Open Access

Blow-up for the weakly dissipative generalized Camassa-Holm equation

Journal of Inequalities and Applications20142014:514

https://doi.org/10.1186/1029-242X-2014-514

Received: 17 September 2014

Accepted: 8 December 2014

Published: 18 December 2014

Abstract

The main goal of this paper is to investigate the blow-up phenomena of solutions to a weakly dissipative generalized Camassa-Holm equation, which contains higher power nonlinear dispersion terms and a convection term. We give a sufficient condition on the initial data such that the strong solution blows up at a finite time, and then we establish an estimate of the blow-up time. Finally, we give a global existence result of the strong solution.

MSC:35A35, 35B30, 35G25, 35Q53.

Keywords

generalized Camassa-Holm equationweak dissipativityblow-upglobal existence

1 Introduction

In recent years, following the research of the Burgers equation, the KdV equation, and the BBM equation [1], the generalized Camassa-Holm equation
u t u x x t + ( k + 2 ) u k u x ( k + 1 ) u k 1 u x u x x u k u x x x = 0
(1.1)

has attracted much attention in the study of mathematical physics, where k 1 , k N .

For k = 1 , (1.1) is reduced to the classical Camassa-Holm equation
u t u x x t + 3 u u x 2 u x u x x u u x x x = 0 ,
(1.2)

which describes the unidirectional propagation of waves at the free surface of shallow water. u ( t , x ) stands for the fluid velocity at time t in the spatial direction x. The Camassa-Holm equation (1.2) is bi-Hamiltonian and admits an infinite number of conservation laws [2]. The Camassa-Holm equation (1.2) has been extensively studied by Constantin and Escher [36], Lai and Wu [7], and so on. The well-posedness of the Camassa-Holm shallow water equation has been established, and some blow-up scenarios were derived by Constantin and Escher [8], Wu and Yin [9, 10], Lai and Wu [11], Zhou [12], Xin and Zhang [13, 14].

For k = 2 , (1.1) becomes the Novikov equation
u t u x x t + 4 u 2 u x 3 u u x u x x u 2 u x x x = 0 ,
(1.3)

which was recently discovered by Novikov [15]. Since the Novikov equation possesses a matrix Lax pair and has a bi-Hamiltonian structure as the Camassa-Holm equation, this equation has been studied by many researchers in the past few years, the well-posedness and persistence properties were studied by Lai et al. [16], Ni and Zhou [17], Zhao et al. [18]. Jiang and Ni [19] considered the blow-up phenomena for the integrable Novikov equation, Yan et al. [20] gave the global existence and blow-up phenomena for the weakly dissipative Novikov equation.

In this paper, we investigate the Cauchy problem for a generalized weakly dissipative Camassa-Holm equation
{ u t u x x t + ( k + 2 ) u k u x ( k + 1 ) u k 1 u x u x x u k u x x x + λ ( u u x x ) = 0 , u ( 0 , x ) = u 0 ( x ) , x R ,
(1.4)

where λ > 0 , k 1 is a positive integer. Equations (1.4) and (1.1) have similar properties as regards the local well-posedness and blow-up phenomena, but they are different as regards the long time behavior. For example, when k = 1 , (1.1) is completely integrable and has an infinite number of conservation laws, but for the corresponding equation (1.4), ( u 2 + u x 2 ) d x is not conservative.

Zhao et al. [21] studied the existence of global weak solutions to the Cauchy problem of the generalized Novikov equation (1.1). Liu and Yin [22] investigated the blow-up phenomena for the Degasperis-Procesi equation
u t u t x x + 4 u u x = 3 u x u x x + u u x x x , t > 0 , x R ,
(1.5)

it is very similar with (1.4), but (1.4) contains the higher power nonlinear dispersion terms ( k + 1 ) u k 1 u x u x x , u k u x x x , and the nonlinear convection term ( k + 2 ) u k u x .

Compared to [22], the main difficulty in this paper comes from the nonlinear effect of higher power nonlinear dispersion terms ( k + 1 ) u k 1 u x u x x , u k u x x x , and the nonlinear convection term ( k + 2 ) u k u x . On the other hand, in the proof of the blow-up property of the solution to (1.4), we need the sign of the term u k 2 ( t , x ) , but u ( t , x ) changes the sign for x R . Compared to the classical Camassa-Holm equation ( k = 1 ) and the classical Novikov equation ( k = 2 ), the term u k 2 ( t , x ) disappears, accordingly. Therefore, we generalized the blow-up property of solutions to the Cauchy problem (1.4).

We first give a sufficient condition on the initial data such that the strong solution of (1.4) blows up at a finite time, and then we establish an estimate of the blow-up time. Finally, we give a global existence result of the strong solution of (1.4).

The paper is organized as follows. In Section 2, we give some preliminaries used in our investigation. In Section 3, we give in our main conclusion the blow-up scenario and global existence result.

