Next Article in Journal
Hyperbolic B-Spline Function-Based Differential Quadrature Method for the Approximation of 3D Wave Equations
Next Article in Special Issue
Solving Biharmonic Equations with Tri-Cubic C1 Splines on Unstructured Hex Meshes
Previous Article in Journal
Applications of the q-Derivative Operator to New Families of Bi-Univalent Functions Related to the Legendre Polynomials
Previous Article in Special Issue
On the Method of Differential Invariants for Solving Higher Order Ordinary Differential Equations
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Regularity and Decay of Global Solutions for the Generalized Benney-Lin Equation Posed on Bounded Intervals and on a Half-Line

by
Nikolai A. Larkin
Departamento de Matemática, Universidade Estadual, de Maringá, Av. Colombo 5790: Agência UEM, Maringá 87020-900, PR, Brazil
Axioms 2022, 11(11), 596; https://doi.org/10.3390/axioms11110596
Submission received: 4 September 2022 / Revised: 21 September 2022 / Accepted: 21 September 2022 / Published: 28 October 2022
(This article belongs to the Special Issue Higher Order Differential Equations)

Abstract

:
Initial-boundary value problems for the generalized Benney-Lin equation posed on bounded intervals and on the right half-line were considered. The existence and uniqueness of global regular solutions on arbitrary intervals as well as their exponential decay for small solutions and for a special choice of a bounded interval have been established.

1. Introduction

This work concerns the existence and uniqueness of global solutions, regularity and exponential decay rates of solutions to some initial- boundary value problems for the generalized Benney-Lin equation
u t + η D x 5 u + β D x 4 u + α D x 3 u + γ D x 2 u + u k u x = 0 ,
where k 1 is a natural number.
This equation was deduced in [1] in connection with applications in fluid mechanics and later was exploited in [2] during research in the theory of liquid films. For various values of the coefficients α , η , β , γ and k = 1 , it presents well-known equations of mathematical physics such as the Korteweg-de Vries equation when η = β = γ = 0 and α = 1 ; the Kawahara equation when γ = β = α = 0 and η = 1 . In the case η = α = 0 and β , γ are positive constants, (1) is the Kuramoto-Sivashinsky equation. In [3], Kuramoto studied the turbulent phase waves and Sivashinsky in [4] obtained an asymptotic equation which simulated the evolution of a disturbed plane flame front. See also [5]. Mathematical results on initial and initial-boundary value problems for various variants of (1) are presented in [6,7,8,9,10,11,12,13,14,15], see references there for more information. In [6,9,12,15], Kuramoto-Sivashinsky type equations have been considered which included u x x x (KdV) term. Initial-boundary value problems for equations of higher orders than (1) have been considered in [16,17,18]. What concerns (1) with k > 1 , η = β = γ = 0 ; α = 1 , called generalized Korteweg-de Vries equations, the Cauchy problem has been studied in [19,20], where it has been proved that for k = 4 , called the critical case, the initial problem is well-posed for small initial data, whereas for arbitrary initial data, solutions may blow-up in a finite time. The generalized KdV equation was intensively studied in order to understand the interaction between the dispersive term and nonlinearity in the context of the theory of nonlinear dispersive evolution equations, see [21]. In [22], an initial-boundary value problem for the generalized KdV equation with an internal damping posed on a bounded interval was studied in the critical case. In [23], an initial-boundary value problem for the generalized KdV equation posed on a half-line was considered, where exponential decay of regular solutions for small initial data has been established. In [24], the Zakharov-Kuznetsov equation with super critical nonlinear term has been studied.
Our goal here are some initial-boundary value problems for the generalized Benney-Lin (Bl) equation
u t D x 5 u + D x 4 u + D x 3 u + D x 2 u + u k u x = 0 ,
where k 1 is a natural number.
First essential problem that arises while one studies stability of solutions for (2), is a destabilizing effect of D x 2 u that may be damped by a dissipative term D x 4 u provided an interval ( 0 , L ) , where (2) is posed, has some specific properties.
The existence and uniqueness of global regular solutions to (2) posed on any bounded interval (0,L) and on the right half-line for k 4 and any positive time interval ( 0 , T ) without smallness conditions for the initial data was established in Theorems 1 and 2. An initial-boundary value problem for (2) posed on a special interval ( 0 , L ) was formulated. The existence of a global regular solution, uniqueness and exponential decay rate of small solutions with k 8 have been established in Theorem 3.

