−→K

γ = K γ1λγ2 K γ31 λ

γ41 ,

α, β ∈ Z4+, α ≤ β ⇐⇒ αi ≤ βi ∀i = 1, 2, 3, 4.

(4.2)

Let (u0,0,0,0, P0,0,0,0) ≡ (u0, P0) be a unique weak solution of problem (P0,0,0,0) (as inTheorem 1) corresponding to

−→K = 0, i.e.,

⎧⎪⎪⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎪⎪⎩

Lu0 = F0,0,0,0 ≡ F(x, t), 0 < x < 1, 0 < t < T,

u0x(0, t) = P0(t), u0(1, t) = 0,

u0(x, 0) = u0(x), u′0(x, 0) = u1(x),

P0(t) = g(t) −∫ t

0k(t − s)u0(0, s)ds,

u0 ∈ L∞(0, T ; V ), u′0 ∈ L∞(0, T ; L2),

u0(0, ·), P0 ∈ H 1 (0, T ) .

(P0,0,0,0)

Let us consider the sequence of weak solutions(uγ , Pγ

), γ ∈ Z

4+, 1 ≤ |γ | ≤ N , defined by thefollowing problems:⎧⎪⎪⎪⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎪⎪⎪⎩

Luγ ≡ uγ t t − uγ x x = Fγ , 0 < x < 1, 0 < t < T,

uγ x (0, t) = Pγ (t), uγ (1, t) = 0,

uγ (x, 0) = u′γ (x, 0) = 0,

Pγ (t) = Cγ (t) −∫ t

0k(t − s)uγ (0, s)ds,

uγ ∈ L∞ (0, T ; V ) , u′γ ∈ L∞(0, T ; L2),

uγ (0, ·) ∈ H 1 (0, T ) , Pγ ∈ L2 (0, T ) ,

(Pγ )

where Fγ , Cγ (t), |γ | ≤ N , defined by the formulas

Page 16: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1541

Fγ =

⎧⎪⎪⎪⎪⎨⎪⎪⎪⎪⎩F, |γ | = 0,

0, 1 ≤ |γ | ≤ N, γ1 = 0, γ2 = 0,

−u′γ1,γ2−1,γ3,γ4

, 1 ≤ |γ | ≤ N, γ1 = 0, γ2 ≥ 1,

−uγ1−1,γ2,γ3,γ4, 1 ≤ |γ | ≤ N, γ1 ≥ 1, γ2 = 0,

−uγ1−1,γ2,γ3,γ4 − u′γ1,γ2−1,γ3,γ4

, 1 ≤ |γ | ≤ N, γ1 ≥ 1, γ2 ≥ 1,

(4.3)

and

Cγ (t) =

⎧⎪⎪⎪⎪⎨⎪⎪⎪⎪⎩g(t), |γ | = 0,

0, 1 ≤ |γ | ≤ N, γ3 = 0, γ4 = 0,

u′γ1,γ2,γ3,γ4−1(0, t), 1 ≤ |γ | ≤ N, γ3 = 0, γ4 ≥ 1,

uγ1,γ2,γ3−1,γ4(0, t), 1 ≤ |γ | ≤ N, γ3 ≥ 1, γ4 = 0,

uγ1,γ2,γ3−1,γ4(0, t) + u′γ1,γ2,γ3,γ4−1(0, t), 1 ≤ |γ | ≤ N, γ3 ≥ 1, γ4 ≥ 1.

(4.4)

Let (u, P) = (uK ,λ,K1,λ1, PK ,λ,K1,λ1

)be a unique weak solution of problem (PK ,λ,K1,λ1). Then

(v, R), with

v = u −∑

|γ |≤N

uγ−→K

γ, R = P −

∑|γ |≤N

Pγ−→K

γ, (4.5)

satisfies the problem⎧⎪⎪⎪⎨⎪⎪⎪⎩vt t − vx x + Kv + λvt = eN (x, t), 0 < x < 1, 0 < t < T,

vx (0, t) = R(t), v(1, t) = 0,

v(x, 0) = vt (x, 0) = 0,

R(t) = eN (t) + K1v(0, t) + λ1v′(0, t) −

∫ t

0k(t − s)v(0, s)ds,

(4.6)

where

eN (x, t) = −∑

|γ |=N

(K uγ + λu′

γ

)−→K

γ, (4.7)

eN (t) =∑

|γ |=N

(K1uγ (0, t) + λ1u′

γ (0, t))−→

Kγ. (4.8)

Then, we have the following lemma.

Lemma 3. Let (H1)–(H3) hold. Then

‖eN ‖L∞(0,T ;L2) ≤ D1N

∥∥∥−→K ∥∥∥N+1, (4.9)

‖eN ‖L2(0,T ) ≤ D2N

∥∥∥−→K ∥∥∥N+1, (4.10)

4

∣∣∣∣∫ t

0eN (s)v′(0, s)ds

∣∣∣∣ ≤ K1v2(0, t) + 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds

+∫ t

0|v(0, s)|2 ds + D3N

∥∥∥−→K ∥∥∥2N+1, (4.11)

where D1N , D2N and D3N are positive constants depending only on the constants∥∥∥−→K ∗

∥∥∥,∥∥uγ

∥∥L∞(0,T ;V )

,∥∥∥u′

γ

∥∥∥L∞(0,T ;L2)

,∥∥uγ (0, ·)∥∥H1(0,T )

,∥∥∥u′

γ (0, ·)∥∥∥

L2(0,T ), (|γ | = N).

Page 17: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1542 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546

Proof. (i) By the boundedness of the functions uγ , u′γ , γ ∈ Z

4+, |γ | = N in the function space

L∞(0, T ; L2), we obtain from (4.7) that

‖eN ‖L∞(0,T ;L2) ≤∑

|γ |=N

(|K |∥∥uγ

∥∥L∞(0,T ;V )

+ λ

∥∥∥u′γ

∥∥∥L∞(0,T ;L2)

) ∣∣∣−→K γ∣∣∣ . (4.12)

On the other hand, using the Holder inequality xα11 xα2

2 xα33 xα4

4 ≤ α1x1 + α2x2 + α3x3 + α4x4,∀xi ≥ 0, ∀αi ≥ 0, α1 + α2 + α3 + α4 = 1 with x1 = K 2, x2 = λ2, x3 = K 2

1 , x4 = λ21, α1 = γ1

N ,α2 = γ2

N , α3 = γ3N , α4 = γ4

N , we obtain

∣∣∣−→K γ∣∣∣ = ∣∣K γ1λγ2 K γ3

1 λγ41

∣∣ =(

K 2γ1N λ2

γ2N K

2γ3N

1 λ2

γ4N

1

) N2

≤(γ1

NK 2 + γ2

Nλ2 + γ3

NK 2

1 + γ4

Nλ2

1

) N2

≤(

K 2 + λ2 + K 21 + λ2

1

) N2 =

∥∥∥−→K ∥∥∥N, (4.13)

for all γ ∈ Z4+, |γ | = N .

Therefore, it follows from (4.12) and (4.13) that

‖eN ‖L∞(0,T ;L2) ≤ D1N

∥∥∥−→K ∥∥∥N+1, (4.14)

where

D1N =∑

|γ |=N

(∥∥uγ

∥∥L∞(0,T ;V )

+∥∥∥u′

γ

∥∥∥L∞(0,T ;L2)

). (4.15)

(ii) With eN (t), then, we obtain from (4.8), in a manner similar to that of the above part,that

‖eN ‖H2(0,T ) ≤ D2N

∥∥∥−→K ∥∥∥N+1, (4.16)

where

D2N =∑

|γ |=N

(∥∥uγ (0, ·)∥∥H1(0,T )+∥∥∥u′

γ (0, ·)∥∥∥

L2(0,T )

). (4.17)

(iii) From (4.8), we have

4∫ t

0eN (s)v′(0, s)ds = 4

∑|γ |=N

K1

∫ t

0uγ (0, s)v′(0, s)ds

−→K

γ

+ 4∑

|γ |=N

λ1

∫ t

0u′

γ (0, s)v′(0, s)ds−→K

γ

≡ J1N (t) + J2N (t). (4.18)

We shall estimate respectively the following terms on the right-hand side of (4.18).