2 Preliminaries

We first review some notations. The convolution between two functions f ( x ) and g ( x ) :
( f g ) ( x ) = R f ( x y ) g ( y ) d y , f , g S ,
where S is the Schwartz class. For any f ( x ) S , the Fourier transform of f ( x ) is defined by F ( f ( x ) ) = f ˆ ( ξ ) , the inverse Fourier transform of f ˆ ( ξ ) denoted by F 1 ( f ˆ ( ξ ) ) . If f ( x ) H s , s R , then the norm of f ( x ) is
f H s = ( R ( 1 + | ξ | 2 ) s | f ˆ ( ξ ) | 2 d ξ ) 1 2 .
Set y = u u x x , the Cauchy problem (1.4) becomes
{ y t + u k y x + ( k + 1 ) u k 1 u x y + λ y = 0 , x R , t > 0 , u ( 0 , x ) = u 0 ( x ) , x R .
(2.1)
Since G ( x ) = 1 2 e | x | is the Green’s function of the differential equation u u x x = δ ( x ) , for all f ( x ) L 2 ( R ) , G f ( x ) = ( 1 x 2 ) 1 f ( x ) , and G y = u ( t , x ) , and thus the Cauchy problem (2.1) can be rewritten as
{ u t + u k u x + G [ k ( k 1 ) u k 2 u x 3 + ( 2 k 1 ) u k 1 u x u x x + ( k + 1 ) u k u x ] + λ u = 0 , u ( 0 , x ) = u 0 ( x ) .
(2.2)

Zhao et al. [18, 21] gave the local and global existence of solutions to the Cauchy problem (2.2), it is crucial in our discussion.

Lemma 2.1 [18]

Given u 0 ( x ) H s ( R ) , s > 3 2 , then there exist a constant T = T ( u 0 ) > 0 and a unique solution u ( t , x ) to (2.2) such that
u ( t , x ) C ( [ 0 , T ) ; H s ( R ) ) C 1 ( [ 0 , T ) ; H s 1 ( R ) ) .

Moreover, the mapping u 0 u ( , u 0 ) : H s ( R ) C ( [ 0 , T ) ; H s ( R ) ) C 1 ( [ 0 , T ) ; H s 1 ( R ) ) is Hölder continuous.

We now describe some properties of solutions of the following initial value problem:
{ q t ( t , x ) = u k ( t , q ( t , x ) ) , t > 0 , x R , q ( 0 , x ) = x , x R ,
(2.3)

where u ( t , x ) is a solution to the Cauchy problem (2.2). The following important properties are immediate consequence of the classical results in the theory of ordinary differential equations.

Lemma 2.2 Let u 0 ( x ) H s ( R ) , s > 3 2 , and T = T ( u 0 ) > 0 be the maximal existence time of the corresponding solution u ( t , x ) to (2.2), then the problem (2.3) has a unique solution q C 1 ( [ 0 , T ) × R ; R ) . Moreover, the map q ( t , ) is an increasing diffeomorphism of R with
q x ( t , x ) = exp ( k 0 t u k 1 u x ( s , x ) d s ) , ( t , x ) [ 0 , T ) × R .
(2.4)
Lemma 2.3 Let u 0 ( x ) H s ( R ) , s > 3 2 , and T = T ( u 0 ) > 0 be the maximal existence time of the corresponding solution u ( t , x ) to (2.2). For y ( t , x ) = u u x x , we have
y ( t , q ( t , x ) ) q x k + 1 k ( t , x ) = y 0 ( x ) e λ t , ( t , x ) [ 0 , T ) × R .
(2.5)
Proof Let P ( t ) = y ( t , q ( t , x ) ) q x k + 1 k ( t , x ) . Thanks to (2.1) and (2.3), we have
d P d t = y t ( t , q ( t , x ) ) q x k + 1 k ( t , x ) + y x ( t , q ( t , x ) ) q t ( t , x ) q x k + 1 k ( t , x ) + k + 1 k y ( t , q ( t , x ) ) q x 1 k ( t , x ) q x t ( t , x ) = q x k + 1 k ( t , x ) [ y t ( t , q ( t , x ) ) + ( k + 1 ) u k 1 u x y ( t , q ( t , x ) ) + y x ( t , q ( t , x ) ) u k ] = λ y ( t , q ( t , x ) ) q x k + 1 k ( t , x ) = λ P ( t ) ,
the solution of ordinary differential equation is P ( t ) = P ( 0 ) e λ t . Since q ( 0 , x ) = x , q x ( 0 , x ) = 1 , we have
y ( t , q ( t , x ) ) q x k + 1 k ( t , x ) = y 0 ( x ) e λ t .

This concludes the proof. □

Lemma 2.4 Let u ( t , x ) be the solution to (2.2). Then we have
R ( u 2 + u x 2 ) d x = e 2 λ t R ( u 0 2 + u 0 x 2 ) d x .
(2.6)

Proof When λ = 0 , Lemma 2.4 is a case of Lemma 2.8 in Zhao et al. [21]. The proof carries over with a slight modification and we present it here for the reader’s convenience.

Thanks to y = u u x x and integrating by parts, we have
R y u d x = R ( u 2 + u x 2 ) d x ,
thus,
d d t R ( u 2 + u x 2 ) d x + 2 λ R ( u 2 + u x 2 ) d x = d d t R y u d x + 2 λ R y u d x = R ( y u t + u y t ) d x + 2 λ R y u d x .
Together with (2.1) and (2.2), on integration by parts we have
R ( y u t + u y t ) d x + 2 λ R y u d x = R [ u k + 1 y x + ( k + 1 ) u k u x y ] d x R u k u x y R y G [ k ( k 1 ) u k 2 u x 3 + ( 2 k 1 ) u k 1 u x u x x + ( k + 1 ) u k u x ] d x = 0 .
Therefore,
d d t R ( u 2 + u x 2 ) d x + 2 λ R ( u 2 + u x 2 ) d x = 0 .