2. Notations and Auxiliary Facts

Let L be a positive number and x ( 0 , L ) . We use the standard notations of Sobolev spaces W k , p , L p and H k for functions and the following notations for the norms [25,26]:
R + = { t R 1 ; t > 0 } , f 2 = ( f , f ) = 0 L | f | 2 d x , f L p p = 0 L | f | p d x ,
f W k , p = 0 α k D x α f L p , D x α = d α d x α , D x = D x 1 , D x 0 u = u .
We will use also the standad notations:
D x f = f x , D x 2 f = f x x , t f = f t .
When p = 2 , W k , p = H k is a Hilbert space with the scalar product
( ( u , v ) ) H k = | j | k ( D j u , D j v ) , u = u L ( 0 , L ) = e s s sup ( 0 , L ) | u ( x ) | .
We use the notation H 0 k ( 0 , L ) to represent the closure of C 0 ( 0 , L ) , the set of all C functions with compact support in ( 0 , L ) , with respect to the norm of H k ( 0 , L ) .
Lemma 1. 
(Steklov’s Inequality [27].) Let v H 0 1 ( 0 , L ) . Then
π 2 L 2 v 2 v x 2 .
Lemma 2. 
(See [18], Lemma 2.2.) Let u be either u H 0 2 ( 0 , L ) or u H 2 ( R + ) , u ( 0 ) = D x ( 0 ) = 0 . Then the following inequality holds:
u 2 D 2 u 1 4 u 3 4 .
Lemma 3. 
(See [28], p. 125.) Suppose u and D m u , m N , belong to L 2 ( 0 , L ) . Then for the derivatives D i u , 0 i < m , the following inequality holds:
D i u A 1 D m u i m u 1 i m + A 2 u ,
where A 1 , A 2 are constants depending only on L, m, i.
Lemma 4. 
(Differential form of the Gronwall Inequality.) Let I = [ t 0 , t 1 ] . Suppose that functions a , b : I R are integrable and a function a ( t ) may be of any sign. Let u : I R be a differentiable function satisfying
u t ( t ) a ( t ) u ( t ) + b ( t ) , for t I a n d u ( t 0 ) = u 0 ,
then
u ( t ) u 0 e t 0 t a ( τ ) d τ + t 0 t e t 0 s a ( r ) d r b ( s ) d s .
Lemma 5.
Let f ( t ) be a continuous positive function such that f ( t ) is a measurable, integrable function and
f ( t ) + ( α k f n ( t ) ) f ( t ) 0 , t > 0 , n N ,
α k f n ( 0 ) > 0 , k > 0 .
Then
f ( t ) < f ( 0 )
for all t > 0 .
Proof. 
Obviously, f ( 0 ) + ( α k f n ( 0 ) ) f n ( 0 ) 0 . Since f is continuous, there exists T > 0 such that f ( t ) < f ( 0 ) for every t [ 0 , T ) . Suppose that f ( 0 ) = f ( T ) . Integrating (3), we find
f ( T ) + 0 T ( α k f n ( t ) ) f ( t ) d t f ( 0 ) .
Since
0 T ( α k f n ( t ) ) f ( t ) d t > 0 ,
then f ( T ) < f ( 0 ) . This contradicts that f ( T ) = f ( 0 ) . Therefore, f ( t ) < f ( 0 ) for all t > 0 .

