Darboux Transformation and New Multi-Soliton Solutions of the Whitham-Broer-Kaup System ()
1. Introduction
Studying of the nonlinear models in shallow water wave is very important, such as Korteweg-de Vries (KdV) equation [1] [2] , Kadomtsev-Petviashvili (KP) equation [3] [4] , Boussinesq equation [5] [6] , etc. There are many methods to study these nonlinear models, such as the inverse scattering transformation [7] , the Bäcklund transformation (BT) [8] , the Hirota bilinear method [9] , the Darboux transformation (DT) [10] , and so on. Among those various approaches, the DT is a useful method to get explicit solutions.
In this paper, we investigate the Whitham-Broer-Kaup (WBK) system [11] -[13] for the dispersive long water in the shallow water
(1)
where
is the field of the horizontal velocity, and
is the height that deviates from equilibrium position of the liquid. The constants
and
represent different diffusion powers. If
and
, the WBK system (1) reduces to the classical long-wave system that describes the shallow water wave with diffusion [14] . If
and
, the WBK system (1) becomes the modified Boussinesq-Burgers equation [7] .
Many solutions have been obtained for the WBK system (1), such as the analytical solution, the soliton-like solution, the soliton solutions, the periodic solution, the rational solution, and so on [15] -[19] .
In this paper, through a proper transformation
(2)
the WBK system (1) is transformed into the parameter Levi system
(3)
Based on the obtained Lax pair, we construct the N-fold DT of the parameter Levi system (3) and then get the N-fold DT of the WBK system (1). Resorting to the obtained DT, we get new multi-soliton solutions of the WBK system.
The paper is organized as follows. In Section 2, we construct the N-fold DT of the Levi system and the WBK system. In Section 3, DT will be applied to generate explicit solutions of the WBK system (1).
2. Darboux Transformation
In this section, we first construct the N-fold DT of the parameter Levi system, and then get explicit solutions of the WBK system.
We consider the following spectral problem corresponding to the Levi system (3)
(4)
and its auxiliary problem
(5)
where
is a spectral parameter and
. The compatibility condition
yields a zero curvature equation
which leads to the Levi system (3) by a direct computation.
Now we introduce a transformation of (4) and (5)
, (6)
where
is defined by
(7)
Then the Lax pair (4) and (5) are transformed into
(8)
(9)
where
,
have the same form as
,
, except replacing
,
,
,
with
,
,
,
, respectively.
In order to make the Lax pair (4) and (5) invariant under the transformation (6), it is necessary to find a matrix
.
Let the matrix
in (6) be in the form of
(10)
with
![]()
where
are functions of
and
.
Let
,
be two basic solutions of the spectral problem (4) and use them to define a linear algebraic system
(11)
with
(12)
where the constants
,
are suitably chosen such that the determinant of the coefficients of (11) are nonzero. If we take
(13)
then
are uniquely determined by (11).
From (10), we have
(14)
We note that (11) can be written as a linear algebraic system
(15)
and
,
which implies that
are
roots of
, that is
, (16)
where
is independent of
. From the above facts, we can prove the following propositions.
Proposition 1. Let
satisfy the following first-order differential equation
(17)
Then the matrix
determined by Equation (7) is the same form as
:
,
where the transformations from the old potentials
,
to
,
are given by
(18)
Proof: Let
and
, (19)
where
denotes the adjoint matrix of
. It is easy to see that
and
are
th-order polynomials in
, while
,
are (2N − 1)th-order polynomials in
. From (4) and (12), we get
(20)
By using (16) and (20), we can prove that
are the roots of
. From (15), we have
![]()
Hence, together with (19), we have
(21)
that is
(22)
with
![]()
where
are independent of
. By comparing the coefficients of
,
and
in (22), we find
(23)
(24)
(25)
(26)
Substituting (17) into (24)-(26) yields
(27)
From (7) and (22), we find that
. The proof is completed. ![]()
Remark. When
, assuming that
, the DT can be rewritten as
(28)
Let the basic solution
,
of (4) satisfy (5) as well. Through a similar way as Proposition 1, we can prove that
has the same form as
under the transformation (6) and (18). We get the following proposition.
Proposition 2. Suppose
satisfy the following equation
(29)
Then the matrix
defined by (9) has the same form as
, that is
![]()
where
and
are given by (18).
The proof of Proposition 2 is similar with Proposition 1, but it is much more tedious and then we omit the proof for brevity. For the similar proof we can also refer to [20] [21] .
According to Proposition 1 and 2, the Lax pair (4) and (5) is transformed into the Lax pair (8) and (9), then the transformation (6) and (18):
is called the DT of the Lax pair (4) and (5). The Lax pair leads to the parameter Levi system (3) and then the transformation (6) and (18):
is also called DT of the parameter Levi system (3). On the other hand, together with the transformation (2), the parameter Levi system (3) is transformed into the WBK system (1), then we get the solutions of the WBK system (1).
Theorem 1. If
is a solution of the parameter Levi system (3),
with
(30)
is another solution of the parameter Levi system (3), where
,
,
,
are determined by (11) and (13).
From the transformation (2), we find that
Theorem 2. If
is a solution of the WBK system (1),
with
(31)
is another solution of the WBK system (1), where
is determined by (30). Then the transformation
is also called the DT of the WBK system (1).
3. New Solutions
In this section, we take a trivial solution
as the “seed” solution, to obtain multi-soliton solutions of the WBK system (1).
Substituting
into the Lax pair (4) and (5), the two basic solutions are
(32)
with
,![]()
.
According to (12), we get
(33)
For simplicity, we discuss the following two cases, i.e.
and
.
As
, let
, solving the linear algebraic system (11) and (13), we have
(34)
according to (28), we get
(35)
Substituting (35) into (31), we obtain the solution of the WBK system (1) as
(36)
with ![]()
By choosing proper parameters (such as
,
,
,
), we find that
is a bell-type- soliton and
is a M-type-soliton.
As
, let
, together with (11) and (13), we have
(37)
with
![]()
With the help of (30), we get
![]()
(a) (b)
Figure 1. Plots of the three-soliton solution (39).
(38)
Then we get another solution of the WBK system (1) by using of (31)
(39)
with ![]()
When we take
,
,
,
,
,
,
,
,
is a three- bell-type-soliton solution with two overtaking solitons and one head-on soliton (see Figure 1(a)) and
is a three-M-type-soliton solution with two overtaking solitons and one head-on soliton (see Figure 1(b)). We note that by the obtained DT, we can get
soliton solutions which are different from those in [19] which are
-soliton solutions.
Acknowledgements
This work is supported by Nurture Funds of National Project of University of Shanghai for Science and Technology (no. 14XPQ09).