Integrating with respect to t from 0 to t, we get the desired conclusion. □

Following the proof of Lemma 2.9 given by Zhao et al. in [21], we can obtain a similar blow-up result of the solution to the Cauchy problem (2.2).

Theorem 2.1 Let u 0 ( x ) H s ( R ) , s > 3 2 , and T = T ( u 0 ) > 0 be the maximal existence time of the corresponding solution u ( t , x ) to (2.2), then u ( t , x ) blows up if and only if
lim sup t T u x ( t , x ) L = + .
(2.7)

3 Blow-up and global existence

Following the local existence Theorem 2.1, we will give our main result on the blow-up property of solution to (2.2). We first give a sufficient condition to guarantee that the solution blows up at a finite time.

Theorem 3.1 Let u 0 ( x ) H s ( R ) , s > 3 2 , and T = T ( u 0 ) > 0 be the maximal existence time of the corresponding solution u ( t , x ) to (2.2). Assume k = 1 or k = 2 n , n is a positive integer, if there exists an x 0 R such that y 0 ( x ) = ( 1 x 2 ) u 0 ( x ) satisfies
y 0 ( x ) 0 for  ( , x 0 ) and y 0 ( x ) 0 for  ( x 0 , + ) ,
(3.1)
and
u 0 k 1 ( x 0 ) u 0 x ( x 0 ) < k λ 1 2 k u 0 H 1 2 k + k 2 λ 2 .
(3.2)
Then the corresponding solution to (2.2) with initial data u 0 ( x ) blows up at finite time T with
T min { 2 ( 1 δ ) m ( 0 ) , 1 A ln m ( 0 ) A m ( 0 ) + A } ,
where 0 < δ < 1 , such that
δ m ( 0 ) = 1 2 k u 0 H 1 2 k + k 2 λ 2 ,
and
A = 1 2 k u 0 H 1 2 k + k 2 λ 2 , m ( 0 ) = u 0 k 1 ( x 0 ) u 0 x ( x 0 ) + k λ .

Proof For k = 1 , the result can be found in Wu and Yin [10]. We just show that the results hold for k = 2 n , n N , and the initial data u 0 H 3 ( R ) , for the general case we can use the smooth approximate technique and denseness.

Let T > 0 be the maximal existence time of the solution u ( t , x ) to (2.2) with initial data u 0 ( x ) . Thanks to (2.4), (2.5), and (3.1), we have y ( t , q ( t , x 0 ) ) = 0 , and, for all t > 0 , we have
y ( t , q ( t , x ) ) 0 , for  x ( , x 0 ) , y ( t , q ( t , x ) ) 0 , for  x ( x 0 , + ) .
(3.3)
With the help of u ( , x ) = G y ( , x ) , x R , we have
u ( t , x ) = 1 2 e x x e ξ y ( t , ξ ) d ξ + 1 2 e x x + e ξ y ( t , ξ ) d ξ
(3.4)
and
u x ( t , x ) = 1 2 e x x e ξ y ( t , ξ ) d ξ + 1 2 e x x + e ξ y ( t , ξ ) d ξ .
(3.5)
After direct calculations we get
u ( t , x ) + u x ( t , x ) = e x x e ξ y ( ξ , t ) d ξ ,
(3.6)
u ( t , x ) u x ( t , x ) = e x x e ξ y ( ξ , t ) d ξ ,
(3.7)
and
u ( t , x ) u x ( t , x ) = 1 4 e 2 x ( x e ξ y ( t , ξ ) d ξ ) 2 + 1 4 e 2 x ( x + e ξ y ( t , ξ ) d ξ ) 2 .
(3.8)
Thanks to Lemma 2.1, u 0 ( x ) H 3 ( R ) implies that
u ( t , x ) C ( [ 0 , T ) ; H 3 ( R ) ) C 1 ( [ 0 , T ) ; H 2 ( R ) ) ,

then u ( t , ) C 2 ( R ) , u t ( t , x ) C ( [ 0 , T ) ; H 2 ( R ) ) , and u x t ( t , x ) C ( [ 0 , T ) ; H 1 ( R ) ) .