3. Generalized Benney-Lin Equation Posed on Bounded Intervals

Problem 1.
Define an interval
D = x ( 0 , L ) , L > 0 ; Q t = ( 0 , t ) × D .
Lemma 6.
Let f H 2 ( D ) H 0 1 ( D ) . Then
a f 2 f x 2 , a 2 f 2 f x x 2 , a f x 2 f x x 2 ,
where a = π 2 L 2 .
Proof. 
Making use of Steklov’s inequalities, we get
f x 2 π 2 L 2 f 2 = a f 2 .
On the other hand,
a f 2 f x 2 = 0 L f f x x d x f x x f .
This implies
a f f x x and a 2 f 2 f x x 2 .
Consequently, a f x 2 f x x 2 .
Proof of Lemma 6 is complete. □
In Q t consider the following initial-boundary value problem:
u t D x 5 u + D x 4 u + D x 3 u + D x 2 u + u k u x = 0 ,
D x i u ( 0 ) = D x i u ( L ) = D x 2 u ( L ) = 0 , i = 0 , 1 ;
u ( x , 0 ) = u 0 ( x ) , x ( 0 , L ) .
Theorem 1.
Let T , L be arbitrary positive numbers and a natural k 4 . Given u 0 H 5 ( D ) H 0 2 ( D ) , D x 2 u 0 ( L ) = 0 . Then the problem (6)–(8) has a unique regular solution
u L ( ( 0 , T ) ; H 0 2 ( D ) ) L 2 ( ( 0 , T ) ; H 7 ( D ) ) ;
u t L ( ( 0 , T ) ; L 2 ( D ) ) L 2 ( ( 0 , T ) ; H 0 2 ( D ) ) .
Proof. 
Define the space W = { f H 5 ( D ) H 0 2 ( D ) , D x 2 f ( L ) = 0 } and let { w i ( x ) , i N } be a countable dense set in W. We can construct approximate solutions to (6)–(8) in the form
u N ( x , t ) = i = 1 N g i ( t ) w i ( x ) .
Unknown functions g i ( t ) satisfy the following initial problems:
d d t ( u N , w i ) ( t ) + ( D x 4 u N , w i ) ( t ) + ( D x 2 u N , w i ) ( t )
( D x 5 u N , w i ) ( t ) + ( D x 3 u N , w i ) ( t ) + ( ( u N ) k D x u N , w i ) ( t ) = 0 ,
g i ( 0 ) = g i 0 , i = 1 , 2 , . . . .
By Caratheodory’s existence theorem, see [29], there exist solutions of (9)–(10), at least locally in t, hence for all finite N, we can construct an approximate solution u N ( x , t ) of (6)–(8). In [17,18], the existence of local regular solutions to problems similar to (6)–(8) has been proved. Taking this into account, all the estimates will be proved on smooth solutions of (6)–(8). Naturally, the same estimates are true also for approximate solutions u N .
Estimate I. Multiply (6) by 2 u to obtain
d d t u 2 ( t ) + 2 D x 2 u 2 ( t ) + 2 ( D x 2 u , u ) ( t ) + | D x 2 ( 0 , t ) | 2 = 0 .
Estimating third term in (11) by the Cauchy inequality, we get
d d t u 2 ( t ) + D x 2 u 2 ( t ) + | D x 2 u ( 0 , t ) | 2 u 2 ( t ) .
Dropping second and third terms in (12) and applying Lemma 4, we find
u 2 ( t ) e T u 0 2 , t ( 0 , T ) .
Returning to (12), we obtain
0 T D x 2 u 2 ( t ) + | D x 2 u ( 0 , t ) | 2 d t ( 1 + T e T ) u 0 2 .
Here and henceforth, T is an arbitrary positive number.
Estimate II.
Differentiate (6) with respect to t, then multiply the result by 2 u t to get
d d t u t 2 ( t ) + D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2
u t 2 ( t ) + 2 ( u k u t , u x t ) ( t ) .
The case k < 4 .
Making use of Lemmas 2 and 6 for an arbitrary ϵ > 0 , we estimate
I = 2 ( u k u t , u x t ) ( t ) 2 sup D | u ( x , t ) | k u t ( t ) u x t ( t )
2 2 1 / 2 D x 2 u 1 / 4 ( t ) u 3 / 4 ( t ) k u t ( t ) 1 a 1 / 2 u x x t ( t )
ϵ u x x t 2 ( t ) + 2 k a ϵ D x 2 u k / 2 ( t ) u 3 k / 2 ( t ) u t 2 ( t )
ϵ u x x t 2 ( t ) + 2 k a ϵ k D x 2 u 2 ( t ) 4 + 4 k 4 u 6 k / ( 4 k ) ( t ) u t 2 ( t ) .
Taking ϵ = 1 2 , and substituting I into (15), we obtain
d d t u t 2 ( t ) + 1 2 D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2
( 3 2 + 2 k 1 a [ k D x 2 u 2 ( t ) 4
+ 4 k 4 ( e T u 0 2 ) 3 k / ( 4 k ) ] ) u t 2 ( t ) .
By (13), (14), D x 2 u 2 ( t ) L 1 ( O , T ) and u ( t ) L ( 0 , T ) , whence, dropping second and third terms in (16) and making use of Lemma 4, we find that
u t 2 ( t ) C 1 u t 2 ( 0 ) , t ( 0 , T ) ,
where
C 1 = exp { 0 T 3 2 + 2 k 1 a k D x 2 u 2 ( t ) 4 + 4 k 4 ( e T u 0 2 ) 3 k / ( 4 k ) d t } .
Returning to (16), we obtain
