Existence for a Higher Order Coupled System of Korteweg-de Vries Equations ()
1. Introduction
It is well known that the form of the coupled nonlinear Korteweg-de Vries equations is as follows
(1.1)
where r is a real constant,
are real-valued functions of x and t,
is a coupling parameter and P, Q satisfy
for a small function H. Model represents the physical problem of describing the strong interaction of two-dimensional long internal gravity waves propagating on neighboring pycnoclines in a stratified fluid. In this paper we consider a special case of (1.1), namely the following system of nonlinear evolution equations:
(1.2)
We always look for solutions of (1.2) of the form
(1.3)
where
. Through calculating, if
we get
(1.4)
Now we consider it in higher dimensional cases, as follows:
(1.5)
where
,
and
,
is a coupling parameter.
The system of (1.1) has been analysed many times. For example, see the recently derived model by Gear and Grimshaw [1], considering
(1.6)
where
are constants. Moreover, the system (1.2) has been extensively studied in recent years and is also a special case of a general class of nonlinear evolution equations considered in [2] in the inverse scattering context. More properties of the system (1.2) have been proved. Alarcon and Montenegro proved the local and global well-posedness results for the initial-value problem for (1.2) with
in [3] and [4]. Panthee and Scialom improved the well-posedness results obtained in the case when
in [2].
As we know, many analyses about higher order equation have been done many years ago including the third and fifth order KdV equation. Firstly, it is already well known that the third order KdV equation describes the evolution of weakly nonlinear and weakly dispersive shallow waves in physical contexts such as plasma, ion-acoustic waves, stratified internal, and atmospheric waves and it has been analysed during the last decades. For the fifth order equation, the results are less than the third. But it has attracted increasing attentions (see [5] - [22] ) and is used to model many physical phenomena such as gravity-capillary waves on a shallow layer and magnetosome propagation in plasmas. For example, Baker took the work and published in 1903; Li Xiaofeng proved the existence of solitary wave solutions of fifth order KdV equations in recent years. Santosh Bhattarai proved the existence of travelling-wave solutions of coupled KdV equations when it loses the compactness, using the method of concentrate compactness principle of Lions in 2015.
We know that the system of higher order equations is rare. We only can find other similar fourth-order systems studying the interaction of the long and short waves have appeared. P. Lvarez-Caudevilla and E. Colorado researched the coupled nonlinear Schrodinger Equations (1.7) and the system of Schrodinger and Korteweg-de Vries Equations (1.8).
(1.7)
and
(1.8)
They proved the existence of equations using the variation approach and minimization techniques on Nehari manifold and the multiplicity of the equations by fibering map.
But we know that there is not previous mathematics work analyzing a higher order system as (1.3) and we get the multiplicity of the equations by a bifurcation theory which is not founded in other higher order equations article.
We organize the paper as follows. In Section 2, we introduce the notation, establish the functional framework, define the Nehari manifold and give the main theorem. In Section 3, we construct semi-trivial solutions and show the properties depending on the coupling parameter. Moreover, we devoted to proving the main results of the paper by the variation principle and mountain-pass theorem. In Section 4, using the Crandall-Rabinowitz local bifurcation theory, we show the multiplicity of the ground state solutions.
2. Preliminaries and Main Theorems
In
, we define the following equivalent norm and scalar product:
(2.1)
Accordingly, the inner product and induced norm on
are given by
(2.2)
We define
the radially symmetric functions in
and
. In addition, we define the energy functional associated with system (1.5) by
(2.3)
and
are the energy functionals of the uncoupled equations. Then, we define
(2.4)
Now, we restrict the Nehari Manifold to the setting, denoting it as
Remark 2.1. (see [23] [24] [25] )
Let
Then we have the following Sobolev embedding:
Proposition 2.1. We are going to prove some properties for
and
on
.
1)
is a locally smooth manifold.
2)
is a complete metric space.
3)
is a critical point of
if and only if
is a critical point of
constrained on
.
4)
is bounded from below on
.
Proof. 1) Differentiating expression (2.4) yields
(2.5)
and because of
, we have the fact that
.
Then, we obtain
(2.6)
Then,
is a locally smooth manifold near any point
with
.
2) Let
be a sequence such that
as
. By the embedding theorem, we have
and
for
. It is clear that
(2.7)
Since
and
, applying Holder's inequality, we get
(2.8)
So we have
. Because of
, we get
. Using
and
, we get
. Hence
and
is a complete metric space.
Taking the derivative of the functional
in the direction
, we find
The second derivative of
is given by
So, we have
which is positive definite so that
is a strict minimum critical point of
. As a consequence, we have that
is a smooth complete manifold, and there exists a constant
such that
(2.9)
3) Assume that
is a critical point of
and with
, then there is a Lagrange multiplier
such that
(2.10)
Apply both sides to
and we can get
(2.11)
Combining (2.6) and (2.11), we get
. Now (2.10) gives
, i.e.
, is a critical point of
.
4) The functional constrained on
takes the form combining (2.3) and (2.4)
(2.12)
using (2.9) and (2.12), we can get
(2.13)
So,
is bounded from below on
. □
Lemma 2.1. For every
, there is a number
such that
.
Proof. For
and
, we have
On the one hand, we have
and
for a small enough t. On the other hand, we have
as
. So there is a maximum point
of t. Moreover we get
and deduce
. □
Lemma 2.2. Assume that
, then
satisfies the PS condition constrained on
.