From (3.6) to (3.8) we have
d d t [ ( k u k 1 u x ) ( t , q ( t , x 0 ) ) ] = k ( k 2 ) u k 2 u x u t + k u k 2 d d t ( u u x ) = k ( k 2 ) u k 2 u x u t + k u k 2 d d t [ 1 4 e 2 q ( t , x 0 ) ( q ( t , x 0 ) e ξ y ( ξ , t ) d ξ ) 2 + 1 4 e 2 q ( t , x 0 ) ( q ( t , x 0 ) + e ξ y ( ξ , t ) d ξ ) 2 ] = k ( k 2 ) u k 2 u x u t + k u k 2 [ 1 2 e 2 q ( t , x 0 ) ( q ( t , x 0 ) e ξ y ( ξ , t ) d ξ ) 2 q t ( t , x 0 ) 1 2 e q ( t , x 0 ) ( q ( t , x 0 ) e ξ y ( ξ , t ) d ξ ) y ( t , q ( t , x 0 ) ) q t ( t , x 0 ) 1 2 e 2 q ( t , x 0 ) ( q ( t , x 0 ) e ξ y ( ξ , t ) d ξ ) ( q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ ) + 1 2 e 2 q ( t , x 0 ) ( q ( t , x 0 ) + e ξ y ( ξ , t ) d ξ ) 2 q t ( t , x 0 ) + 1 2 e q ( t , x 0 ) ( q ( t , x 0 ) e ξ y ( ξ , t ) d ξ ) y ( t , q ( t , x 0 ) ) q t ( t , x 0 ) + 1 2 e 2 q ( t , x 0 ) ( q ( t , x 0 ) + e ξ y ( ξ , t ) d ξ ) ( q ( t , x 0 ) + e ξ y t ( ξ , t ) d ξ ) ] .
(3.9)
Notice y ( t , q ( t , x 0 ) ) = 0 , using (2.3), (3.6), and (3.7) we have
d d t [ ( k u k 1 u x ) ( t , q ( t , x 0 ) ) ] = k ( k 2 ) u k 2 u x u t + k u k 2 [ 1 2 e 2 q ( t , x 0 ) ( q ( t , x 0 ) e ξ y ( ξ , t ) d ξ ) 2 q t ( t , x 0 ) 1 2 e 2 q ( t , x 0 ) ( q ( t , x 0 ) e ξ y ( ξ , t ) d ξ ) ( q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ ) + 1 2 e 2 q ( t , x 0 ) ( q ( t , x 0 ) + e ξ y ( ξ , t ) d ξ ) 2 q t ( t , x 0 ) + 1 2 e 2 q ( t , x 0 ) ( q ( t , x 0 ) + e ξ y ( ξ , t ) d ξ ) ( q ( t , x 0 ) + e ξ y t ( ξ , t ) d ξ ) ] = k ( k 2 ) u k 2 u x u t + k 2 u 2 k 2 ( u u x ) 2 + k 2 u 2 k 2 ( u + u x ) 2 k 2 u k 2 ( u u x ) e q ( t , x 0 ) q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ + k 2 u k 2 ( u + u x ) e q ( t , x 0 ) q ( t , x 0 ) + e ξ y t ( ξ , t ) d ξ .
(3.10)
Now we calculate the first term on the right hand side of (3.10). From (2.3), (3.4), and y ( t , q ( t , x 0 ) ) = 0 , we obtain
k ( k 2 ) ( u k 2 u x u t ) ( t , q ( t , x 0 ) ) = k ( k 2 ) u k 2 u x [ 1 2 e q ( t , x 0 ) q t ( t , x 0 ) q ( t , x 0 ) e ξ y ( ξ , t ) d ξ + 1 2 y ( t , q ( t , x 0 ) ) q t ( t , x 0 ) + 1 2 e q ( t , x 0 ) q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ + 1 2 e q ( t , x 0 ) q t ( t , x 0 ) q ( t , x 0 ) + e ξ y ( ξ , t ) d ξ 1 2 y ( t , q ( t , x 0 ) ) q t ( t , x 0 ) + 1 2 e q ( t , x 0 ) q ( t , x 0 ) + e ξ y t ( ξ , t ) d ξ ] = k ( k 2 ) u 2 k 2 u x 2 + k ( k 2 ) u k 2 u x [ 1 2 e q ( t , x 0 ) q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ + 1 2 e q ( t , x 0 ) q ( t , x 0 ) + e ξ y t ( ξ , t ) d ξ ] .
(3.11)
Substituting (3.11) into (3.10) yields