0 T D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2 d t C ( T , u 0 ) u t 2 ( 0 ) ,
where u t ( 0 ) C ( u 0 W ) can be estimated directly from (6) on t = 0 .
The case k = 4 .
Consider (15) for k = 4 :
d d t u t 2 ( t ) + D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2
u t 2 ( t ) + 2 ( u 4 u t , u x t ) ( t ) .
Acting as in the case k < 4 , we estimate
I 2 = 2 ( u 4 u t , u x t ) ( t ) 2 u 4 ( t ) u t ( t ) u x t ( t )
2 3 a 1 / 2 D x 2 u ( t ) u 3 ( t ) u t ( t ) D x 2 u t ( t )
ϵ D x 2 u t 2 ( t ) + 2 4 a ϵ D x 2 u 2 ( t ) u 6 ( t ) u t 2 ( t ) .
Taking ϵ = 1 2 and substituting I 2 into (18), we get
d d t u t 2 ( t ) + 1 2 D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2
3 2 + 2 5 e 3 T a D x 2 u 2 ( t ) u 0 6 u t 2 ( t ) .
By (14), D x 2 u 2 ( t ) L 1 ( O , T ) , whence, dropping second and third terms in (19) and making use of Lemma 4, we find that
u t 2 ( t ) C 2 u t 2 ( 0 ) , t ( 0 , T ) ,
where
C 2 = 0 T 3 2 + 2 5 e 3 T a D x 2 u 2 ( t ) u 0 6 d t .
Returning to (19), we obtain
0 T D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2 d t C ( T , u 0 ) u t 2 ( 0 ) .
Here u t ( 0 ) C ( u 0 W ) can be estimated directly from (6) on t = 0 .
Proposition 1. 
e s s sup Q T | u ( x , t ) | M < + , t T .
Proof. 
D x 2 u 2 ( t ) D x 2 u 0 2 = 0 t d d s D x 2 u 2 ( s ) d s
0 t 2 D x 2 u ( s ) D x 2 u s ( s ) d s 0 T D x 2 u 2 ( s ) + D x 2 u s 2 ( s ) d s .
Estimates (13), (14) and (21) prove that D x 2 u ( t ) L ( 0 , T ) and Lemma 2 completes the proof of Proposition 1.
Estimates (13), (14), (20) and (21) and Proposition 1 imply that
u L ( ( 0 , T ) ; H 0 2 ( D ) ) , u t L ( ( 0 , T ) ; L 2 ( D ) ) L 2 ( ( 0 , T ) ; H 0 2 ( D ) ) .
It is easy to see that these inclusions do not depend on N, hence, by standard arguments, we can pass to the limit as N in (9), (10) and to prove the existence of weak soluitons to (6)–(8) { u ( x , t ) } satisfying the following identity:
( u t , ϕ ) ( t ) + ( D x 2 u , D x 2 ϕ ) ( t ) + ( D x 2 u , ϕ ) ( t )
+ ( D 2 u x , D x 3 ϕ ) ( t ) ( D x 2 u , ϕ x ) ( t ) + ( u k u x , ϕ ) ( t ) = 0 , t > 0 ,
where ϕ ( x , y ) is an arbitrary function from H 0 3 ( D ) .
We can rewrite (22) in the form of a distribution on ( 0 , L )
D 5 u x D x 4 u D x 3 u = u t + D x 2 u + u k u x f ( x , t ) .
Due to properties of a weak solution u ( x , t ) , f L ( ( 0 , T ) ; L 2 ( D ) ) . This implies that
I = D x 5 u D x 4 u D x 3 u L ( ( 0 , T ) ; L 2 ( D ) ) .
Making use of Lemma 3, we find that for an arbitrary ϵ > 0
( 1 2 ϵ ) D x 5 u ( t ) C ϵ ( u ( t ) + f ( t ) ) .
Choosing 4 ϵ = 2 and taking into account that f , u L ( ( 0 , T ) ; L 2 ( D ) ) , we get
u L ( ( 0 , T ) ; H 5 ( D ) ) .
Taking this into account and that u t L 2 ( ( 0 , T ) ; H 0 2 ( D ) ) , we find from (23)
D x 5 u = u t + D x 4 u + D x 3 u + D x 2 u + u k u x L 2 ( ( 0 , T ) ; H 1 ( D ) ) ,
whence
u L ( ( 0 , T ) ; H 5 ( D ) ) L 2 ( ( 0 , T ) ; H 6 ( D ) ) .
Returning to (24), one can observe that D 5 u L 2 ( ( 0 , T ) ; H 2 ( D ) ) . hence
u L ( ( 0 , T ) ; H 5 ( D ) ) L 2 ( ( 0 , T ) ; H 7 ( D ) ) .
This proves the existence part of Theorem 1.
Remark 1.
Assertions of Theorem 1 are true for arbitrary positive numbers T , L , but estimates of solutions depend on T. It means that one can not pass to the limit as L , T + . Hence, we do not have stability results. On the other hand, we have not any smallness restrictions for u 0 ( x ) .
Lemma 7.
The regular solution of (6)–(8) is unique.
Proof. 
Let u and v be two distinct solutions to (6)–(8). Denoting w = u v , we come to the following problem:
w t D x 5 w + D x 4 w + D x 3 w + D x 2 w + 1 k + 1 D x [ u k + 1 v k + 1 ] = 0 ,
D x i w ( 0 ) = D x i w ( L ) = D x 2 w ( L ) = 0 , i = 0 , 1 ;
w ( x , 0 ) = 0 .