Proof. Suppose
is a sequence i.e.
(2.14)
From (2.4) and (2.9), we can get
is bounded, then we have a weakly convergent subsequence
(for convenience denoted again by
). Since H is compactly embedded into
for
, we infer that
Moreover, using the fact that
and (2.3), we have
which implies that
. Letting
(2.15)
denotes the constrained gradient of
on
with
. Taking the scalar product with
and with
we can get
then, taking into account (2.6) and (2.7), we deduce that
as
. We also have that
is bounded, so with (2.13) and the fact
as
, we obtain
(2.16)
So we deduce that
To finish the proof, since
as
, it follows that
strongly. □
Theorem 2.1. Suppose
. The infimum of
on
is attained at some
with
and both components
.
Theorem 2.2.
1)Let
be the principal eigenvalue of
and let
be the corresponding positive eigenfunction. Then there exists
such that when
, (1.3) has solutions
of the form
and
2). There exists
such that when
, (1.3) has solutions
of the form
3. Existence Results of Semi-Trivial Solutions and Non-Trivial Solutions
System (1.5) admits two kinds of semi-trivial solutions of the form
and
. So we take
and
, where
and
are radially symmetric ground state solutions of the equation
in
. Moreover, if we denote w a radially symmetric ground state solution of (3.1)
(3.1)
then, by scaling, we can get
(3.2)
Hence, system (1.5) has two kinds of semi-trivial solutions
and
with lowest energy among all the semi-trivial solutions.
Definition 3.1.
1) We define new Nehari manifold corresponding to the equations of (1.5)by
where
Let us define the tangent space to
on
and
by
And define the tangent space to
on
and
on
by
We can prove the following equivalence:
(3.3)
If we denote by
the second derivative of
constrained on
,using that
is a critical point of
,plainly we obtain that
2) We define the following Sobolev constants related to
and
,
(3.4)
and
Proposition 3.1.
1) If
then
is a strict local minimum of
constrained on
.
2) If either
,then
is a saddle point of
constrained on
. Moreover
(3.5)
Proof. 1) Suppose
.
For
one has that
For one thing, since
and the definition of
, there exists
such that
(3.6)
For another thing, using (3.3) and the fact that
is a minimum of
on
, there exists a constant
such that
(3.7)
Hence, using (3.6) and (3.7) we get
(3.8)
proving that
is a strict local minimum of
on
.
When
, we can obtain the same result by using the same argument as above.
2) Assume
In this case, we choose an element
, such that
Now, taking
we get
And taking
we get
Hence,
is a saddle point of
on
.
When
we can obtain the same result using the same argument as above and obviously inequality (3.5) holds. □
Next, we will give the proof of Theorem 2.1 and Theorem 2.2.
Proof. By the Ekelands variational principle [26], there exists a minimizing sequence
, i.e.,
Due to the Lemma 2.2, there exists
such that
so,
is a minimum point of
on
. □
We have
. Then there exists
such that
. So we get
Combining
we get
. According to the definition of
,
with
, we get
is a nonnegative ground state solution of the system. We can conclude that both components of
are non-trivial. In fact, if the second component
, then
. So
is the non-trivial solution of the system (1.5), Hence, we have
However, this is a contradiction due to the fact that
is a radial ground state solution of
. We conclude the first component
using the same way. Lastly, taking into account Proposition 3.1-(2) and
we have
(3.9)
4. Bifurcation of Nontrivial Solutions
In this subsection, we prove the existence of nontrivial solution of (1.5) by using local bifurcation theory (see [27] ). The main results follow.
Proof. Consider the eigenvalue problem
(4.1)
It is well known that this problem admits a sequence of eigenvalues
(4.2)
Moreover, we infer from [28] that the first eigenvalue
is simple and the principle eigenfunction
is a positive function. Set
, We shall consider the bifurcation of nontrivial solution of (1.5) from the semitrivial branch
near
. To accomplish this, we apply the bifurcation results of Crandall and Rabinowitz. First, we define F by
(4.3)
Clearly, for
, one sees that
(4.4)
We define
(4.5)
From (4.1) and (4.2), we get that the null space
. The solution space of
in
is
. Hence, the null space
is trivial. So the null space
, and
is the principal eigenfunction of (4.1). The range space of L is defined by
(4.6)
Thus,
. Since
, it follows from (5.6) that
Thus, we can apply the result of [27] to conclude that the set of positive solutions to (1.5) near
is a smooth curve
(4.7)
such that
,
,
, where
is a small constant. Moreover,
can be calculated as (see, for example, [29] [30] )
(4.8)
where
is a linear functional on
defined as
. Hence, we infer from (4.7) and (4.8) that for
(4.9)
and
Now, we give the proof of (2). As we know,
is the unique positive solution of (1.5) with
. Recalling the map defined in (4.3), we have
(4.10)
It is well known that
and
are both invertible; hence,
is nondegenerate in X2r, i.e.,
exists. We infer from the implicit function theorem that there exists
and
such that for any
,
. Moreover, we can compute
. Differentiating
by
at
, because of
, we get
(4.11)
so
(4.12)
This gives the expression of
.
5. Summary
In the paper, we studied the positive radial solutions for a higher order coupled system of Korteweg-de Vries equations in Theorem 2.1. Moreover, we proved the multiplicity of the equations by a bifurcation theory in Theorem 2.2.