d d t [ ( k u k 1 u x ) ( t , q ( t , x 0 ) ) ] = k ( k 2 ) u 2 k 2 u x 2 + k 2 u 2 k 2 ( u u x ) 2 + k 2 u 2 k 2 ( u + u x ) 2 k 2 u k 2 [ u ( k 1 ) u x ] e q ( t , x 0 ) q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ + k 2 u k 2 [ u + ( k 1 ) u x ] e q ( t , x 0 ) q ( t , x 0 ) + e ξ y t ( ξ , t ) d ξ .
(3.12)
By (2.1), integration by parts gives
e q ( t , x 0 ) q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ = e q ( t , x 0 ) [ q ( t , x 0 ) e ξ ( y u k ) ξ d ξ + q ( t , x 0 ) e ξ d y u k 1 u ξ ξ + λ q ( t , x 0 ) e ξ y d ξ ] = e q ( t , x 0 ) q ( t , x 0 ) e ξ y u k 1 ( u u ξ ) d ξ λ e q ( t , x 0 ) q ( t , x 0 ) e ξ y d ξ .
Thanks to y = u u x x and (3.7), we get
e q ( t , x 0 ) q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ = e q ( t , x 0 ) q ( t , x 0 ) e ξ ( u u ξ ξ ) u k 1 ( u u ξ ) d ξ λ e q ( t , x 0 ) q ( t , x 0 ) e ξ y d ξ = e q ( t , x 0 ) q ( t , x 0 ) e ξ u k 1 ( u 2 u u ξ ξ u u ξ + u ξ u ξ ξ ) d ξ λ e q ( t , x 0 ) q ( t , x 0 ) e ξ y d ξ = e q ( t , x 0 ) q ( t , x 0 ) e ξ ( u k 1 u ξ u ξ ξ u k u ξ ξ u k u ξ + u k + 1 ) d ξ λ e q ( t , x 0 ) q ( t , x 0 ) e ξ y d ξ .
Since
( u k u x ) x = k u k 1 u x 2 + u k u x x , ( u k 1 u x 2 ) x = ( k 1 ) u k 2 u x 3 + 2 u k 1 u x u x x ,
we have
e q ( t , x 0 ) q ( t , x 0 ) e ξ u k 1 u ξ u ξ ξ d ξ = 1 2 e q ( t , x 0 ) q ( t , x 0 ) e ξ [ ( u k 1 u ξ 2 ) ξ ( k 1 ) u k 2 u ξ 3 ] d ξ = 1 2 ( u k 1 u x 2 ) ( t , q ( t , x 0 ) ) 1 2 e q ( t , x 0 ) q ( t , x 0 ) e ξ [ u k 1 u ξ 2 + ( k 1 ) u k 2 u ξ 3 ] d ξ ,
and
e q ( t , x 0 ) q ( t , x 0 ) e ξ u k u ξ ξ d ξ = e q ( t , x 0 ) q ( t , x 0 ) e ξ [ ( u k u ξ ) ξ k u k 1 u ξ 2 ] d ξ = ( u k u x ) ( t , q ( t , x 0 ) ) + e q ( t , x 0 ) q ( t , x 0 ) e ξ ( u k u ξ + k u k 1 u ξ 2 ) d ξ ,
therefore,
e q ( t , x 0 ) q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ = ( u k u x ) ( t , q ( t , x 0 ) ) + 1 2 ( u k 1 u x 2 ) ( t , q ( t , x 0 ) ) + e q ( t , x 0 ) q ( t , x 0 ) e ξ ( u k + 1 + 2 k 1 2 u k 1 u ξ 2 k 1 2 u k 2 u ξ 3 ) d ξ λ ( u u x ) ( t , q ( t , x 0 ) ) = ( u k u x ) ( t , q ( t , x 0 ) ) + 1 2 ( u k 1 u x 2 ) ( t , q ( t , x 0 ) ) λ ( u u x ) ( t , q ( t , x 0 ) ) + e q ( t , x 0 ) q ( t , x 0 ) e ξ 2 [ u k + 1 + u k 2 ( u u ξ ) 2 ( u ( k 1 ) u ξ ) + ( k + 1 ) u k u ξ ] d ξ = ( u k u x ) ( t , q ( t , x 0 ) ) + 1 2 ( u k 1 u x 2 ) ( t , q ( t , x 0 ) ) λ ( u u x ) ( t , q ( t , x 0 ) ) + 1 2 u k + 1 ( t , q ( t , x 0 ) ) + e q ( t , x 0 ) q ( t , x 0 ) e ξ 2 u k 2 ( u u ξ ) 2 [ u ( k 1 ) u ξ ] d ξ .
(3.13)
Thanks to (3.3),
y ( t , q ( t , x ) ) 0 , for  x ( , x 0 ) , t 0 , y ( t , q ( t , x ) ) 0 , for  x ( x 0 , + ) , t 0 ,
together with (3.6) and (3.7),
u ( t , q ( t , x 0 ) ) + u x ( t , q ( t , x 0 ) ) = e q ( t , x 0 ) q ( t , x 0 ) e ξ y ( ξ , t ) d ξ 0 , u ( t , q ( t , x 0 ) ) u x ( t , q ( t , x 0 ) ) = e q ( t , x 0 ) q ( t , x 0 ) e ξ y ( ξ , t ) d ξ 0 ,