Multiplying (26) by 2 w and taking into account that
D x w 2 = ( D x 2 w , w ) 1 2 D x 2 w 2 + w 2 ,
we find
d d t w 2 ( t ) + | D x 2 w ( 0 , t ) | 2 + 2 D x 2 w 2 ( t ) + 2 ( D x 2 w , w ) ( t )
= 2 k + 1 ( [ u k + 1 v k + 1 ] , D x w ) ( t ) D x w 2 ( t )
+ 4 ( k + 1 ) 2 u k + 1 v k + 1 2 ( t )
1 2 D x 2 w 2 ( t ) + 1 2 w 2 ( t ) + 4 ( k + 1 ) 2 u k + 1 v k + 1 2 ( t ) .
Making use of the functional mean value theorem and Proposition 1, we get
I = u k + 1 v k + 1 2 ( k + 1 ) 2 sup Q T | u | + sup Q T | v | 2 k w 2
( k + 1 ) 2 2 2 k M 2 k w 2 .
Substituting I into (29), we obtain
d d t w 2 ( t ) + 1 2 D x 2 w 2 ( t ) C w 2 ( t ) ,
where the constant C depends on M. Applying Lemma 4, we find that
w ( t ) 0 .
This proves Lemma 7 and consequently Theorem 1. □
Benney-Lin equation posed on R + .
In Q t = ( 0 , t ) × R + , t ( 0 , T ) , consider the following problem:
u t D x 5 u + D x 4 u + D x 3 u + D x 2 u + u k u x = 0 ,
D x i u ( 0 , t ) = 0 , i = 0 , 1 ; lim x + [ | u ( x ) |
+ | u x ( x ) | + | u x x ( x ) | ] = 0 ; t ( 0 , T ) ,
u ( x , 0 ) = u 0 ( x ) , x R + .
Theorem 2.
Let T be an arbitrary positive number and a natural k 4 . Given u 0 H 5 ( R + ) , u ( 0 ) = D x ( 0 ) = 0 . Then the problem (31) and (33) has a unique regular solution
u L ( ( 0 , T ) ; H 2 ( R + ) ) L 2 ( ( 0 , T ) ; H 7 ( R + ) ) ;
u t L ( ( 0 , T ) ; L 2 ( R + ) ) L 2 ( ( 0 , T ) ; H 2 ( R + ) ) .
Proof. 
Define the space W = { f H 5 ( R + ) , f ( 0 ) = D x f ( 0 ) = 0 } and let { w i ( x ) , i N } be a countable dense set in W. We can construct approximate solutions to (31)–(33) in the form
u N ( x , t ) = i = 1 N g i ( t ) w i ( x ) .
Unknown functions g i ( t ) satisfy the following initial problems:
d d t ( u N , w i ) ( t ) + ( D x 4 u N , w i ) ( t ) + ( D x 2 u N , w i ) ( t )
( D x 5 u N , w i ) ( t ) + ( D x 3 u N , w i ) ( t ) + ( ( u N ) k D x u N , w i ) ( t ) = 0 ,
g i ( 0 ) = g i 0 , i = 1 , 2 , . . . .
By Caratheodory’s existence theorem, see [29], there exist solutions of (34) and (35), at least locally in t, hence for all finite N, we can construct an approximate solution u N ( x , t ) of (31)–(33). Taking this into account, all the estimates we will prove will be done on smooth solutions of (31)–(33). Naturally, the same estimates are true also for approximate solutions u N .
Estimate I. Multiplying (31) by 2 u and acting in the same manner as by proving (13) and (14), we obtain
u 2 ( t ) e T u 0 2 , t ( 0 , T ) ;
0 T D x 2 u 2 ( t ) + | D x 2 u ( 0 , t ) | 2 d t ( 1 + T e T ) u 0 2 .
Here and henceforth, T is an arbitrary positive number,
u 2 ( t ) = R + u 2 ( x , t ) d x .
Estimate II. Differentiate (31) with respect to t, then multiply the result by 2 u t to get:
The case k < 4 .
u t 2 ( t ) C 1 u t 2 ( 0 ) , t ( 0 , T ) ,
where
C 1 = exp { 0 T 3 2 + 2 k 2 a k D x 2 u 2 ( t ) + 4 k 4 ( e T u 0 ) 3 k / ( 4 k ) d t } ,
0 T D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2 d t C ( T , u 0 ) u t 2 ( 0 )
and u t ( 0 ) C ( u 0 W ) can be estimated directly from (31) on t = 0 . The case k = 4 .
Acting similarly to the case k < 4 , we find
u 2 ( t ) e T u 0 2 , t ( 0 , T ) ,
0 T D x 2 u 2 ( t ) + | D x 2 u ( 0 , t ) | 2 d t ( 1 + T e T ) u 0 2 ,
u t 2 ( t ) C 3 u t 2 ( 0 ) , t ( 0 , T ) ,
where
C 2 = 0 T 3 2 + 2 5 e 3 T a D x 2 u 2 ( t ) u 0 6 d t ,
0 T D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2 d t C ( T , u 0 ) u t 2 ( 0 ) .
Here u t ( 0 ) C ( u 0 W ) can be estimated directly from (31) on t = 0 . Independent of N estimates (39)–(42) and standard arguments allow us to prove the existence of weak solutions to (31)–(33). Regularity and uniqueness of a regular solution can be proved in the same manner as in the proof of Theorem 1. The proof of Theorem 2 is complete.

