You are on page 1of 23

Electronic Journal of Qualitative Theory of Differential Equations

2010, No. 18, 1-23; http://www.math.u-szeged.hu/ejqtde/

On the viscous Burgers equation in unbounded domain

j. lı́maco 1 , h. r. clark 1, 2 & l. a. medeiros 3

Abstract
In this paper we investigate the existence and uniqueness of global solutions, and
a rate stability for the energy related with a Cauchy problem to the viscous Burgers
equation in unbounded domain R × (0, ∞). Some aspects associated with a Cauchy
problem are presented in order to employ the approximations of Faedo-Galerkin in
whole real line R. This becomes possible due to the introduction of weight Sobolev
spaces which allow us to use arguments of compactness in the Sobolev spaces.

Key words: Unbounded domain, global solvability, uniqueness, a rate decay estimate
for the energy.

AMS subject classifications codes: 35B40, 35K15, 35K55, 35R35.

1 Introduction and Formulation of the Problem
We are concerned with the existence of global solutions – precisely, global weak
solutions, global strong solutions and regularity of the strong solutions –, uniqueness
of the solutions and the asymptotic stability of the energy for the nonlinear Cauchy
problem related to the classic viscous Burgers equation

ut + uux − uxx = 0

established in R × (0, T ), for an arbitrary T > 0. More precisely, we consider the real
valued function u = u(x, t) defined for all (x, t) ∈ R × (0, T ) which is the solution of the
Cauchy problem
(
ut + uux − uxx = 0 in R × (0, T ),
(1.1)
u(x, 0) = u0 (x) in R.
1
Universidade Federal Fluminense, IM, RJ, Brasil
2
Corresponding author: hclark@vm.uff.br
3
Universidade Federal do Rio de Janeiro, IM, RJ, Brasil, lmedeiros@abc.org.br

EJQTDE, 2010 No. 18, p. 1

The Burgers equation has a long history. We briefly sketch this history by citing
one of the pioneer work by Bateman [2] about an approximation of the flux of fluids.
Later, Burgers published the works [5] and [6] which are also about flux of fluids or
turbulence. In the classic fashion the Burgers equation has been studied by several
authors, mainly in the last century, and excellent papers and books can be found in
the literature specialized in PDE. One can cite, for instance, Courant & Friedrichs [8],
Courant & Hilbert [9], Hopf [10], Lax [13] and Stoker [17].
Today, the equation (1.1)1 ((1.1)1 refers to the first equation in (1.1)) is known as
viscous Burgers equations and perhaps it is the simplest nonlinear equation associating
the nonlinear propagation of waves with the effect of the heat conduction.
The existence of global solutions for the Cauchy problem (1.1) will be obtained
employing the Faedo-Galerkin and Compactness methods. The Faedo-Galerkin method
is probably one of the most effective methods to establish existence of solutions for non-
linear evolution problems in domains whose spatial variable x lives in bounded sets. To
spatial unbounded sets, there exist few results about existence of solutions established
by the referred method. Thus, as the non-linear problem (1.1) is defined in R, in order
to reach our goal through this method we will also need to use compactness’ argument,
as in Aubin [1] or Lions [15]. In order to apply the Compactness method we employ
a suitable theory on weight Sobolev spaces to be set as follows. In fact, in the sequel
H m (R) represents the Sobolev space of order m in R, with m ∈ N. The space L2 (R)
is the Lebesgue space of the classes of functions u : R → R with square integrable on
R. Assuming that X is a Banach space, T is a positive real number or T = +∞ and
1 ≤ p ≤ ∞, we will denote by Lp (0, T ; X) the Banach space of all measurable mapping
u :]0, T [−→ X, such that t 7→ ku(t)kX belongs to Lp (0, T ). For more details on the
functional spaces above cited the reader can consult, for instance, the references [3] and
[15]. In this work we will also use the following weight vectorial spaces
n Z o
2 2
L (K) = φ ∈ L (R); |φ(y)|2R K(y)dy < ∞ ,
n R o
H (K) = φ ∈ H (R); D i φ ∈ L2 (K) with i = 1, 2, . . . , m, m ∈ N,
m m

where K is a weight function given for

K(y) = exp{y 2 /4}, y ∈ R. (1.2)

EJQTDE, 2010 No. 18, p. 2

The inner product and norm of L2 (K) and Hm (K) are defined by Z Z 2 (φ. ψ) = φ(y)ψ(y)K(y)dy. |φ| = |φ(y)|2R K(y)dy. R R m Z X m X .

i .

2 ((φ. ψ))m = D i φ(y)Di ψ(y)K(y)dy. kφk2m = .

D φ.

1) is to transform it to another equivalent one proposed in the suitable functional spaces by using a change of variables defined by x z(y. S) with σ(x. es − 1) and u(x. p. H) with 1 ≤ p ≤ ∞. ln(t + 1) . i=1 R i=1 respectively. it yields −1  ∂z ∂y ∂z ∂s  ut = (t + 1)−3/2 z + (t + 1)−1/2 + 2 ∂y ∂t ∂s ∂t  z yz  y = (t + 1)−3/2 − − + zs  z yz 2 2  y = e−3s/2 − − + zs . Therefore. Results on compactness of space of spherically symmetric functions that vanishes at infinity were proven by Strauss [18]. t) = (t + 1)−1/2 z x/(t + 1)1/2 .   z(y. see [16]. s) = es/2 u(es/2 y. Differentiating u with respect to t. T ) → Ry × (0. 18. Some properties of the spaces L2 (K) and Hm (K) as the compactness of the inclu- sion Hm (K) ֒→ L2 (K) and Poincaré inequality with the weight (1. s) and S = ln(T + 1). We will also use the functional structure of the spaces Lp (0. T .3) defines a diffeomorphism σ : Rx × (0. 3 . By D(R) it denotes the class of C ∞ functions in R with compact support and convergence in the Laurent Schwartz sense. . In this direction one can see some results in Kurtz [12]. (1. s) = (t + 1)1/2 u(x. t) = (y. The vector spaces L2 (K) and Hm (K) are Hilbert spaces with the above inner products. The method used to prove the existence of solutions for the Cauchy problem (1. t) where y = and s = ln(t + 1). From (1. 2 2 EJQTDE. where H is one of the spaces: L2 (K) or Hm (K).3) we have t = es − 1 and x = es/2 y. 2010 No.2) has been proven in Escobedo-Kavian [11].3) (t + 1)1/2 The changing of variable (1.