we have u x ( t , q ( t , x 0 ) ) 0 .

Noticing k = 2 n , n is a positive integer, we have
u ( k 1 ) u x u u x 0 , x ( , q ( t , x 0 ) ) ,
hence
q ( x 0 , t ) e ξ 2 u k 2 ( u u ξ ) 2 [ u ( k 1 ) u ξ ] d ξ 0 .
(3.14)
This implies
e q ( t , x 0 ) q ( t , x 0 ) e ξ y t ( ξ , t ) d ξ ( u k u x ) ( t , q ( t , x 0 ) ) + 1 2 ( u k 1 u x 2 ) ( t , q ( t , x 0 ) ) + 1 2 u k + 1 ( t , q ( t , x 0 ) ) λ ( u u x ) ( t , q ( t , x 0 ) ) .
(3.15)
Similarly, we repeat the above calculations and obtain
e q ( t , x 0 ) q ( t , x 0 ) + e ξ y t ( ξ , t ) d ξ = ( u k u x ) ( t , q ( t , x 0 ) ) 1 2 ( u k 1 u x 2 ) ( t , q ( t , x 0 ) ) λ ( u + u x ) ( t , q ( t , x 0 ) ) 1 2 u k + 1 ( t , q ( t , x 0 ) ) e q ( t , x 0 ) q ( t , x 0 ) + e ξ 2 u k 2 ( u + u ξ ) 2 [ u + ( k 1 ) u ξ ] d ξ .
(3.16)
Since u x ( t , q ( t , x 0 ) ) 0 , then
u + ( k 1 ) u x u + u x 0 , x ( q ( t , x 0 ) , + ) ,
we have
e q ( t , x 0 ) q ( t , x 0 ) + e ξ y t ( ξ , t ) d ξ ( u k u x ) ( t , q ( t , x 0 ) ) 1 2 ( u k 1 u x 2 ) ( t , q ( t , x 0 ) ) 1 2 u k + 1 ( t , q ( t , x 0 ) ) λ ( u + u x ) ( t , q ( t , x 0 ) ) .
(3.17)
Inserting (3.15) and (3.17) into (3.12), we get
d d t [ ( u k 1 u x ) ( t , q ( t , x 0 ) ) ] ( k 2 ) u 2 k 2 u x 2 + 1 2 u 2 k 2 ( u u x ) 2 + 1 2 u 2 k 2 ( u + u x ) 2 1 2 u k 2 [ u ( k 1 ) u x ] [ u k u x + 1 2 u k 1 u x 2 + 1 2 u k + 1 λ ( u u x ) ] + 1 2 u k 2 [ u + ( k 1 ) u x ] [ u k u x 1 2 u k 1 u x 2 1 2 u k + 1 λ ( u + u x ) ] = 1 2 u 2 k ( t , q ( t , x 0 ) ) 1 2 ( u 2 k 2 u x 2 ) ( t , q ( t , x 0 ) ) λ k ( u k 1 u x ) ( t , q ( t , x 0 ) ) .
(3.18)
Thanks to the Cauchy-Schwartz inequality,
2 u 2 ( x , t ) = 2 ( x u u x d x x + u u x d x ) x ( u 2 + u x 2 ) d x + x + ( u 2 + u x 2 ) d x = u ( x , t ) H 1 2 ,
from Lemma 2.4, we have
2 u 2 ( t , x ) u ( t , x ) H 1 2 = e 2 λ t u 0 H 1 2 u 0 H 1 2 ,
(3.19)
then
u ( t , x ) L 2 2 u 0 H 1 .
(3.20)
Combining (3.18) with (3.20), we have
d d t [ ( u k 1 u x ) ( t , q ( t , x 0 ) ) ] 1 2 k + 1 u 0 H 1 2 k 1 2 ( u 2 k 2 u x 2 ) ( t , q ( t , x 0 ) ) λ k ( u k 1 u x ) ( t , q ( t , x 0 ) ) .
(3.21)
We now define a function
m ( t ) = ( u k 1 u x ) ( t , q ( t , x 0 ) ) + λ k ,
since ( u k 1 u x ) ( t , q ( t , x 0 ) ) is continuously differentiable on [ 0 , T ) , m ( t ) is continuously differentiable on [ 0 , T ) , from (3.21), we obtain
d m ( t ) d t 1 2 k + 1 u 0 H 1 2 k 1 2 ( m ( t ) λ k ) 2 λ k ( m ( t ) λ k ) = 1 2 m 2 ( t ) + 1 2 k + 1 u 0 H 1 2 k + λ 2 k 2 2 .
(3.22)
By the assumption
m ( 0 ) = u 0 k 1 ( x 0 ) u 0 x ( x 0 ) + k λ < 1 2 k u 0 H 1 2 k + k 2 λ 2 ,

we have m 2 ( 0 ) > 1 2 k u 0 H 1 2 k + k 2 λ 2 .

We claim that
m ( t ) < 1 2 k u 0 H 1 2 k + k 2 λ 2 , t [ 0 , T ) .
(3.23)
Otherwise, if (3.23) is not true, by the continuity of m ( t ) , there exists a t 0 ( 0 , T ) such that, for all t [ 0 , t 0 ) ,
m 2 ( t ) > 1 2 k u 0 H 1 2 k + k 2 λ 2 ,
(3.24)
and
m 2 ( t 0 ) = 1 2 k u 0 H 1 2 k + k 2 λ 2 .
(3.25)
Combining (3.22) and (3.24), we have, for all t [ 0 , t 0 ] ,
d m ( t ) d t 0 ,
(3.26)
since m ( t ) is continuously differentiable on [ 0 , t 0 ] , integrating (3.26) with respect to t from 0 to t 0 , we have
m ( t 0 ) m ( 0 ) = u 0 k 1 ( x 0 ) u 0 x ( x 0 ) + k λ < 1 2 k u 0 H 1 2 k + k 2 λ 2 .
(3.27)

Recalling (3.24), we get the desired contradiction, which concludes the proof of the claim.