Page 18: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1543

First term J1N (t). Integrating by parts, we have

J1N (t) = 4∑

|γ |=N

K1

∫ t

0uγ (0, s)v′(0, s)ds

−→K

γ

= 4K1v(0, t)∑

|γ |=N

uγ (0, t)−→K

γ − 4K1

∑|γ |=N

∫ t

0u′

γ (0, s)v(0, s)ds−→K

γ. (4.19)

By (4.13), it follows from (4.19) that

J1N (t) ≤ K1v2(0, t) + 4K1

( ∑|γ |=N

∣∣uγ (0, t)∣∣ ∣∣∣−→K γ

∣∣∣)2

+ 4K1

∑|γ |=N

(∫ t

0

∣∣∣u′γ (0, s)

∣∣∣2 ds

)1/2(∫ t

0|v(0, s)|2 ds

)1/2 ∣∣∣−→K γ∣∣∣

≤ K1v2(0, t) +

∫ t

0|v(0, s)|2 ds

+ 4

⎡⎣( ∑|γ |=N

∥∥uγ

∥∥L∞(0,T ;V )

)2

+( ∑

|γ |=N

∥∥∥u′γ (0, ·)

∥∥∥L2(0,T )

)2 ∥∥∥−→K ∗∥∥∥⎤⎦

×∥∥∥−→K ∥∥∥2N+1

≤ K1v2(0, t) +

∫ t

0|v(0, s)|2 ds

+ 4(

1 +∥∥∥−→K ∗

∥∥∥)[ ∑|γ |=N

(∥∥uγ

∥∥L∞(0,T ;V )

+∥∥∥u′

γ (0, ·)∥∥∥

L2(0,T )

)]2

×∥∥∥−→K ∥∥∥2N+1

, (4.20)

where∥∥∥−→K ∥∥∥ ≤

∥∥∥−→K ∗∥∥∥, with

−→K ∗ = (K∗, λ∗, K1∗, λ1∗).

Second term J2N (t). Using again (4.13), we obtain

J2N (t) = 4∑

|γ |=N

λ1

∫ t

0u′

γ (0, s)v′(0, s)ds−→K

γ ≤ 4λ1

∑|γ |=N

∫ t

0

∣∣∣u′γ (0, s)v′(0, s)

∣∣∣ ds∣∣∣−→K γ

∣∣∣≤ 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds + 2λ1

( ∑|γ |=N

∥∥∥u′γ (0, ·)

∥∥∥L2(0,T )

)2 ∥∥∥−→K ∥∥∥2N

≤ 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds + 2

( ∑|γ |=N

∥∥∥u′γ (0, ·)

∥∥∥L2(0,T )

)2 ∥∥∥−→K ∥∥∥2N+1. (4.21)

Combining (4.18), (4.20) and (4.21), we then have

Page 19: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1544 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546∣∣∣∣4 ∫ t

0eN (s)v′(0, s)ds

∣∣∣∣ ≤ K1v2(0, t) + 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds

+∫ t

0|v(0, s)|2 ds + D3N

∥∥∥−→K ∥∥∥2N+1, (4.22)

where

D3N = 4(

1 +∥∥∥−→K ∗

∥∥∥)[ ∑|γ |=N

(∥∥uγ

∥∥L∞(0,T ;V )

+∥∥∥u′

γ (0, ·)∥∥∥

L2(0,T )

)]2

+ 2

( ∑|γ |=N

∥∥∥u′γ (0, ·)

∥∥∥L2(0,T )

)2

. (4.23)

The proof of Lemma 3 is complete. �Next, we obtain the following theorem.