4. Stability Intervals. Small Solutions

Our goal in this section is to determine intervals ( 0 , L ) which guarantee decay of small solutions as t + .
In Q t = ( 0 , t ) × ( D ) , D = ( 0 , L ) , consider the following initial-boundary value problem:
u t D x 5 u + D x 4 u + D x 3 u + D x 2 u + u k u x = 0 ,
D x i u ( 0 ) = D x i u ( L ) = D x 2 u ( L ) = 0 , i = 0 , 1 ;
u ( x , 0 ) = u 0 ( x ) , x ( 0 , L ) .
Theorem 3.
Let
a = π 2 L 2 > 1 , θ = 1 1 a > 0 , n a t u r a l k 8 .
Given u 0 H 5 ( D ) H 0 2 ( D ) , D x 2 u 0 ( L ) = 0 such that
θ 2 k 2 a θ [ k u 0 2 u t 2 ( 0 ) θ 2
+ ( 8 k ) u 0 12 k / ( 8 k ) ] > 0 for k < 8 ,
θ 2 8 a θ 3 u 0 14 u t 2 ( 0 ) > 0 for k = 8 .
Then (43)–(45) has a unique regular solution
u L ( R + ; H 5 ( D ) H 0 2 ( D ) ) L 2 ( R + ; H 7 ( D ) ) ;
u t L ( R + ; L 2 ( D ) ) L 2 ( R + ; H 0 2 ( D ) ) .
Moreover, u satisfies the following inequalities:
u 2 ( t ) u 0 2 exp { 2 a 2 θ t } ,
u 2 ( t ) + 2 θ 0 t D x 2 u 2 ( τ ) d τ u 0 2 ,
u t 2 ( t ) + θ 2 0 t D x 2 u τ 2 ( τ ) d τ u t 2 ( 0 ) , t > 0 ;
u t 2 ( t ) u t 2 ( 0 ) exp { a 2 θ 2 t } .
Here u t ( 0 ) 2 = C ( u 0 H 5 ( D ) ) .
Proof. 
To establish the results of Theorem 3, we use as before the same approach based on Faedo-Galerkin‘s method and we start with estimates of smooth solution of (43)–(45).
Estimate I. Multiply (43) by 2 u to obtain
d d t u 2 ( t ) + 2 D x 2 u 2 ( t ) 2 D x u 2 ( t ) + | D x 2 ( 0 , t ) | 2 = 0 .
Making use of conditions of Theorem 3 and Lemma 6, we get
d d t u 2 ( t ) + 2 θ D x 2 u 2 ( t ) 0 .
Again, by Lemma 6,
d d t u 2 ( t ) + 2 a 2 θ u 2 ( t ) 0 .
This implies
u 2 ( t ) u 0 2 exp { 2 a 2 θ t } .
Returning to (53), we obtain
u 2 ( t ) + 2 θ 0 t D x 2 u 2 ( τ ) d τ u 0 2 .
Estimate II. Differentiate (43) by t and multiply the result by 2 u t to obtain
d d t u t 2 ( t ) + 2 θ D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2
2 ( u k u t , u x t ) ( t ) .
The case k < 8 .
Maiking use of Lemma 2, for an arbitrary ϵ > 0 , we estimate
I = 2 ( u k u t , u x t ) ( t ) 2 sup D | u ( x , t ) | k u t ( t ) u x t ( t )
2 2 1 / 2 D x 2 u 1 / 4 ( t ) u 3 / 4 ( t ) k u t ( t ) 1 a 1 / 2 u x x t ( t )
ϵ u x x t 2 ( t ) + 2 k a ϵ D x 2 u k / 2 u 3 k / 2 ( t ) u t 2 ( t )
ϵ u x x t 2 ( t ) + 2 k a ϵ k D x 2 u 4 ( t ) 8 + 8 k 8 u 12 k / ( 8 k ) ( t ) u t 2 ( t ) .
Rewrite (53) as
D x 2 u 2 ( t ) 1 θ u t ( t ) u ( t ) .
Substitute this into I , to get
I = 2 ( u k u t , u x t ) ( t ) 2 sup D | u ( x , t ) | k u t ( t ) u x t ( t )
ϵ u x x t 2 ( t ) + 2 k 3 a ϵ [ k u 0 2 u t 2 ( t ) θ 2
+ ( 8 k ) u 0 12 k / ( 8 k ) ] u t 2 ( t ) .
Taking 2 ϵ = θ and substituting I into (56), we obtain
d d t u t 2 ( t ) + θ 2 D x 2 u t 2 ( t ) + ( θ 2 k 2 a θ [ k u 0 2 u t 2 ( t ) θ 2
+ ( 8 k ) u 0 12 k / ( 8 k ) ] ) u t 2 ( t ) 0 .
Making use of (46), (57), positivity of second term in (58) and Lemma 5, we get
θ 2 k 2 a θ k u 0 2 u t 2 ( t ) θ 2 + ( 8 k ) u 0 12 k / ( 8 k ) > 0 , t > 0 .
Hence (58) becomes
d d t u t 2 ( t ) + θ 2 D x 2 u t 2 ( t ) 0 .
Integration gives
u t 2 ( t ) + θ 2 0 t D x 2 u τ 2 ( τ ) d τ u t 2 ( 0 ) , t > 0 .
On the other hand, by Lemma 6, (59) can be rewritten as
d d t u t 2 ( t ) + a 2 θ 2 u t 2 ( t ) 0 .
Hence
u t 2 ( t ) u t 2 ( 0 ) exp { a 2 θ 2 t } .
The case k = 8 .
Since (54), (55) are true for all natural k, consider (56) for k = 8 :
d d t u t 2 ( t ) + 2 θ D x 2 u t 2 ( t ) + | D x 2 u t ( 0 , t ) | 2
+ 2 ( u 8 u t , u x t ) ( t ) .
Making use of Lemma 2, we estimate
I = 2 ( u 8 u t , u x t ) ( t ) 2 sup D | u ( x , t ) | 8 u t ( t ) u x t ( t ) 2 2 1 / 2 D x 2 u 1 / 4 ( t ) u 3 / 4 ( t ) 8 u t ( t ) 1 a 1 / 2 u x x t ( t ) ϵ u x x t 2 ( t ) + 2 8 a ϵ D x 2 u 4 u 12 ( t ) u t 2 ( t ) .
Taking 2 ϵ = θ , making use of (60) and substituting I into (63), we obtain
d d t u t 2 ( t ) + θ 2 D x 2 u t 2 ( t )
+ θ 2 10 a θ 3 u 0 14 u t 2 ( t ) u t 2 ( t ) 0 .
Making use of (47), positivity of second term in (64) and Lemma 5, we get
d d t u t 2 ( t ) + θ 2 D x 2 u t 2 ( t ) 0 .
By Lemma 6, this implies
u t 2 ( t ) u t 2 ( 0 ) exp { a 2 θ 2 t } ;
u t 2 ( t ) + θ 2 0 t D x 2 u τ 2 ( τ ) d τ u t 2 ( 0 ) , t > 0 .
Acting in the same manner as in the case k 4 , we find
u L ( ( 0 , T ) ; H 5 ( D ) ) L 2 ( ( 0 , T ) ; H 7 ( D ) ) .
This proves the existence part of Theorem 3. Uniqueness of this regular solution has been proved in Lemma 7. It means that the proof of Theorem 3 is complete.