Since m ( t ) is continuously differentiable and strictly decreasing on [ 0 , T ) , we can choose δ ( 0 , 1 ) such that
δ m ( 0 ) = 1 2 k u 0 H 1 2 k + k 2 λ 2 ,
(3.28)
thanks to (3.22) and (3.28), we have, for all t [ 0 , T ) ,
d m ( t ) d t 1 δ 2 m 2 ( t ) .
(3.29)
Since m ( t ) is continuously differentiable and strictly negative on [ 0 , T ) , hence 1 m ( t ) is continuously differentiable on [ 0 , T ) , and
d d t ( 1 m ( t ) ) = 1 m 2 ( t ) d m ( t ) d t > 1 δ 2 , t [ 0 , T ) .
(3.30)
Integrating with respect to t over [ 0 , T ] on both sides of (3.30) yields
1 m ( t ) 1 m ( 0 ) > 1 δ 2 t , t [ 0 , T ) .
(3.31)
Since m ( t ) < 0 on [ 0 , T ) , we know that the maximal existence time is
T = 2 ( 1 δ ) m ( 0 ) < + ,
(3.32)
such that
lim t T m ( t ) lim t T 2 ( 1 δ ) ( t T ) = .
Since
inf x R ( u k 1 u x ) ( t , x ) ( u k 1 u x ) ( t , q ( t , x 0 ) ) = m ( t ) λ k ,
(3.33)
this implies
lim t T inf x R ( u k 1 u x ) ( x , t ) lim t T [ 2 ( 1 δ ) ( t T ) λ k ] = .
For A = 1 2 k u 0 H 1 2 k + k 2 λ 2 > 0 , from (3.22), we have
1 2 A ( 1 m ( t ) A 1 m ( t ) + A ) d m ( t ) = d m ( t ) m 2 ( t ) A 2 1 2 d t ,
that is,
( 1 m ( t ) A 1 m ( t ) + A ) d m ( t ) A d t ,
(3.34)
integrating with respect to t over [ 0 , t ] yields
ln m ( t ) A m ( t ) + A ln m ( 0 ) A m ( 0 ) + A A t .
As m ( t ) < A < 0 , ln m ( t ) A m ( t ) + A > 0 , we have
t 1 A ln m ( 0 ) A m ( 0 ) + A ,
(3.35)
due to m ( 0 ) < A , ln m ( 0 ) A m ( 0 ) + A > 0 , from (3.32) and (3.35), we can choose
T min { 2 ( 1 δ ) m ( 0 ) , 1 A ln m ( 0 ) A m ( 0 ) + A } .

This completes the proof. □

Remark 3.1 The result in Theorem 3.1 contains the cases for k = 1 : the weakly dissipative Camassa-Holm equation and k = 2 : the weakly dissipative Novikov equation. We used the method developed by Liu and Yin [22] to deal with the Degasperis-Procesi equation (1.5): u t u t x x + 4 u u x = 3 u x u x x + u u x x x , but (1.4) contains higher power nonlinear dispersion terms ( k + 1 ) u k 1 u x u x x , u k u x x x , and the nonlinear convection term ( k + 2 ) u k u x . When the local solution u ( t , x ) of (2.2) exists, in the proof of its blow-up property we need the sign of u k 2 ( t , x ) ; see the last term in (3.13). In general, u ( t , x ) changes the sign for x R so we give the condition on the power of nonlinear term k = 2 n , n N in (1.4). For k = 1 , the last term in (3.13) disappears; for k = 2 , the last term in (3.13) does not contain u k 2 ( t , x ) . Therefore, we generalized the blow-up property of the solutions to the Cauchy problem (1.4).

Finally we give a global existence result, thanks to Theorem 2.1, this will be done if we can estimate u x ( x , t ) L is finite.

Theorem 3.2 Let u 0 ( x ) H s ( R ) , s > 3 2 . If y 0 ( x ) = ( 1 x 2 ) u 0 ( x ) does not change sign on R, then the problem (2.2) has a strong solution
u ( x , t ) C ( [ 0 , + ) ; H s ( R ) ) C 1 ( [ 0 , + ) ; H s 1 ( R ) ) .

Proof We just consider s = 3 , otherwise we can use the smooth approximate technique and denseness. When y 0 ( x ) = ( 1 x 2 ) u 0 ( x ) 0 , then from Lemma 2.2 and Lemma 2.3, we can derive that y ( t , x ) 0 , for all [ 0 , T ) .

Due to the positivity of the Green’s function G ( x ) and u ( t , x ) = G ( x ) y ( t , x ) , we obtain u ( t , x ) 0 , for all t 0 , u ( t , x ) + u x ( t , x ) 0 , and u ( t , x ) u x ( t , x ) 0 , and these imply that, for all ( t , x ) [ 0 , T ) × R ,
| u x ( t , x ) | u ( t , x ) u ( t , x ) L 2 2 u ( t , x ) H 1 = 2 2 e λ t u 0 ( x ) H 1 u 0 ( x ) H 1 ,

we obtain u ( t , x ) C ( [ 0 , + ) ; H s ( R ) ) C 1 ( [ 0 , + ) ; H s 1 ( R ) ) by Theorem 2.1.

When y 0 ( x ) = ( 1 x 2 ) u 0 ( x ) 0 , thanks to Lemma 2.2 and Lemma 2.3, we obtain y ( t , x ) 0 , for all [ 0 , T ) . Since u ( t , x ) = G ( x ) y ( t , x ) and due to the positivity of G ( x ) , we obtain u ( t , x ) 0 , for all t 0 , u ( t , x ) + u x ( t , x ) 0 , and u ( t , x ) u x ( t , x ) 0 , and these imply that, for all ( t , x ) [ 0 , T ) × R ,
u x ( t , x ) L u ( t , x ) u ( t , x ) L 2 2 u ( t , x ) H 1 = 2 2 e λ t u 0 ( x ) H 1 u 0 ( x ) H 1 ,

we obtain u ( t , x ) C ( [ 0 , + ) ; H s ( R ) ) C 1 ( [ 0 , + ) ; H s 1 ( R ) ) by Theorem 2.1.

Therefore, we find that the solution exists globally in time. □

Declarations

Acknowledgements

This work was supported by National Natural Science Foundation of China (grant Nos. 11471129, 11272277).