Theorem 4. Let (H1)–(H3) hold. Then, for every (K , λ, K1, λ1) ∈ R × R3+, with |K | ≤ K∗,

0 ≤ λ ≤ λ∗, 0 ≤ K1 ≤ K1∗, 0 ≤ λ1 ≤ λ1∗, problem (PK ,λ,K1,λ1) has a unique weak solution(u, P) = (

uK ,λ,K1,λ1, PK ,λ, K1,λ1

)satisfying the asymptotic estimations up to order N + 1

2 asfollows∥∥∥∥∥u′ −

∑|γ |≤N

u′γ−→K

γ

∥∥∥∥∥L∞(0,T ;L2)

+∥∥∥∥∥u −

∑|γ |≤N

uγ−→K

γ

∥∥∥∥∥L∞(0,T ;V )

+√λ1

∥∥∥∥∥u′(0, ·) −∑

|γ |≤N

u′γ (0, ·)−→K γ

∥∥∥∥∥L2(0,T )

≤ D∗N

∥∥∥−→K ∥∥∥N+ 12, (4.24)

and ∥∥∥∥∥P −∑

|γ |≤N

Pγ−→K

γ

∥∥∥∥∥L2(0,T )

≤ D∗∗N

∥∥∥−→K ∥∥∥N+ 12, (4.25)

for all (K , λ, K1, λ1) ∈ R × R3+, |K | ≤ K∗, 0 ≤ λ ≤ λ∗, 0 ≤ K1 ≤ K1∗, 0 ≤ λ1 ≤ λ1∗, the

functions (uγ , Pγ ) being the weak solutions of problems (Pγ ), γ ∈ Z4+, |γ | ≤ N.

Remark 2. In [10], as in this special case for problem (1.1)–(1.5), Long, Ut, and Truc haveobtained a result about the asymptotic expansion of the solutions with respect to two parameters(K , λ) up to order N + 1.

Proof. Put{Z(t) = ∥∥v′(t)

∥∥2 + ‖vx (t)‖2 ,

CT = 1 + |K | + 2 |k(0)| + 2 ‖k‖2L2(0,T )

+ T∥∥k ′∥∥2

L2(0,T ).

(4.26)

Using Lemma 2 again, we prove, in a manner similar to that of Theorem 2, that

Z(t) + 2K1v2(0, t) + 4λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds

≤ 2CT

∫ t

0Z(s)ds + 4

∫ t

0

⟨eN (s), v′(s)

⟩ds + 4

∫ t

0eN (s)v′(0, s)ds. (4.27)

Page 20: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546 1545

Using Lemma 3, it follows from (4.27) that

Z(t) + 2K1v2(0, t) + 4λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds

≤ 2CT

∫ t

0Z(s)ds + 4T D2

1N

∥∥∥−→K ∥∥∥2N+2 +∫ t

0

∥∥v′(s)∥∥2

ds

+ K1v2(0, t) + 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds +

∫ t

0|v(0, s)|2 ds + D3N

∥∥∥−→K ∥∥∥2N+1

= K1v2(0, t) + 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds + 2

(1 + CT

) ∫ t

0Z(s)ds

+(

4T D21N

∥∥∥−→K ∗∥∥∥+ D3N

) ∥∥∥−→K ∥∥∥2N+1, (4.28)

where∥∥∥−→K ∥∥∥ ≤

∥∥∥−→K ∗∥∥∥, with

−→K ∗ = (K∗, λ∗, K1∗, λ1∗). Hence

Z(t) + K1v2(0, t) + 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds

≤(

4T D21N

∥∥∥−→K ∗∥∥∥+ D3N

) ∥∥∥−→K ∥∥∥2N+1 + 2(1 + CT

) ∫ t

0Z(s)ds. (4.29)

Using Gronwall’s lemma, it follows from (4.29) that

Z(t) + 2K1v2(0, t) + 2λ1

∫ t

0

∣∣v′(0, s)∣∣2 ds

≤(

4T D21N

∥∥∥−→K ∗∥∥∥+ D3N

) ∥∥∥−→K ∥∥∥2N+1exp

(2(1 + CT

)T). (4.30)

Hence∥∥v′∥∥L∞(0,T ;L2)

+ ‖v‖L∞(0,T ;V ) +√λ1∥∥v′(0, ·)∥∥L2(0,T )

≤ D∗N

∥∥∥−→K ∥∥∥N+ 12, (4.31)

where

D∗N = 3

√4T D2

1N + D3N exp((

1 + CT)

T), (4.32)

or ∥∥∥∥∥u′ −∑

|γ |≤N

u′γ−→K

γ

∥∥∥∥∥L∞(0,T ;L2)

+∥∥∥∥∥u −

∑|γ |≤N

uγ−→K

γ

∥∥∥∥∥L∞(0,T ;V )

+√λ1

∥∥∥∥∥u′(0, ·) −∑

|γ |≤N

u′γ (0, ·)−→K γ

∥∥∥∥∥L2(0,T )

≤ D∗N

∥∥∥−→K ∥∥∥N+ 12. (4.33)

On the other hand, it follows from (4.6)4, (4.31) that

‖R‖L2(0,T ) ≤ ‖eN ‖L2(0,T ) +(

K1 + √T∥∥k ′∥∥

L2(0,T )

)‖v(0, ·)‖L2(0,T )

+ λ1∥∥v′(0, ·)∥∥L2(0,T )

Page 21: A wave equation associated with mixed nonhomogeneous conditions: Global existence and asymptotic expansion of solutions

1546 N.T. Long, V.G. Giai / Nonlinear Analysis 66 (2007) 1526–1546

≤[

D2N

√∥∥∥−→K ∗∥∥∥+

(K1 + √

T∥∥k ′∥∥

L2(0,T )

)D∗

N +√λ1∗ D∗

N

] ∥∥∥−→K ∥∥∥N+ 12

≡ D∗∗N

∥∥∥−→K ∥∥∥N+ 12, (4.34)

where

D∗∗N =

√∥∥∥−→K ∗∥∥∥D2N +

(K1 + √

T∥∥k ′∥∥

L2(0,T )+√

λ1∗)

D∗N , (4.35)

and hence∥∥∥∥∥P −∑

|γ |≤N

Pγ−→K

γ

∥∥∥∥∥L2(0,T )

≤ D∗∗N

∥∥∥−→K ∥∥∥N+ 12, (4.36)

Theorem 4 is proved completely. �

Acknowledgement

We wish to acknowledge the referee for constructive remarks and corrections to themanuscript.

References

[1] D.D. Ang, A.P.N. Dinh, Mixed problem for some semilinear wave equation with a nonhomogeneous condition,Nonlinear Anal. 12 (1988) 581–592.

[2] N.T. An, N.D. Trieu, Shock between absolutely solid body and elastic bar with the elastic viscous frictionalresistance at the side, J. Mech. NCSR Vietnam 13 (2) (1991) 1–7.

[3] M. Bergounioux, N.T. Long, A.P.N. Dinh, Mathematical model for a shock problem involving a linear viscoelasticbar, Nonlinear Anal. 43 (2001) 547–561.

[4] A.P.N. Dinh, N.T. Long, Linear approximation and asymptotic expansion associated to the nonlinear wave equationin one dimension, Demonstratio Math. 19 (1986) 45–63.

[5] J.L. Lions, Quelques methodes de resolution des proble mes aux limites nonlineaires, Gauthier-Villars, Paris,Dunod, 1969.

[6] N.T. Long, A.P.N. Dinh, On the quasilinear wave equation: utt − �u + f (u, ut ) = 0 associated with a mixednonhomogeneous condition, Nonlinear Anal. 19 (1992) 613–623.

[7] N.T. Long, A.P.N. Dinh, A semilinear wave equation associated with a linear differential equation with Cauchydata, Nonlinear Anal. 24 (1995) 1261–1279.

[8] N.T. Long, T.N. Diem, On the nonlinear wave equation utt − uxx = f (x, t, , u, ux , ut ) associated with the mixedhomogeneous conditions, Nonlinear Anal. 29 (1997) 1217–1230.

[9] N.T. Long, A.P.N. Dinh, T.N. Diem, On a shock problem involving a nonlinear viscoelastic bar, J. Boundary ValueProblems, Hindawi Publishing Corporation 2005 (3) (2005) 337–358.

[10] N.T. Long, L.V. Ut, N.T.T. Truc, On a shock problem involving a linear viscoelastic bar, Nonlinear Anal. 63 (2)(2005) 198–224.