5. Conclusions

In this work, we studied initial-boundary value problems for the generalized Benney-Lin Equation (2) posed on bounded intervals and on a half-line. In the case of an interval ( 0 , L ) , where L is an arbitrary positive number, we proved the existence and uniqueness of a regular solution without smallness conditions on the initial data for k 4 and all t ( 0 , T ) , where T is an arbitrary positive number. For a special choice of ( 0 , L ) , we proved for k 8 the existence and uniqueness of small regular solutions as well as their exponential decay as t + . In the case of the right half-line, we proved for k 4 and arbitrary positive number T the existence and uniqueness of a regular solution without smallness conditions for the initial data.

Funding

This research received no external funding.

Data Availability Statement

All the data that I have used are mathematical theorems available in published articles and books, presented in References and propérly cited in the text.

Acknowledgments

The author appreciated very much attention and support of my submission by anonymous editor as well as concrete and profound comments of the reviewers.

Conflicts of Interest

The author declares that there is no conflict of interest regarding the publication of this paper.

References

  1. Benney, D.J. Long waves on liquid films. J. Math. Phys. 1966, 45, 150–155. [Google Scholar] [CrossRef]
  2. Lin, S.P. Finite amplitude side band stability of a viscous film. J. Fluid Mech. 1974, 63, 417–429. [Google Scholar] [CrossRef]
  3. Kuramoto, Y.; Tsuzuki, T. On the formation of dissipative structures in reaction-diffusion systems. Progr. Theor. Phys. 1975, 54, 687–699. [Google Scholar] [CrossRef] [Green Version]
  4. Sivashinsky, G.I. Nonlinear analysis of hydrodinamic instability in laminar flames-1. Derivation of basic equations. Acta Astronauica 1977, 4, 1177–1206. [Google Scholar] [CrossRef]
  5. Cross, M.C. Pattern formation outside of equilibrium. Rev. Mod. Phys. 1993, 65, 851–1086. [Google Scholar] [CrossRef] [Green Version]
  6. Biagioni, H.A.; Bona, J.L.; Iorio, R.J., Jr.; Scialom, M. On the Korteweg de Vries-Kuramoto-Sivashinsky Equation. Adv. Differ. Equations 1996, 1, 1–20. [Google Scholar]
  7. Cousin, A.T.; Larkin, N.A. Kuramoto-Sivashinsky equation in domains with moving boundaries. Port. Math. 2002, 59, 335–349. [Google Scholar]
  8. Doronin, G.G.; Larkin, N.A. Kawahara equation in a bounded domain. Discret. Contin. Dyn. Syst. Ser. B 2008, 10, 783–799. [Google Scholar] [CrossRef]
  9. Feng, B.F.; Malomed, B.A.; Kawahara, T. Stable Periodic Waves in Coupled Sivashinsky-Korteweg-de Vries Equations. arXiv 2002, arXiv:nlin/0209003. [Google Scholar] [CrossRef] [Green Version]
  10. Kato, T. On the Cauchy problem for the (generalized) Korteweg-de Vries equation. Stud. Appl. Math. Adv. Math. Suppl. Stud. 1983, 8, 93–128. [Google Scholar]
  11. Larkin, N.A. Correct initial boundary value problems for dispersive equations. J. Math. Anal. Appl. 2008, 344, 1079–1092. [Google Scholar] [CrossRef] [Green Version]
  12. Larkin, N.A. Korteweg de Vries and Kuramoto-Sivashinsky equations in bounded domains. J. Math. Anal. Appl. 2004, 297, 169–185. [Google Scholar] [CrossRef]
  13. Larkin, N.A.; Simões, M.H. The Kawahara equation on bounded intervals and on a half-line. Nonlinear Anal. 2015, 127, 397–412. [Google Scholar] [CrossRef]