Authors’ Affiliations

(1)
School of Mathematics and Statistics, Huazhong University of Science and Technology
(2)
School of Mathematics and Statistics, Xuchang University

References

  1. Guo YT, Wang M, Tang YB: Higher regularity of global attractor for a damped Benjamin-Bona-Mahony equation on  R . Appl. Anal. 2014. 10.1080/00036811.2014.946561Google Scholar
  2. Camassa R, Holm D: An integrable shallow water equation with peaked solitons. Phys. Rev. Lett. 1993, 71: 1661-1664. 10.1103/PhysRevLett.71.1661MathSciNetView ArticleMATHGoogle Scholar
  3. Constantin A, Escher J: Global weak solutions for a shallow water equation. Indiana Univ. Math. J. 1998, 47: 1527-1545.MathSciNetView ArticleMATHGoogle Scholar
  4. Constantin A, Escher J: Wave breaking for nonlocal shallow water equations. Acta Math. 1998, 181: 229-243. 10.1007/BF02392586MathSciNetView ArticleMATHGoogle Scholar
  5. Constantin A, Escher J: Global existence and blow-up for a shallow water equation. Ann. Sc. Norm. Super. Pisa, Cl. Sci. 1998,XXVI(4):303-328.MathSciNetGoogle Scholar
  6. Constantin A: Existence of permanent and breaking waves for a shallow water equation: a geometric approach. Ann. Inst. Fourier (Grenoble) 2000, 50: 321-362. 10.5802/aif.1757MathSciNetView ArticleMATHGoogle Scholar
  7. Lai S, Wu Y: The local well-posedness and existence of weak solutions for a generalized Camassa-Holm equation. J. Differ. Equ. 2010, 248: 2038-2063. 10.1016/j.jde.2010.01.008MathSciNetView ArticleMATHGoogle Scholar
  8. Constantin A, Escher J: On the blow-up rate and the blow-up set of breaking waves for a shallow water equation. Math. Z. 2000, 233: 75-91. 10.1007/PL00004793MathSciNetView ArticleMATHGoogle Scholar
  9. Wu S, Yin Z: Blowup, blowup rate and decay of the solution of the weakly dissipative Camassa-Holm equation. J. Math. Phys. 2006, 47: 1-12.MathSciNetGoogle Scholar
  10. Wu S, Yin Z: Global existence and blow-up phenomena for the weakly dissipative Camassa-Holm equation. J. Differ. Equ. 2009, 246: 4309-4321. 10.1016/j.jde.2008.12.008MathSciNetView ArticleMATHGoogle Scholar
  11. Lai S, Wu Y: Global solutions and blow-up phenomena to a shallow water equations. J. Differ. Equ. 2010, 249: 693-706. 10.1016/j.jde.2010.03.008MathSciNetView ArticleMATHGoogle Scholar
  12. Zhou Y: Blowup of solutions to the DGH equation. J. Funct. Anal. 2007, 250: 227-248. 10.1016/j.jfa.2007.04.019MathSciNetView ArticleMATHGoogle Scholar
  13. Xin Z, Zhang P: On the uniqueness and large time behavior of the weak solution to a shallow water equation. Commun. Partial Differ. Equ. 2002,27(9-10):1815-1844. 10.1081/PDE-120016129MathSciNetView ArticleMATHGoogle Scholar
  14. Xin Z, Zhang P: On the weak solution to a shallow water equation. Commun. Pure Appl. Math. 2000, 53: 1411-1433. 10.1002/1097-0312(200011)53:11<1411::AID-CPA4>3.0.CO;2-5MathSciNetView ArticleMATHGoogle Scholar
  15. Novikov V: Generalizations of the Camassa-Holm equation. J. Phys. A 2009. 42: Article ID 342002Google Scholar
  16. Lai S, Li N, Wu Y: The existence of global strong and weak solutions for Novikov equation. J. Math. Anal. Appl. 2013, 399: 682-691. 10.1016/j.jmaa.2012.10.048MathSciNetView ArticleMATHGoogle Scholar
  17. Ni L, Zhou Y: Well-posedness and persistence properties for the Novikov equation. J. Differ. Equ. 2011, 250: 3002-3201. 10.1016/j.jde.2011.01.030MathSciNetView ArticleMATHGoogle Scholar
  18. Zhao Y, Li Y, Yan W: Local well-posedness and persistence property for the generalized Novikov equation. Discrete Contin. Dyn. Syst., Ser. A 2014, 34: 803-820.MathSciNetView ArticleMATHGoogle Scholar
  19. Jiang Z, Ni L: Blow-up phenomena for the integrable Novikov equation. J. Math. Anal. Appl. 2012, 385: 551-558. 10.1016/j.jmaa.2011.06.067MathSciNetView ArticleMATHGoogle Scholar
  20. Yan W, Li Y, Zhang Y: Global existence and blow-up phenomena for the weakly dissipative Novikov equation. Nonlinear Anal. 2012, 75: 2464-2473. 10.1016/j.na.2011.10.044MathSciNetView ArticleMATHGoogle Scholar
  21. Zhao Y, Li Y, Yan W: The global weak solutions to the Cauchy problem of the generalized Novikov equation. Appl. Anal. 2014. 10.1080/00036811.2014.930826Google Scholar
  22. Liu Y, Yin Z: On the blow-up phenomena for the Degasperis-Procesi equation. Int. Math. Res. Not. 2007. 10.1093/imrn/rnm117Google Scholar

Copyright

© Guo and Tang; licensee Springer. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.