Thus the initial data becomes u0 (x) = u(x. S). ∂y ∂x ∂y (t + 1)1/2 Differentiating again with respect to x.4) 2 2 Moreover. 2 (1. 4 . we obtain yzy z zs − zyy − − + zzy = 0 in R × (0.5) and (1. Therefore.7)  z(y. 18. In Section 3 we establish the same properties of Section 2 for the strong solutions. we investigate the existence of global weak solutions of (1. ψ))2 (1.5) For use later and a better understanding we will modify the equation (1. it yields ∂ 2 z ∂y uxx = (t + 1)−1 = (t + 1)−3/2 zyy = e−3s/2 zyy . its uniqueness and as well as analysis of the decay of these solutions. ∞) with real values and the energy associated with this solution is asymptotically stable. (1. 2 Weak Solution Setting the initial data u0 ∈ L2 (K) we are able to show that the Cauchy problem (1. we study the regularity of the strong solutions. it yields ∂z ∂y ∂z 1 ux = (t + 1)−1/2 = (t + 1)−1/2 = (t + 1)−1 zy = e−s zy . ψ) = (φy . 2010 No.4) as follows: yφy one defines the operator L : H2 (K) −→ R by φ 7−→ Lφ = −φyy − 2 . S).1)1 . 0) = z(y. (1.1) is equivalent by σ to   zs + Lz − z + zzy = 0 in R × (0. from (1. for t = 0. 0) = z0 (y). ψy ) = ((φ.1). ∂y 2 ∂x Inserting the three last identities in (1. (1.4). 0) = u0 (y) in R. which satisfies: 1 Lφ = − (Kφy )y and (Lφ. t) defined in R×(0.1) is established in the following sense EJQTDE. p. we have by definition of y that x = y.Differentiating u with respect to x.6) K for all φ ∈ H2 (K) and ψ ∈ H1 (K). The purpose of this work is: in Section 2.6)1 the Cauchy problem (1.1) has a unique global weak solution u = u(x. In Section 4. The concept of the solutions for (1.

(4π) (4π)1/4 √ (5) Finally.1) and (1. . T ). (2. p. energy E(t) = 12 |u(t)|2 associated with this solution satisfies E(t) ≤ E(0)(t + 1)−3/4 . H 1 (R)). . one has the Poincaré inequality |v| ≤ 2 |vy | for v ∈ H1 (K) As the two Cauchy problems (1.2) The following proposition. Proposition 2. will be useful throughout this paper. whose proof has have been done in Escobedo & Kavian [11]. A global weak solution for the Cauchy problem (1.1) is a real valued function u = u(x. . ∞) such that  u ∈ L2loc (0. R R (2) The immersion H1 (K) ֒→ L(K) is compact. the function u satisfies the identity integral Z TZ Z TZ Z T Z − [uv]ϕt dxdt + [uux v]ϕdxdt + [ux vx ]ϕdxdt = 0.2 and Theorem 2. Moreover. ut ∈ L2loc 0. ∞. then there exists a unique global solution u of (1.1. One has the results Z Z (1) |y|2 |v(y)|2R Kdy ≤ 16 |vy (y)|2R Kdy for all v ∈ H1 (K).2. (4) The eigenvalues of L are positive real numbers λj = j/2 for j = 1.1) 0 R 0 R 0 R for all v ∈ H 1 (R) and for all ϕ ∈ D(0. 5 .Definition 2. 2 .1) in the sense of Definition 2. Moreover. 0) = u0 (x) for all x ∈ R.1.1.1.1 and Theorem 2. 2010 No. The existence of solution of (1. and   the related space with λj is N (L − λj I) = Dj ω1 with 1 1 ω1 (y) = 1/4 [K(y)]−1 = exp{−y 2 /4}. t) defined in R × (0.7) are equivalent the Definition 2. u satisfies the initial condition u(x. 18.1) in the precedent sense is guaranteed by the fol- lowing theorem Theorem 2. ∞. EJQTDE. [H 1 (R)]′ . (3) L : H1 (K) −→ [H1 (K)]′ is an isomorphism.1 are also equivalent to Definition 2. Suppose u0 ∈ L2 (K). (2.

the energy E(s) = 12 |z(s)|2 satisfies E(s) ≤ E(0) exp [−s/4] . Theorem 2. [H1 (K)]′ . 18.7) in the sense of Definition 2.2-item (b). then there exists a unique solution z of (1. v) for all v ∈ H1 (K) and j ∈ N. we first introduce the following property. . The existence of solutions for system (1. p. .2. v) = λj (ωj . ∞.2. s) defined in R × (0. ω2 . .2. (2.Definition 2. we are ready to state the following result. Suppose z0 ∈ L2 (K) ∩ {ω1 }⊥ .7) will be shown by means of Faedo-Galerkin method.2. it suffices to prove the Theorem 2.1 and 2. z satisfies the initial condition z(y. S). v) = 0 . Now. provided |z0 | < √1 holds.2 are equivalent.5) Since Theorems 2. ∞. ωN of (ωj )j∈N . In fact. where C1 is 4 3C1 a positive real constant defined below in the Proposition 2. 6 . 0) = z0 (y) for all y ∈ R. (2. ∞) such that  z ∈ L2loc (0. Moreover.7) is a real valued function z = z(y. In these conditions one defines VN as the subspace of L2 (K) spanned by the N − eigenfunction ω1 . Moreover. (2. .4). the function z satisfies the identity integral Z SZ Z S Z − [zv]ϕs Kdyds + [zzy v]ϕKdyds+ 0 R 0 R Z S Z Z S Z 1 [zy vy ]ϕKdyds − [zv]ϕKdyds = 0. since H1 (K) ֒→ L2 (K) is compactly imbedding there exist ωj solutions of the spectral problem associated with the operator L in H1 (K). zs ∈ L2loc 0. 2010 No.4)  Fixed the first eigenfunction ω1 of L we set {ω1 }⊥ = v ∈ L2 (K).3) 0 R 2 0 R for all v ∈ H1 (K) and for all ϕ ∈ D(0. Moreover. (ω1 . This means that (Lωj . being ωj with j ∈ N defined by (2. H1 (K)). which will be useful later: EJQTDE. as L2 (K) is a separable Hilbert space there exists a orthogonal hilber- tian basis (ωj )j∈N of L2 (K). Before this. A global weak solution for the Cauchy problem (1.

Considering v in H1 (K) we have . Proposition 2.2.

.

(a) K 1/2 v ∈ L∞ (R) and .

K 1/2 v .

√ (c) kvk1 ≤ 3|vy |. . (b) |v|L∞ (R) ≤ C1 kvk1 .L∞ (R) ≤ C1 kvk1 .

.

.

.

where C1 = 4C and C > 0 is defined by .

K 1/2 v .

L∞ (R) ≤ C .

K 1/2 v .

Moreover. for some C Proof . 2 e1 > 0 established to follow at the end of the proof below. if v ∈ H2 (K) then √ (d) |vy | ≤ 2|Lv|.H 1 (R) . (e) kvk ≤ C e1 |Lv|.As .

.

Z Z h 2 i .

1/2 .

2 |y| .

K v .

≤ |v(y)|2R Kdy + |v(y)|2R K + 2|vy (y)|2R K dy. H 1 (R) R R 8 then from Proposition 2.1 one has .

.

Z Z .

1/2 .

2 .

K v .

1 ≤ |v(y)|R Kdy + 4 |vy (y)|2R Kdy ≤ 4kvk21 . we have K 1/2 v ∈ L∞ (R) and . 2 H (R) R R As the continuous immersion H 1 (R) ֒→ L∞ (R) holds.

.

.

.

there exists C > 0 such that .

K 1/2 v .

L∞ (R) ≤ C .

K 1/2 v .

then from (a) one gets (b).H 1 (R) . Proposition 2.1-item (5). Proposition 2. Defining w = K 1/2 v one can write   1 y2 wyy = + w − K 1/2 Lv.1-item (5) one gets (d). The statement (c) is an immediate consequence from Proposition 2.1-item (5) and Proposition 2. 18. one has h i |vyy |2R ≤ C2 |w|2R + y 2 |w|2R + y 4 |w|2R + y 2 |wy |2R + |wyy |2R K −1 . v) . one has Z h i 1 y4 y2 1 y2 |wyy |2R + |w|2R + |w|2R + |w|2R + |wy |2R + |wy |2R dy = R 16 256 32 2 8 Z h i 1 3 K|Lv|R − ywy w dy ≤ |Lv|. As K 1/2 ≥ 1. R 4 2 On the other hand.2-item (c). vy ) = (Lv. Notice that for all v ∈ H2 (K) one has (vy . 4 16 From this. 7 . 2010 No. p. From these two above inequalities. From this and Proposition 2. Finally.1-item (5) and (d) one obtains (e) EJQTDE. This proves the statement (a). let v ∈ H2 (K) and Lv = f with f ∈ L2 (K).

Setting ω = z N (s) ∈ VN in (2.7) in the sense of Definition 2. the solution of the i=1 system of ordinary differential equations        z N (s). see for instance. Coddington-Levinson [7]. ωj )ωj .2.4) and this consists in finding z N (s. y) = giN (s)ωi (y) ∈ VN .Proof of Theorem 2. it yields Z 1 d . In fact.  s XN N N (2. ω − 1/2 z N (s). S[ for all S > 0 and to obtain subsequences that converge. ω + z N (s)zyN (s). The estimates to be proven later allow us to extend the solutions z N to whole interval [0.2 . The System (2. the approximate system is obtained from N P (2.6) has local solution z N in 0 ≤ s < sN .6)1 . ω = 0.6)   z (0) = z0 =  (z0 . ω + Lz N (s). j=1 for all ω belong to VN . Estimate 1. to the solution of (1. in convenient spaces.We will employ the Faedo-Galerking approximate method to prove the existence of solutions.

.

N .

.

2 .

.

N .

.

2 1 .

.

N .

.

by using Hölder inequality. In fact. 2 ds 2 R The integral above is upper bounded.2-item (b) we can write .2 z (s) + zy (s) − z (s) + z N (s)zyN (s)z N (s)Kdy = 0. Proposition 2.1-item (5) and Proposition 2.

Z .

√ .

.

2 .

.

.

.

.

z N (s)zyN (s)Lz N (s)Kdy .

≤ 3C1 .

zyN (s).

.

z N (s).

R R From this and from precedent identity we get 1 d . .

.

N .

.

2 1 .

.

N .

.

2 1  .

.

N .

.

2 .

.

N .

.

2  √ .

.

2 .

.

z (s) + zy (s) + zy (s) − z (s) ≤ 3C1 .

zyN (s).

.

z N (s).

7) 2 ds 2 2 By using (1.6)2 .4) we have N N . the fact that basis (ωj ) is orthonormal and (2. . (2.

N .

2 X X .

zy (s).

j=1 j=1 By using these two identities we are able to prove that 1 N  |zy (s)|2 − |z N (s)|2 ≥ 0 for all N ∈ N. 18.8) 2 In fact. p. (2. 2 2 2 j=2 EJQTDE. 8 . 2010 No. note that 1 N  1 N 1X |zy (s)|2 − |z N (s)|2 = (g1N (s))2 (λ1 − 1) + (gjN (s))2 (λj − 1) . = (gjN (s))2 λj and |z N (s)|2 = (gjN (s))2 .

10) is null. because Z h 2 i N  1 1 z (s)zyN (s). z N (s)zyN (s). ω1 = 0.1-item (4) the second statement in (2. ω1 = 0. In fact. ω1 = z (s). Therefore. Taking into account |y|−→∞  these facts in (2. Setting ω = ω1 ∈ VN in (2.Next one can prove that N X g1N (s) = 0 and (gjN (s))2 (λj − 1) ≥ 0 for all s and N. we will show that (zN (s).1-item (4) one can writes  1 N  Lz N (s). (2. 2 The non-linear term of (2. ω1 . ω1 = (zN (s). ω1 ) = 0. ω1 − z (s). 2 (4π)1/4 R y we have used above Proposition 2.10) 2 By using (2. first. this completes the proof of statement of (2.6)1 . (2. it suffices to prove that g1N (s) = 0 for all s and N. that is. ω1 = z N (s) dy. j=1 Thus. (4π)1/4 2 2 From this. ω1 = 0 for all N and s. Thus. the inequality (2. ω1 + Lz N (s). it yields zsN (s). as ωj ∈ H1 (K) then z N and zyN belong to L1 (R) and consequently  lim z N (y. 1 ω1 (y) = exp{−y 2 /4}.6)2 and hypothesis   on z0 we get z N (s). ω1 = z N (0). it yields   1 N   zsN (s). Therefore. Thus.9) is obvious.1-item (4).9) Since (2.7) is reduced to 1 d . note that X N  g1N (s) = gjN (s)ωj . t) = 0.10).8) is true. ω1 = 0.9) j=2 From Proposition 2. by using (2. ω1 ) . ω1 + z N (s)zyN (s).4) and Proposition 2.

.

N .

.

2 1 .

.

N .

.

2 .

.

N .

.

2  1 √ .

.

 z (s) + zy (s) + zy (s) − 3C1 .

z N (s).

≤ 0. 2010 No.12) 4 3C1 EJQTDE. (2. we will prove that 1 |z N (s)| < √ for all s ≥ 0. 18. 9 . p.11) 2 ds 4 4 Next. (2.

4 3C1 4 3C1 Integrating (2. Then there exists s∗ such that 1 1 |z N (s)| < √ for all 0 ≤ s < s∗ and |z N (s∗ )| = √ .11) from 0 to s∗ . suppose it is not true. it yields Z s∗ Z s∗ 1 √  1 N ∗ 2 1 1 |z (s )| + |zyN (s)|2 ds + |zyN (s)|2 − 3C1 |z N (s)| ds ≤ |z0 |2 . . 2 4 0 0 4 2 From hypothesis on z0 we have √ |z N (s∗ )| < 1/4 3C1 .In fact.

.

√ This contradicts .

z N (s∗ ).

14) 2 EJQTDE. 2 i=1 i=1 i=1 i=1 From this and definition of PN one can write   1   PN zsN + PN Lz N − PN z N + PN z N (s)zyN (s) = 0. ωi )ωi + N (Lz . ωi )ωi − (z . Therefore.4) and (2. ωi )ωi + (z N (s)zyN (s). then 1  zsN = −Lz N + z N − PN z N (s)zyN (s) . p. 18. (2. 10 .12) it is true. In this estimate we will use the projection operator N X PN : L2 (K) −→ VN defined by v 7−→ PN (v) = (v. ωi )ωi . (2. 2010 No.13) 2 0 4 3C1 Estimate 2.6)2 .11) from 0 to s and by using (2. it yields Z s N 1 2 1 |z (s)| + |zyN (τ )|2 dτ ≤ |z0 |2 ≤ √ . Lz N ∈ VN . integrating (2. = 1/4 3C1 . (2. i=1 Thus. zsN . Thus. 2 As PN VN ⊂ VN and z N .6)1 we have N X N X N N 1X N X (zsN . from (2. ωi )ωi = 0.

S. H1 (K) sense. analyzing each term on the right-hand side of (2.14) is verified in the L∞ 0.14) we prove this statement as one can see: .   ′  The identity (2. In fact.

 .

.

.

.

N .

N |Lz (s)|[H1 (K)] = ′ sup .

zy (s). vy L2 (K) .

≤ .

zyN (s).

(2. .15) v∈H1 (K) R ||v||1 ≤1 .

 .

.

.

.

N .

|z N (s)|[H1 (K)]′ = sup .

v L2 (K) . z (s).

≤ .

z N (s).

. (2.16) v∈H1 (K) R ||v||1 ≤1 .

.

.

N .

.

N .

.

PN z N (s)zyN (s) .

1 ≤ C1 sup .

z (s).

.

zy (s).

k(PN v)k 1 [H (K)]′ H (K) v∈H1 (K) ||v||1 ≤1 .

.

.

.

≤ .

z N (s).

.

zyN (s).

17) are similar.16) and (2. we will just make the last one. (2. .15). In fact.17) As the proof of the three identities (2. . (2.

.

.

 .

.

PN z N (s)zyN (s) .

1 .

.

[H (K)]′ = sup .

PN z N (s)zyN (s) . v [H 1 (K)]′ ×H 1 (K) .

v∈H1 (K) R ||v||≤1 .

  .

.

.

= sup .

PN z N (s)zyN (s) . v L2 (K) .

v∈H1 (K) R ||v||≤1 .

 .

.

N .

= sup .

z (s)zyN (s). PN v L2 (K) .

v∈H1 (K) R ||v||≤1 .

N .

.

N .

≤ C1 sup .

z (s).

.

zy (s).

2010 No. i=1 λi EJQTDE. ωiy ) i=1 i=1 N X N X 2 = (v. p. k(PN v)k 1 H (K) . ωi ) (Lωi . √ = vy . 18. √ i=1 λi i=1 λi X∞  2 ωiy ≤ vy . ωi ) = (v. 11 . v∈H1 (K) ||v||≤1 On the other hand. XN 2 XN k(PN v)k2H1 (K) = (v. ωi )2 λi i=1 i=1 XN  2 N X  2 Lωi ωiy = v. ωi )ωi = (v. √ = kvk2H1 (K) ≤ 1. ωi )2 (ωiy .

Integrating this inequality from 0 to S and again using Estimate (2.13) and (2. we get  2 1 N  N |zsN (s)|2[H1 (K)]′ ≤ N |z (s)| + 1 + C1 |z (s)| |zy (s)| 2 " √ ! #2 2 ≤ + 1 + C1 |z N (s)| |zyN (s)| 2 " √ ! #2 2 C1 N ≤ + 1 + p√ |zy (s)| .18) we can extract subsequences of (z N ).17). from (2. 2 2 3C1 2 3C1 The limit in the approximate problem (2.15)-(2.18) 0 where !2 √ 2 C1 1 C= + 1 + p√ √ . more pre- cisely. 2 2 3C1 where we have used in the two last step the Poincaré inequality and Estimate (2. which one will denote by (z N ).13).Inserting this inequality in the precedent one we get (2.6): By Estimates 1 and 2. we obtain Z S |zsN (s)|2[H1 (K)]′ ds ≤ C.17) in (2. (2. By using (2.14). and a function z : R × (0. S) → R satisfying .13).

 .

L2 (K) . S. zN ⇀ z weak star in L∞ 0. .

.

 .

(2. H1 (K) . zN ⇀ z weak in L2 0. S.19) .

  ′  .

.

S. 0 0 EJQTDE. z N φ ds (2. In fact. from (2. From these convergence we are able to pass to the limits in the linear terms of (2.20) On the other hand. z N − z φ ds + zyN . p. see Aubin [1]. H1 (K) . s) = v(x)θ(s) with v ∈ H1 (K) and θ ∈ D(0. S. Browder [4]. The nonlinear term needs careful analysis. 2010 No. L2 (K)). for all φ(x. φ(s) ds = zyN . (2. 12 . zsN ⇀ zs weak in L2 0. zφ ds. 18. which one will denote by (z N ).6). S) we have Z S Z S N N   z (s)zy (s). Lions [15] or Lions [14]. such that zN → z strongly in L2 (0.21) 0 0 Z S Z S     = zyN . 3 and Aubin’s compactness result.19)1. we can extract a subsequences of (z N ).

Next. In fact. the first one can be upper bounded as follows .21) converge. we will show that the last two integrals on the right-hand side of (2.

Z S  N   .

.

.

N .

zy . z − z φ ds.

≤ 0 R Z S .

N.

.

.

.

zy .

|φ|L∞ (R) .

z N − z .

ds ≤ 0 .

.

.

.

C1 |φ|L∞ (0.S.H1 (K)) .

zyN .

L2 (0.S.L2 (K)) .

z N − z .

suppose z and zb are two solutions of (1.L2 (0. 0 0 Taking these two limits in (2. S > 0. 18. Thus.22)1 with ϕ we obtain 1 d 1 |ϕ(s)|2 + |ϕy (s)|2 = |ϕ(s)|2 − (z(s)ϕy (s).21) we get Z S Z S N  z (s)zyN (s). ϕ(s)) − (b zy (s)ϕ(s). we have Z S Z S  zyN (s).7): The global weak solutions of the initial value problem (1. S.13) and (2. zi[H1 (K)]′ ×H1 (K) makes sense.19)1 and (2. z N (s) − z(s) φ(s) ds −→ 0 as N −→ ∞.20) we have Z S    zyN (s).22) 2 Taking the duality paring on the both sides of (2. φ(s) ds −→ (z(s)zy (s). (2. (2. 2010 No. L2 (K)  and because φz ∈ L2 0. (2. In fact.23) 2 ds 2 EJQTDE. S]. from (2.S. φ(s)) ds as N −→ ∞ 0 0 Uniqueness of solutions of (1. z(s)φ(s) ds −→ (zy (s). z(s)φ(s)) ds as N −→ ∞.7) is unique for all s ∈ [0.19)3 the duality hzs . in particular. then ϕ satisfies ϕ ϕs + Lϕ − = − (zzy − zb zby ) . From this.19)2 we have. L2 (K) . ϕ(0) = 0. that  zyN ⇀ zy weak in L2 0.7) and let ϕ = z − zb. 13 . 0 The second integral also converges because from (2. p.L2(K)) . S. ϕ(s)) . Therefore.

as s = ln(t + 1).3) we obtain |z(s)|2 = (t + 1)1/2 |u(t)|2 and |z0 |2 = |u0 |2 . then exp[−s/4] = (t + 1)−1/4 . H1 (K) . Moreover. p. as s → ∞. 2 ds 2 2  From (2.11). S] 1 2 Asymptotic behavior: The asymptotic behavior. from (2.5). ds 4 As a consequence from this inequality we get the inequality (2.7) is established as a consequence of inequality (2.From Proposition 2.12) and Banach-Steinhauss theorem we get that the limit function z defined by (2. S.1-item (5) we obtain d 1 |z(s)|2 + |z(s)|2 ≤ 0.5) directly Remark 2. 2010 No.19) satisfies the inequality 1 d 1 |z(s)|2 + |zy (s)|2 ≤ 0. 18. 2 ds 4 From Proposition 2. from (1. applying the Gronwall inequality one gets ϕ(s) = 0 in [0. The inequality (2. ϕ(s))|R ≤ |z(s)|L∞ (R) |ϕy (s)| |ϕ(s)| + |ϕ(s)|L∞ (R) |b zy (s)| |ϕ(s)| ≤ √   C1 3 |ϕy (s)| |zy (s)| + |b zy (s)| |ϕ(s)| ≤ 1   |ϕy (s)|2 + 3C12 |zy (s)|2 + |bzy (s)|2 |ϕ(s)|2 . of E(s) = 2 |z(s)| given by the unique solution of the Cauchy problem (1.23) yields    1 d 1 1 2 2 2 2 |ϕ(s)| + |ϕy (s)| ≤ 2 + 3C1 |zy (s)| + |b zy (s)| |ϕ (s)|2 . In fact.11). 14 . 2 Substituting this inequality in (2.1. ϕ(s)) + (b zy (s)ϕ(s).2) is a consequence of (2.19)2 one has that z and zb belong to L2 0.5) we have 1 |u(t)|2 = |z(s)|2 ≤ |u0 |2 (t + 1)−3/4 t+1 EJQTDE. (2. Therefore. from (2. Therefore.2-item (b) and (c) one obtains |(z(s)ϕy (s). In fact.

∞.1 and The- orem 3.1) 0 R 0 R 0 R  for all ϕ ∈ L2 0.1 are also equivalent to Definition 3. The existence of solution of (1. Definition 3.3 Strong Solution Setting the initial data u0 ∈ H1 (K) we are able to show that the Cauchy problem (1. ∞.2. ∞. p. Definition 3. T ) for an arbitrary T > 0. 0) = u0 (x) for all x ∈ R. Z T Z Z TZ Z TZ ut ϕdxdt + uux ϕdxdt + ux ϕx dxdt = 0. Moreover. 15 .1) has a unique real valued strong solution u = u(x. H (R)). ut ∈ Lloc (0.1. 0) = z0 (y) for all y ∈ R.2 and Theorem 3. Moreover. u satisfies the initial condition u(x. EJQTDE. zs ∈ Lloc (0. S. t) defined in R × (0. (3. T . T ) for all T > 0. the Definition 3. H (K)) for S > 0.7) is a real valued function z = z(y. such that u ∈ L∞ 2 2 1 loc (0. 18. A global strong solution of the initial-boundary value problem (1.1.1) is a real valued function u = u(x.1) is defined as follows. the strong solution of (1. Moreover. S) for arbitrary S > 0. H (R)). As the Cauchy problems (1.7) are equivalent. Precisely. 2 0 R 0 R  for all ϕ ∈ L2 0. such that z ∈ L∞ 2 2 1 loc (0. H1 (K) .1. ∞. Theorem 3.2. t) defined in R × (0. 2010 No. then there exists a unique global solution u of (1. energy E(t) = 12 |u(t)|2 associated with this solution satisfies E(t) ≤ E(0)(t + 1)−3/4 . H (K)). Z SZ Z SZ zs ϕKdyds + LzϕKdyds+ (3.1) in the precedent sense is guaranteed by the fol- lowing theorem.2) 0 R 0 R Z S Z Z S Z 1 − zϕKdyds + zzy ϕKdyds = 0.1) in the sense of Definition 3. Suppose u0 ∈ H1 (K).1) and (1. z satisfies the initial condition z(y. H 1 (R) . A global strong solution for the initial value problem (1. s) defined in R × (0.

the energy E(s) = 12 |z(s)|2 satisfies E(s) ≤ E(0) exp [−s/4] . then there exists a unique solution z √ of (1.7) will be also shown by means of Faedo- Galerkin method and by using the special basis defined as solutions of spectral problem  (2. (ω1 . v) = 0 . Since Theorems 3.2. In fact. Theorem 3.2. Setting ω = Lz N (s) ∈ VN in (2.2.2 are equivalent it is suffices to prove Theorem 3. The existence of solutions of the system (1. Under these conditions one defines VN as in Section 2.1 and 3.2 . We need to establish two estimates. Now we state the following theorem. Moreover. Suppose z0 ∈ H1 (K) ∩ {ω1 }⊥ . Estimate 3. it yields 1 d N |z (s)|2 + |Lz N (s)|2 ≤ 2 ds y 1 N 1 .7) in the sense of Definition 2. Proof of Theorem 3. provided |z0y | < 1/4 6C1 holds.4) and the first eigenfunction ω1 of L such that {ω1 }⊥ = v ∈ L2 (K).6)1 .

Z .

.

.

|z (s)| + |Lz (s)| + .

z N (s)zyN (s)Lz N (s)Kdy .

Proposition 2.2-items (b). In fact. (c) and (d) we can write . we will find the upper bound of the last term on the right-hand side of the above inequality. 2 N 2 8 2 R R Next. by using Hölder inequality. .

Z .

√ .

.

.

.

2 .

.

.

z N (s)zyN (s)Lz N (s)Kdy .

≤ C1 6 .

zyN (s).

.

Lz N (s).

18. R R From this we have 1 d N 1 1 N  |zy (s)|2 + |Lz N (s)|2 + |Lz (s)|2 − |z N (s)|2 + 2 ds 8 8 1 √  (3. 4 Use a similar argument as in Estimate 1 we are able to prove that 1 N  |Lz (s)|2 − |z N (s)|2 ≥ 0 for all N ∈ N.3) |Lz N (s)|2 − 6C1 |zyN (s)| ≤ 0. p. 2010 No.4) 8 EJQTDE. (3. . 16 .

8 8 8 j=2 From (2.In fact. zsN (s) + z (s).2 we have   2 .5). 2 Next.6) in (3. it yields  1 N   |zsN (s)|2 = − Lz N (s). see Estimate 1.7) 4 0 4 6C1 Estimate 4. (3.4) is true.4). Setting ω = zsN (s) ∈ VN in (2.9) we have N X  (gjN (s))2 λ2j − 1 ≥ 0 and g1N (s) = 0 for all s and N. (3. In fact.5) 2 ds 8 4 Next. note that 1 N  1 N 2  1X  2 N 2 2 |Lz (s)| − |z (s)| = (g1N (s)) λ1 − 1 + (gjN (s))2 λ2j − 1 . we will estimate the three inner product on the right-hand side of the above identity.12) we will prove that √ |zyN (s)| < 1/4 6C1 for all s ≥ 0.6) Therefore.3) is reduced to 1 d N 1 1 √  |zy (s)|2 + |Lz N (s)|2 + |Lz N (s)|2 − 6C1 |zyN (s)| ≤ 0. by usual inequalities and Proposition 2. zsN (s) − z N (s)zyN (s). j=2 From this we obtain (3. the inequality (3. it yields 1 d N 1 |zy (s)|2 + |Lz N (s)|2 ≤ 0. 2 ds 8 Integrating from 0 to s and using the hypothesis on the initial data we obtain Z 1 s 1 |zyN (s)|2 + |Lz N (τ )|2 dτ ≤ |z0y |2 < √ . proceeding as (2. Since (3.6)1 . (3. zsN (s) . by using (3.

.

N .

1 .

N .

.

N .

2 2 N 2 .

.

|zs (s)| ≤ C2 Lz (s) + z (s) + zy (s) .

.

.

2 √ where C2 = max{1. . 3C1 }. From this we get a constant C > 0 independent of N and s such that Z S .

N .

2 .

zs (s).

ds ≤ C. 18. (3. 2010 No. 17 .8) 0 EJQTDE. p.

However. such that .7) and (3.6).8) we can take the limit on the approximate system (2. 1/4 6C1 and of the con- stants of immersions established in Proposition 2. From (3.6) is similar to those of Section 2. the nonlinear term is made as follows.2.7). (3. the analysis of the limit as N −→ ∞ in the linear terms of (2. From (3. In fact. √ where C depends on the constant of Estimate (3. that is.8) and Aubin-Lions theorem one extracts subsequences of (z N ).7). which will be denoted by (z N ).

 .

.

S. . H2 (K) as N −→ ∞. zN ⇀ z weak in L2 0.

 .

. zN → z strong in L2 0. L2 (K) as N −→ ∞. S.

.

 (3.9) .

L2 (K) as N −→ ∞. . zyN → zy strong in L2 0. S.

 .

.

2 one has Z . S. From usual inequalities and Proposition 2. zsN ⇀ zs weak in L2 0. L2 (K) as N −→ ∞.

N .

.

zy (s)z N (s) − zy (s)z(s).

2 Kdy = R R Z .

 N  N  N .

.

zy (s) − zy (s) z (s) + zy (s) z (s) − z(s) .

2 Kdy ≤ R R .

.

2 .

N .

.

.

2 .

z N (s).

∞ .

zy (s) − zy (s).

2 + 2 |zy (s)|2 .

z N (s) − z(s).

2 ≤ L (R) h.

.

2 .

.

2 i C .

zyN (s) − zy (s).

+ .

z N (s) − z(s).

it yields Z S . Taking (3. .9) in this inequality.

N .

.

zy (s)z N (s) − zy (s)z(s).

2 ds ≤ 0 Z h.

(3.10) S .

.

.

i C .

z N (s) − zy (s).

2 + .

z N (s) − z(s).

2 ds −→ 0 as N −→ ∞. y 0 Therefore. p. S. 2010 No.2 is completed by using a similar argument as in Section 2. 18 . the uniqueness of solutions and the exponential decay rate of the energy are established in a similar way as in Section 2 EJQTDE. Finally. 18. the proof of Theorem 3. one has  zyN z N −→ zy z strong in L2 0. L2 (K) as N −→ ∞ Thus.

We will achieve this goal by means of the following regularity result Theorem 4. Let z = z(y. ω − 1/2 v N (s). Setting ω = v N (s) ∈ VN in (4. Proof: We will show that z is the limit of a Cauchy sequence.4 Regularity of Strong Solutions Our goal here is to prove a result of regularity for strong solutions established in Section 3. ω =     (4. from (2. suppose M. H1 (K) .1) PM z M (s)zyM (s) − PN z N (s)zyN (s) . then z ∈ C 0 [0. z M are two solutions of (1. T ].6). PM are projection operators defined in L2 (K) with values in V N . 2 Therefore.7). ω . which is guaran-  teed by Theorem 3. v N = z N − z M satisfies 1   vsN (s) + Lv N (s) − v N (s) = PM z M (s)zyM (s) − PN z N (s)zyN (s) in L2 (K). Thus. for all ω ∈ V N ⊂ L2 (K). N ∈ N fixed with N > M and z N .1.2. ω . ω + Lv N (s). In fact. where PN . V M respectively.1) we get 1 d . one has that    vsN (s). Estimate 5. s) be a strong solution of problem (1.7).

.

N .

.

2 .

.

N .

.

2 1 .

.

N .

.

2 v (s) + vy (s) − v (s) ≤ .

2 ds 2 .

N N  .

PN z (s)zy (s) − z(s)zy (s) + PN (z(s)zy (s)) −  .

.

.

.

PM (z(s)zy (s)) + PM z(s)zy (s) − z M (s)zyM (s) .

.

v N (s).

≤ .

N .

.

z (s)zyN (s) − z(s)zy (s).

2 + |PN (z(s)zy (s)) − PM (z(s)zy (s))|2 + .

.

.

.

.

z(s)zy (s) − z M (s)zyM (s).

2 + 3 .

v N (s).

2 . 4 Integrating form 0 to S one has Z S Z S 1 .

.

N .

.

2 .

N .

2 .

.

1 .

.

N .

.

2 .

N .

.

z (s)z N (s) − z(s)zy (s).

2) Z S 0 Z S .2 ds+ v (s) + vy (s) ds ≤ v0 + y 2 0 Z 2 0 S |PN (z(s)zy (s)) − PM (z(s)zy (s))|2 ds+ (4.

.

.

N .

2 .

z(s)zy (s) − z M (s)zyM (s).

2 ds + 5 .

v (s).

2010 No. 19 . ds. 18. 0 4 0 EJQTDE. p.

S.  Therefore. Therefore. Hence we have that (4. S] as N −→ ∞. (4.The task now is to show that Z S |PN (z(s)zy (s)) − PM (z(s)zy (s))|2 ds −→ 0 as N −→ ∞. |PN (z(s)zy (s)) − z(s)zy (s)| ≤ 2 |z(s)zy (s)| and |z(s)zy (s)| ∈ L2 (0. L2 (K) . as zzy ∈ L2 0. (PN (zzy ))N ∈N is a Cauchy sequence in L2 0. in [0. L2 (K) as N −→ ∞. S. then z(s)zy (s) ∈ L2 (K) a. PN (z(s)zy (s)) −→ z(s)zy (s) in a.2) and Granwall inequality one gets Z S 1 . in [0. S] as N −→ ∞.3) is true On the other hand. Thus. S. it yields Z S |PN (z(s)zy (s)) − z(s)zy (s)|2 ds −→ 0 as N −→ ∞. That is. e. from (4. e. e. in [0. L2 (K) . S]. Moreover. applying Lebesgue’s dominated convergence theorem. 0 In other words  PN (zzy ) −→ zzy in L2 0.3) 0  In fact. |PN (z(s)zy (s)) − z(s)zy (s)| −→ 0 in a. S).

.

N .

.

2 .

N .

2 .

vy (s).

ds ≤ v (s) + 2 0 h 1 .

.

2 Z S .

.

.

v0N .

+ .

z N (s)zyN (s) − z(s)zy (s).

2 ds+ 2 0 Z S |PN (z(s)zy (s)) − PM (z(s)zy (s))|2 ds+ Z S 0 .

.

i .

z(s)zy (s) − z M (s)zyM (s).

L2 (K) . p.10) and (4. S. (3. EJQTDE. 2010 No. To obtain the desired regularity one needs one more estimate as follows. 0  From this. hypothesis on the initial data. 18. 20 .3) one gets that z N N ∈N is a  Cauchy sequence in C 0 0.2 ds exp {(5/4)S} .

Setting ω = Lv N (s) ∈ VN in (4.1) we get 1 d .Estimate 6.

.

N .

.

2 .

.

N .

.

2 1 .

.

N .

.

2 vy (s) + Lv (s) − vy (s) ≤ .

2 ds  2 .

N N .

PN z (s)zy (s) − z(s)zy (s) + PN (z(s)zy (s)) −  .

.

.

.

PM (z(s)zy (s)) + PM z(s)zy (s) − z M (s)zyM (s) .

.

Lv N (s).

≤ .

N .

.

z (s)zyN (s) − z(s)zy (s).

2 + |PN (z(s)zy (s)) − PM (z(s)zy (s))|2 + .

.

.

.

.

z(s)zy (s) − z M (s)zyM (s).

2 + 3 .

Lv N (s).

2 . one gets Z S 1 . 4 Integrating from 0 to S and using Granwall inequality.

.

N .

.

2 .

N .

2 .

Lv (s).

ds ≤ vy (s) + y 2 0 h 1 .

.

2 Z S.

.

N.

.

v0y + .

z N (s)zyN (s) − z(s)zy (s).

2 ds+ 2 0 Z S |PN (z(s)zy (s)) − PM (z(s)zy (s))|2 ds+ Z S 0 .

.

i .

z(s)zy (s) − z M (s)z M (s).

21 . 2010 No. Dunod. 43 (1915). 163–170. we have the desired regularity. 256 (1963). 18.R. EJQTDE. Paris (1999). H. Mon. p. 5042–5044. (3.3) one gets that z N N ∈N is  a Cauchy sequence in C 0 0. Sci. Acad. S.10) and (4. H. P. And therefore the proof of Theorem 4. y 0  From this. pp. J. Water Rev... hypothesis on the initial data.1 is ended Acknowledgment We want to thank the anonymous referee for a careful reading and helpful suggestions which led to an improvement of the original manuscript. H1 (K) . Thus.2 ds exp {(5/4)S} .. Some recent researches on the motion of fluids. Analyse Fonctionelle (Théorie et Applications). References [1] Aubin. Un theorème de compacité. [2] Bateman. pp.. [3] Brezis. Paris. C.

Vol. Theory of ordinary differential equations. Hyperbolic systems of conservation laws and the mathematical theory of shock waves. TMA.. II. V. Problèmes Non-Linéaires. (1983).. N. EDP Sciences Ed. [10] Hopf. 1 (1948). R. Comm. vol. Sci. 2010 No. Supersonic flow and shock waves Interscience. Proc. pp. Vol. 11. 10 (1987). 1–47. Adv. J. Dunod. (1972). Paris (1969). R. E. 2–12. Canada. [11] Escobedo. L. 2003. Press de l’Université de Montréal. Karman. pp. R.. Nonlinear Analysis. 1962. E. Acad. K. E. No. Variational problems related to self-similar solutions of the heat equation. [13] Lax. 1948. J. 43 (1940). 171–199. D. & Hilbert. O.. L. pp. [5] Burgers. 3 (1950). [7] Coddington.Regional conference series in applied mathematics. SIAM . The partial differential equation ut + uux = ǫuxx .. Pure Appl. No. J. F. 18. [9] Courant. (1966). 1103– 1133.. & Levinson. M. M. pp. 1955.. A mathematical model illustrating the theory of turbulence). pp. Math. J. & Kavian. Diff. Paris. pp. App.. V. N. Oeu- vres Choisies de Jacques Louis Lions.. Problèmes aux limites dans les équations aux derivées partielles. & Friedrichs. [15] Lions.Y. Methods of mathematical physics. Mac- Graw Hill. Interscience. New York. 201–230. R. J. Equations. [6] Burgers. 11. Quelques methodes de resolution des problèmes aux limites non lin- eaires. Mech. Weighted Sobolev spaces with applications to singular nonlinear bound- ary value problems. Mises and T. EJQTDE. Montreal. [8] Courant. M. [4] Browder. Amsterdam. 49.. Application of a model system to illustrate some points of the statistical theory of free turbulence. p. [14] Lions. O. P. New York.. Ed. J. I. C.. 105–123. 22 .. [12] Kurtz..

France.. 55. Existence of solitary waves in higher dimensions. A. J. 23 . (Received December 29. 149–162. Phys.. [17] Stoker. 2009) EJQTDE. Interscience. 2010 No.. pp. 1945. Hermann. Mèthodes mathematiques pour les sciences physiques. Water waves. Paris. L. (1977). New York.[16] Schwartz. p. J. Communs Math. W.. 1957 [18] Strauss. 18.