  14. Nicolaenko, B.; Scheurer, B.; Temam, R. Some global dynamical properties of the Kuramoto-Sivashinsky equations: Nonlinear stability and attractors. Phys. D 1985, 16, 155–183. [Google Scholar] [CrossRef]
  15. Li, J.; Zhang, B.-Y.; Zhang, Z. A nonhomogeneous boundary value problem for the Kuramoto-Sivashinsky equation in a quarter plane. arXiv 2016, arXiv:1607.00506v2. [Google Scholar] [CrossRef] [Green Version]
  16. Faminskii, A.V. Odd-order quasilinear evolution equations with general nonlinearity on bounded intervals. Lobachevskii J. Math. 2021, 42, 875–888. [Google Scholar] [CrossRef]
  17. Faminskii, A.V.; Larkin, N.A. Initial-boundary value problems for quasilinear dispersive equations posed on a bounded interval. Electron. J. Differ. Equ. 2010, 2010, 1–20. [Google Scholar]
  18. Larkin, N.A.; Luchesi, J. Initial-boundary value problems for generalized dispersive equations of higher orders posed on bounded intervals. J. Appl. Math. Optim. 2019, 83, 1081–1102. [Google Scholar] [CrossRef]
  19. Fonseca, G.; Linares, F.; Ponce, G. Global existence for the critical generalized KDV equation. Proc. AMS 2002, 131, 1847–1855. [Google Scholar] [CrossRef]
  20. Martel, Y.; Merle, F. Instability of solutions for the critical generalized Korteweg-de Vries equation. Geom. Funct. Anal. 2001, 11, 74–123. [Google Scholar] [CrossRef]
  21. Jeffrey, A.; Kakutani, T. Weak nonlinear dispersive waves: A discussion centered around the Korteweg-de Vries equation. SIAM Rev. 1972, 14, 582–643. [Google Scholar] [CrossRef]
  22. Linares, F.; Pazoto, A. On the exponential decay of the critical generalized Korteweg-de Vries equation with localized damping. Proc. Am. Math. Soc. 2007, 135, 1515–1522. [Google Scholar] [CrossRef] [Green Version]
  23. Larkin, N.A. Regular global solutions for a generalized KdV equation posed on a half-line. arXiv 2020, arXiv:2006.06645v1. [Google Scholar]
  24. Castelli, M.; Doronin, G. Modified and subcritical Zakharov-Kuznetsov equations posed on rectangles. J. Differ. Equ. 2021, 275, 554–580. [Google Scholar] [CrossRef]
  25. Adams, R.A.; Fournier, J.F. Sobolev Spaces; Elsevier Sciuence Ltd.: Oxford, UK, 2003. [Google Scholar]
  26. Brezis, H. Functional Analysis, Sobolev Spaces and Partial Differential Equations; Springer: New York, NY, USA, 2011. [Google Scholar]
  27. Steklov, A.V. The problem of cooling of an heterogeneous rigid rod. Commun. Kharkov Math. Soc. Ser. 1896, 2, 136–181. (In Russian) [Google Scholar]
  28. Nirenberg, L. On elliptic partial differential equations. In Annali Della Scuola Normale Superiore di Pisa; Classe di Scienze 3ª Série; Springer: Berlin/Heidelberg, Germany, 1959; pp. 115–162. [Google Scholar]
  29. Coddington, E.; Levinson, N. Theory of Ordinary Differential Equations; MacGraw-Hill: New York, NY, USA, 1955. [Google Scholar]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Larkin, N.A. Regularity and Decay of Global Solutions for the Generalized Benney-Lin Equation Posed on Bounded Intervals and on a Half-Line. Axioms 2022, 11, 596. https://doi.org/10.3390/axioms11110596

AMA Style

Larkin NA. Regularity and Decay of Global Solutions for the Generalized Benney-Lin Equation Posed on Bounded Intervals and on a Half-Line. Axioms. 2022; 11(11):596. https://doi.org/10.3390/axioms11110596

Chicago/Turabian Style

Larkin, Nikolai A. 2022. "Regularity and Decay of Global Solutions for the Generalized Benney-Lin Equation Posed on Bounded Intervals and on a Half-Line" Axioms 11, no. 11: 596. https://doi.org/10.3390/axioms11110596

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop