Bifurcation method and traveling wave solution to Whitham-Broer-Kaup equation
-
Upload
independent -
Category
Documents
-
view
3 -
download
0
Transcript of Bifurcation method and traveling wave solution to Whitham-Broer-Kaup equation
Applied Mathematics and Computation 171 (2005) 677–702
www.elsevier.com/locate/amc
Bifurcation method and traveling wavesolution to Whitham–Broer–Kaup equation
Jianwei Shen a,b,*, Wei Xu a, Yanfei Jin a
a Department of Applied Mathematics, Northwestern Polytechnical University, Xi’an,
Shaanxi 710072, Chinab Department of Mathematics, Xuchang University, Xuchang, Henan 461000, China
Abstract
By using bifurcation method to Whitham–Broer–Kaup shallow water equations,
bifurcation parameter sets are shown. Numbers of solitary waves, kink waves and peri-
odic waves are given. Under various parameter conditions, all explicit formulas of sol-
itary wave solutions, kink wave solutions and periodic wave solutions are listed.
� 2005 Elsevier Inc. All rights reserved.
Keywords: Solitary wave solution; Kink and anti-kink wave solution; Periodic wave solution;
WBK equations; Bifurcation theory; Dynamical systems
1. Introduction
Under Boussinesq approximation, Whitham, Broer and Kaup [8–10] ob-
tained WBK equations
ut þ uux þ vx þ buxx ¼ 0; ð1:1aÞ
0096-3003/$ - see front matter � 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.amc.2005.01.078
* Corresponding author.
E-mail addresses: [email protected] (J. Shen), [email protected] (W. Xu).
678 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
vt þ ðuvÞx þ auxxx � bvxx ¼ 0; ð1:1bÞwhere u = u(x, t) is the field of horizontal velocity, v = v(x, t) is the height thatdeviate from equilibrium position of liquid, a, b are constants that represent
different diffusion power, Eq. (1.1) is a very good model to describe dispersive
wave. For the background materials of model equation, we refer to the paper
[8–10] and the references therein. If a = 0 and b 5 0, Eq. (1.1) is classic long-
wave equations that describe shallow water wave with diffusion [12]. If a = 1
and b = 0, Eq. (1.1) is modified Boussinesq equations [13]. Kaup [8] and
Ablowitz [13] studied inverse transformation solution for the special case of
Eq. (1.1), Kupershmidt discussed their symmetries and conservation laws. Byusing of Backlund transformation, Fan [12] found three pairs of solutions of
WBK equation. Xie et al. [14] made use of hyperbolic function method and
Wu elimination method obtained four pairs of solutions of WBK equation.
Chen et al. [15] present a more general transformation and applied it to
WBK equation and obtained some traveling wave solutions. Unfortunately,
these results in [8–15] are not complete since the authors did not consider
the bifurcation behavior of phase portraits for the corresponding traveling
wave equation.In this paper, by using bifurcation theory and methods of dynamical systems
(see [1–4]), we consider bifurcations of Eq. (1.1). Under fixed parameter condi-
tions, we obtain the all explicit formulas of traveling wave solution which
including solitary wave, kink and anti-kink wave, periodic wave solutions.
To find the traveling wave solution of (1.1), we first consider the traveling
wave solutions in the form
uðx; tÞ ¼ uðnÞ; vðx; tÞ ¼ vðnÞ; n ¼ x� ct; ð1:2Þ
where c denotes the wave speed. Therefore (1.1) reduce to be
�cun þ uun þ vn þ bunn ¼ 0; ð1:3aÞ
�cvn þ ðuvÞn þ aunnn � bvnn ¼ 0: ð1:3bÞ
Integrating (1.3) once with respect to n leads to
�cuþ 1
2u2 þ vþ bun ¼ g1; ð1:4aÞ
�cvþ uvþ aunn � bvn ¼ g2: ð1:4bÞwhere g1, g2 2 R are integral constant. Derived from (1.4), we can get
v ¼ cu� 1
2u2 � bun þ g1; ð1:5Þ
vn ¼ cun � uun � bunn: ð1:6Þ
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 679
Substituting (1.5) and (1.6) into (1.4b), we get
unn ¼1
2ðaþ b2Þðu3 � 3cu2 þ 2ðg1 � c2Þuþ 2ðg1cþ g2ÞÞ: ð1:7Þ
Let dudn ¼ y. Then we have the following traveling wave system which is an pla-
nar dynamical system:
dudn
¼ y;dydn
¼ 1
2ðaþ b2Þðu3 � 3cu2 þ 2ðg1 � c2Þuþ 2ðg1cþ g2ÞÞ: ð1:8Þ
Clearly, (1.8) is a Hamiltonian system with Hamiltonian function
Hðu; yÞ ¼ 1
2y2 � 1
2ðaþ b2Þ1
4u4 � cu3 þ ðg1 � c2Þu2 þ 2ðg1cþ g2Þu
� �:
ð1:9ÞFrom (1.8), we can know that we only consider the traveling wave solutions of
(1.8).
The Phase portraits defined by the vector fields of system (1.8) determine all
traveling wave solutions of (1.1). Suppose that u(x, t) = u(x � ct) = u(n) is a
continuous solution of (1.1) for n 2 (�1, 1) and limn!1u(n) = a1 and
limn!�1u(n) = b1. It is well known that (i) u(x, t) is called a solitary wave solu-tion if a1 = b1 and (ii) u(x, t) is called a kink or anti-kink solution if a1 5 b1.
Usually, a solitary wave solution of (1.1) corresponds to a homoclinic orbit
of (1.8). A kink (or anti-kink) wave solution of (1.1) corresponds to a hetero-
clinic orbit (or so-called connecting orbit) of (1.8). Similarly, a periodic orbit of
(1.8) corresponds to a periodically traveling wave solution of (1.1). Thus, to
investigate all bifurcations of solitary wave, kink waves and periodic waves
of (1.1), we should find all periodic annuli, homoclinic and heteroclinic orbits
of (1.8) depending on the parameter space of this system. The bifurcationtheory (see [1–7]) plays an important role in our study.
This paper is organized as follows. In Section 2, we give the bifurcation set
and phase portraits of (1.8). In Section 3, we show all explicit formulae of sol-
itary wave solutions and kink (or anti-kink) wave solutions under given param-
eter conditions. In Section 4, we show all explicit formulae of periodic wave
solutions under given parameter conditions. Our study results give rise to com-
plete description of all traveling waves of (1.1) and contain some results in
[8–15] as special examples.
2. Bifurcations of phase portraits of (1.8)
In this section, we consider bifurcation set and phase portraits of (1.8).
Obviously, on the (u, y)-phase plane, the abscissas of equilibrium points of
680 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
system (1.8) are the zeros of f(u) = u3 � 3cu2 + 2(g1 � c2)u + 2(g1c + g2). Notice
that f 0(u) = 3u2 � 6cu + 2(g1 � c2). Let (u0, 0) be an equilibrium of (1.8), at this
point, the determinant of the linearized system (1.8) has the form Jðu0; 0Þ ¼� 1
2ðaþb2Þ f0ðu0Þ. By using the bifurcation theory of planar dynamical system,
we know that if J(u0, 0) > 0 (or < 0), then the equilibrium (u0, 0) is a center
(or saddle point); if J(u0, 0) = 0 and the Poincare index of (u0, 0) is zero, then
the equilibrium (u0, 0) is a cusp. By using the above facts to do qualitative anal-
ysis, we obtain the following results.
Case I. g1 = c2.
Making the transformation 1ffiffiffiffiffiffiffiffiffiffiffiffi2ðaþb2Þ
p n ¼ s, y ! 1ffiffiffiffiffiffiffiffiffiffiffiffi2ðaþb2Þ
p y for a + b2 > 0, or
1ffiffiffiffiffiffiffiffiffiffiffiffiffiffi�2ðaþb2Þ
p n ¼ s, y ! 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffi�2ðaþb2Þ
p y for a + b2 < 0, (1.8) becomes the following
two-dimensional system:
duds
¼ y;dyds
¼ �ðu3 � 3cu2 þ 2aÞ; ð2:1Þ
where a = g1c + g2, and when a + b2 > 0 (or < 0), then the sign in the right-hand side of second equation of (2.1) is ‘‘+’’ (or ‘‘�’’). System (2.1) has Ham-
iltonian function
Hðu; yÞ ¼ 1
2y2 � 1
4u4 � cu3 þ 2au
� �: ð2:2Þ
By using the above discussion to do qualitative analysis, we obtain the follow-
ing results. On the (a, c)-parametric plane, there are three bifurcation curves
(see Fig. 1):
L1 : a ¼ 0; L2 : a ¼ c3; L3 : a ¼ 2c3: ð2:3Þ
–4
–2
0
2
4
a
–4 –2 2 4c
Fig. 1. The partition of the (a, c)-parameter strip of (2.1).
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 681
From the above discussion we have the partition in (c, a)-parameter plane
by curves Li, i = 1, 2, 3, shown in Fig. 1 where
ðAÞ 0 < c < L1; ðBÞ L2 < c < L3; ðCÞ c > L3 > 0;
ðDÞ L2 < c < L1; ðEÞ L3 < c < L2; ðF Þ c < L3 < 0:
Corresponding to regions (A)–(F) of the bifurcation set of (a, c), phase por-
traits of (2.1) can be shown in Figs. 2 and 3, respectively.
Case II. g1 > c2.Making the transformation
ffiffiffiffiffiffiffiffiffig1�c2
aþb2
qn ¼ f, 1ffiffiffiffiffiffiffiffiffiffiffiffiffi
2ðg1�c2Þp u ¼ /,
ffiffiffiffiffiffiffiffiffiffiffiffiffi0:5jaþb2j
p
ðg1�c2Þ32
y ! y,
(1.8) becomes the following two-dimensional system:
d/df
¼ �y;dydf
¼ /3 � 3b/2 þ /þ d; ð2:4Þ
where
b ¼ cffiffiffi2
pðg1 � c2Þ
12
; d ¼ g1cþ g2
ð2g1 � 2c2Þ32
;
and when a + b2 > 0 (or < 0), the sign of the term y in the right-hand side of the
first equation of (2.4) is ‘‘+’’ (or ‘‘�’’). This system (2.4) is a Hamiltonian sys-tem with Hamiltonian function
Hð/; yÞ ¼ � 1
2y2 � 1
4/4 � b/3 þ 1
2/2 þ d/
� �: ð2:5Þ
Denote that f1(/) = /3 � 3b/2 + / + d, f 0(/) = 3/2 � 6b/ + 1. Thus f 0(/) has
two zeros at /1;2 ¼ 3b�ffiffiffiffiffiffiffiffiffi9b2�3
p3
. Obviously, /1,2 is real if and only if jbj P 1ffiffi3
p .
Hence, in this situation there exist three bifurcation curves (see Fig. 4):
L4 : d ¼ �2
27ð9b2 � 3Þ
32 þ 2b3 � b; L5 : d ¼ 2b3 � b;
L6 : d ¼ 2
27ð9b2 � 3Þ
32 þ 2b3 � b:
ð2:6Þ
From the above discussion we have the partition in (b, d)-parameter plane by
curves Li, i = 4, 5, 6, shown in Fig. 4 where
ðI1Þ d < L4; ðII1Þ L4 < d < L5; ðIII1Þ L5 < d < L6; ðIV 1Þ d > L6:
Corresponding to regions (I1)–(IV1) of the bifurcation set of (b, d), the bifurca-
tions of phase portraits of (2.4) can be shown in Figs. 5 and 6, respectively.
Case III. g1 < c2.
Making the transformationffiffiffiffiffiffiffiffiffic2�g1jaþb2j
qn ¼ 1, 1ffiffiffiffiffiffiffiffiffiffiffiffiffi
2ðc2�g1Þp u ¼ u,
ffiffiffiffiffiffiffiffiffiffiffiffiffi0:5jaþb2j
p
ðc2�g1Þ32
y ! y,
(1.8) becomes the following two-dimensional system:
—4
—2
0
2
4
y
—2 —1 1 2 3 4x
—4
—2
0
2
4
y
–2 —1 1 2 3 4x
–4
—2
0
2
4
y
—2 —1 1 2 3 4
x
—4
—2
0
2
4
y
—2 —1 1 2 3 4x
—4
—2
0
2
4
y
—2 —1 1 2 3 4x
—4
–2
0
2
4
y
–2.5 —2 —1 .5 —1 —0 .5 0 .5 1x
(1) (2)
(3) (4)
(5) (6)
Fig. 2. Phase portraits of (2.1) for g1 = c2 and a + b2 > 0. (1) (a, c) 2 L1, (2) (a, c) 2 (A), (3) (a, c) 2L2, (4) (a, c) 2 (B), (5) (a, c) 2 L3, (6) (a, c) 2 (C), (7) (a, c) 2 L1, (8) (a, c) 2 (D), (9) (a, c) 2 L2, (10)
(a, c) 2 E, (11) (a, c) 2 L3, (12) (a, c) 2 (F).
682 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
—4
—2
0
2
4
y
—2 —1 1 2 3 4x
—4
—2
0
2
4
y
–2 —1 1 2 3 4
x
—4
—2
0
2
4
y
—4 —3 —2 —1 1 2x
—4
—2
0
2
4
y
—4 —3 —2 —1 1 2x
—4
—2
0
2
4
y
—4 —3 —2 —1 1 2x
—4
—2
0
2
4
y
—1 1 2x
(7) (8)
(9) (10)
(11) (12)
Fig. 2 (continued)
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 683
–4
–2
2
4
y
–1 1 2 3 4
x
–3
–2
–1
0
1
2
3
y
–1 1 2 3
x
–3
–2
–1
0
1
2
3
y
–1 1 2 3
x
–4
–2
0
2
4
y
–2 –1 1 2 3
x
–2
–1
0
1
2
y
–1.5 –1 –0.5
x
–3
–2
–1
0
1
2
3
y
–1 1 2 3
x
(1) (2)
(3) (4)
(5) (6)
Fig. 3. Phase portraits of (2.1) for g1 = c2 and a + b2 < 0. (1) (a, c) 2 L1, (2) (a, c) 2 (A), (3) (a, c) 2L2, (4) (a, c) 2 (B), (5) (a, c) 2 L3, (6) (a, c) 2 (C), (7) (a, c) 2 L1, (8) (a, c) 2 (D), (9) (a, c) 2 L2, (10)
(a, c) 2 E, (11) (a, c) 2 L3, (12) (a, c) 2 (F).
684 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
–4
–2
2
4
y
–4 –3 –2 –1 1
x
–3
–2
–1
1
2
3
y
–3 –2 –1 1
x
–2
–1
1
2
y
–3 –2 –1 1
x
–3
–2
–1
1
2
3
y
–3 –2 –1 1
x
–4
–2
0
2
4
y
–3 –2 –1 1 2
x
–2
–1
0
1
2
y
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
x
(7) (8)
(9) (10)
(11) (12)
Fig. 3 (continued)
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 685
–4
–2
0
2
4
d
±4 ±2 2 4
b
Fig. 4. The partition of the (b, d)-parameter strip of (2.4) for g1 > c2.
686 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
dud1
¼ �y;dyd1
¼ u3 � 3eu2 � uþ g; ð2:7Þ
where
e ¼ cffiffiffi2
pðc2 � g1Þ
12
; g ¼ g1cþ g2ffiffiffi2
pðc2 � g1Þ
32
;
and when a + b2 > 0 (or < 0), the sign of the term y in the right-hand side of the
first equation of (2.7) is ‘‘+’’ (or ‘‘�’’). This system (2.7) is a Hamiltonian sys-
tem with Hamiltonian function
Hðu; yÞ ¼ � 1
2y2 � 1
4u4 � eu3 � 1
2u2 þ gu
� �: ð2:8Þ
On the (e, g)-parameter plane there exist three bifurcation curves (see Fig. 7):
L7 : g ¼ � 2
27ð9e2 þ 3Þ
32 þ eþ 2e3; L8 : g ¼ eþ 2e3;
L9 : g ¼ 2
27ð9e2 þ 3Þ
32 þ eþ 2e3:
ð2:9Þ
From the above discussion we have the partition in (e, g)-parameter plane
by curves Li, i = 7, 8, 9, shown in Fig. 7 where
ðI2Þ g < L7; ðII2Þ L7 < g < L8; ðIII2Þ L8 < g < L9; ðIV 2Þ g > L9:
Corresponding to regions (I2)–(IV2) of the bifurcation set of (e, g), phase por-
traits of (2.4) can be shown in Figs. 8 and 9.
From the above discussion, we have
Theorem 1. Corresponding to (1.1), all of the solitary wave, kink (or anti-kink)
wave and periodic wave solutions are smooth.
–4
–2
0
2
4
y
–1 1 2 3 4
x
–4
–2
0
2
4
y
–2 –1 1 2 3
x
–4
–2
0
2
4
y
–2 –1 1 2 3 4
x
–4
–2
0
2
4
y
–2 –1 1 2 3 4
x
–4
–2
0
2
4
y
–2 –1 1 2 3 4
x
–4
–2
0
2
4
y
–2 –1 1 2 3
x
(1) (2)
(3) (4)
(5) (6)
Fig. 5. Phase portraits of (2.4) for g1 > c2 and a + b2 > 0. (1) (b, d) 2 I1, (2) (b, d) 2 (L4), (3) (b, d) 2(II1), (4) (b, d) 2 (L5), (5) (b, d) 2 (III1), (6) (b, d) 2 L6, (7) (b, d) 2 (IV1).
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 687
–4
–2
0
2
4
y
–2 –1.5 –1 –0.5 0.5
x
(7)
Fig. 5 (continued)
688 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
3. Solitary wave and kink wave solutions determined by (1.8)
In this section we will give all explicit solitary wave solutions and kink (or
anti-kink) wave solutions of (1.8) under given parameter conditions. By using
the traveling wave system (1.8) and the Hamiltonian function (1.9) to do the
calculations and table of integral [16], we have the following results:
Case I. g1 = c2.
Denote that u1 < u2 < u3 are the roots of u3 � 3cu2 + 2a = 0,
a1i = 6ui(ui � 2c), b1i = 4(ui � c), q1i ¼ 8ðu2i � 2uic� 2c2Þ, i = 1, 2, 3, n = x � ct.
Suppose that a + b2 > 0.
(1) Corresponding to Fig. 2(2) and (10), (1.1) has one smooth solitary wave
solutionwithpeak form,whichhas the followingparametric representation:
u ¼ u1 �2a11
b11 þffiffiffiffiffiffiq11
psinhð
ffiffiffiffiffiffiffiffiffiffiffiffi0:5a11
psÞ;
s ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2ðaþ b2Þ
q n; v ¼ cu� 1
2u2 � bun þ g1:
ð3:1Þ
(2) Corresponding to Fig. 2(3) and (9), (1.1) has one kink wave and one anti-
kink wave solutions, which has the following parametric representation:
u ¼ c�ffiffiffi3
pc tanhð
ffiffiffiffiffiffiffi1:5
pcsÞ;
s ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2ðaþ b2Þ
q n; v ¼ cu� 1
2u2 � bun þ g1:
ð3:2Þ
–2
–1
0
1
2
y
0.5 1 1.5 2 2.5 3
x
–1.5
–1
–0.5
0
0.5
1
1.5
y
–0.5 0.5 1 1.5 2 2.5
x
–3
–2
–1
1
2
3
y
–1 1 2 3 4
x
–3
–2
–1
0
1
2
3
y
–1 1 2 3 4
x
–4
–2
2
4
y
–2 –1 1 2 3 4
x
–1.5
–1
–0.5
0
0.5
1
1.5
y
–1 –0.5 0.5 1 1.5 2
x
(1) (2)
(3) (4)
(5) (6)
Fig. 6. Phase portraits of (2.4) for g1 > c2 and a + b2 < 0. (1) (b, d) 2 I1, (2) (b, d) 2 (L4), (3) (b, d) 2(II1), (4) (b, d) 2 (L5), (5) (b, d) 2 (III1), (6) (b, d) 2 L6, (7) (b, d) 2 (IV1).
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 689
–2
–1
0
1
2
y
–3 –2.5 –2 –1.5 –1 –0.5
x
(7)
Fig. 6 (continued)
–4
–2
0
2
4
g
±4 ±2 2 4
e
Fig. 7. The partition of the (e, g)-parameter strip of (2.7) for g1 < c2.
690 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
(3) Corresponding to Fig. 2(4) and (8), (1.1) has one smooth solitary wave
solution with valley form, which has the following parametric
representation:
u ¼ u3 �2a13
b13 þffiffiffiffiffiffiq13
psinhð
ffiffiffiffiffiffiffiffiffiffiffiffi0:5a13
psÞ;
s ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2ðaþ b2Þ
q n; v ¼ cu� 1
2u2 � bun þ g1:
ð3:3Þ
Fig. 8. Phase portraits of (2.7) for g1 < c2 and a + b2 > 0. (1) (e, g) 2 I2, (2) (e, g) 2 (L7), (3) (e,
g) 2 (II2), (4) (e, g) 2 (L8), (5) (e, g) 2 (III2), (6) (e, g) 2 L9, (7) (e, g) 2 (IV2).
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 691
Fig. 8 (continued)
692 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
Suppose that a + b2 < 0.
(4) Corresponding to Fig. 3(1) and (7), (1.1) has one smooth solitary wavesolution with peak form, which has the following parametric
representation:
u ¼ 4c1þ 2c2s2
;
s ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�2ðaþ b2Þ
q n; v ¼ cu� 1
2u2 � baun þ g1:
ð3:4Þ
(5) Corresponding to Fig. 3(2), (4), (8) and (10), (1.1) has a pair of smooth
solitary wave solutions, which have the following parametricrepresentation:
u ¼ u2 þ2a12
�b12 þffiffiffiffiffiffiq12
psinhð
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�0:5a12
psÞ;
s ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�2ðaþ b2Þ
q n; v ¼ cu� 1
2u2 � bun þ g1:
ð3:5Þ
(6) Corresponding to Fig. 3(3) and (9), (1.1) has a pair of smooth solitary
wave solutions, which have the following parametric representation:
u ¼ cþffiffiffi6
pc sechð
ffiffiffi3
pcsÞ;
s ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�2ðaþ b2Þ
q n; v ¼ cu� 1
2u2 � bun þ g1:
ð3:6Þ
Fig. 9. Phase portraits of (2.7) for g1 < c2 and a + b2 < 0. (1) (e, g) 2 I2, (2) (e, g) 2 (L7), (3) (e, g) 2(II2), (4) (e, g) 2 (L8), (5) (e, g) 2 (III2), (6) (e, g) 2 L9, (7) (e, g) 2 (IV2).
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 693
Fig. 9 (continued)
694 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
(7) Corresponding to Fig. 3(5) and (7), (1.1) has one smooth solitary wave
solution with valley form, which has the following parametric
representation:
u ¼ 2c� 4c1þ 2c2s2
;
s ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�2ðaþ b2Þ
q n; v ¼ cu� 1
2u2 � bun þ g1:
ð3:7Þ
Case II. g1�c2 > 0.
Denote that /1 < /2 < /3 are the roots of /3 � 3b/2 + / + d = 0, n = x � ct,
a2i ¼ 6/2i � 12/ibþ 2, b2i = 4(/i � b), q2i ¼ 8ð/2
i � 2/ibþ 1� 2b2Þ, i = 1, 2, 3.
Suppose that a + b2 > 0.
(1) Corresponding to Fig. 5(3), (1.1) has one smooth solitary wave solution
with peak form, which has the following parametric representation:
/ ¼ /1 �2a21
b21 þffiffiffiffiffiffiq21
psinhð
ffiffiffiffiffiffiffiffiffiffiffiffi0:5a21
pfÞ:
f ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig1 � c2
jaþ b2j
sn; / ¼ ð2g1 � 2c2Þ�
12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:8Þ(2) Corresponding to Fig. 5(4), (1.1) has one kink and one anti-kink wave
solutions, which has the following parametric representation:
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 695
/ ¼ b�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi3b2 � 1
ptanh
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi0:5ð3b2 � 1Þ
qf
� �
f ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig1 � c2
jaþ b2j
sn; / ¼ ð2g1 � 2c2Þ�
12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:9Þ(3) Corresponding to Fig. 5(5), (1.1) has one smooth solitary wave solution
with valley form, which has the following parametric representation:
/ ¼ /3 �2a23
b23 þffiffiffiffiffiffiq23
psinhð
ffiffiffiffiffiffiffiffiffiffiffiffi0:5a23
pfÞ:
f ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig1 � c2
jaþ b2j
sn; / ¼ ð2g1 � 2c2Þ�
12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:10ÞSuppose that a + b2 < 0.
(4) Corresponding to Fig. 6(2), (1.1) has one smooth solitary wave solution
with peak form, which has the following parametric representation:
/ ¼ b� 1
3
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi9b2 � 3
pþ 4
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi9b2 � 3
p
6þ ð3b2 � 1Þf2;
f ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig1 � c2
jaþ b2j
sn; / ¼ ð2g1 � 2c2Þ�
12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:11Þ(5) Corresponding to Fig. 6(3) and (5), (1.1) has a pair of smooth solitary
wave solutions with peak form and valley form, which have the following
parametric representation:
/ ¼ /2 þ2a22
�b22 þffiffiffiffiffiffiq22
psinhð
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�0:5a22
pfÞ;
f ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig1 � c2
jaþ b2j
sn; / ¼ ð2g1 � 2c2Þ�
12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:12Þ(6) Corresponding to Fig. 6(4), (1.1) has a pair of smooth solitary wave solu-
tions with peak form and valley form, which have the following paramet-
ric representation:
696 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
/ ¼ b�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi6b2 � 2
psech
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi3b2 � 1
pf
� �;
f ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig1 � c2
jaþ b2j
sn; / ¼ ð2g1 � 2c2Þ�
12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:13Þ(7) Corresponding to Fig. 6(6), (1.1) has one smooth solitary wave solution
with valley form, which has the following parametric representation:
/ ¼ 1
3
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi9b2 � 3
pþ b� 4
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi9b2 � 3
p
3þ 2ð3b2 � 1Þf2;
f ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig1 � c2
jaþ b2j
sn; / ¼ ð2g1 � 2c2Þ�
12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:14ÞCase III. g1 � c2 < 0.
Denote that u1 < u2 < u3 are the roots of u3 � 3eu2 � u + g = 0, n = x � ct,
a3i ¼ 2ð3u2i � 6uie� 1Þ, b3i = 4(ui � e), q3i ¼ 8ðu2
i � 2uie� 2e2 � 1Þ, i = 1,
2, 3.
Suppose that a + b2 > 0.
(1) Corresponding to Fig. 8(3), (1.1) has one smooth solitary wave solution
with peak form, which has the following parametric representation:
u ¼ u1 �2a31
b31 þffiffiffiffiffiffiq31
psinhð
ffiffiffiffiffiffiffiffiffiffiffiffi0:5a31
p1Þ;
1 ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffic2 � g1jaþ b2j
sn; u ¼ ð2c2 � 2g1Þ
�12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:15Þ(2) Corresponding to Fig. 8(4), (1.1) has one kink and one anti-kink wave
solution, which has the following parametric representation:
u ¼ e�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi3e2 þ 1
ptanh
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi6e2 þ 2
p
21
� �;
1 ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffic2 � g1jaþ b2j
sn; u ¼ ð2c2 � 2g1Þ
�12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:16Þ(3) Corresponding to Fig. 8(3), (1.1) has one smooth solitary wave solution
with valley form, which has the following parametric representation:
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 697
u ¼ u3 �2a33
b33 þffiffiffiffiffiffiq33
psinhð
ffiffiffiffiffiffiffiffiffiffiffiffi0:5a33
p1Þ;
1 ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffic2 � g1jaþ b2j
sn; u ¼ ð2c2 � 2g1Þ
�12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:17Þ
Suppose that a + b2 < 0.
(4) Corresponding to Fig. 9(2), (1.1) has one smooth solitary wave solutionwith peak form, which has the following parametric representation:
u ¼ e� 1
3
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi9e2 þ 3
pþ 4
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi9e2 þ 3
p
3þ 2ð3e2 þ 1Þ12 ;
1 ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffic2 � g1jaþ b2j
sn; u ¼ ð2c2 � 2g1Þ
�12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:18Þ(5) Corresponding to Fig. 9(3), (1.1) has a pair of smooth solitary wave solu-
tion with peak form and valley form, which have the following paramet-
ric representation:
u ¼ u2 þ2a32
�b32 þffiffiffiffiffiffiq32
psinhð
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�0:5a32
p1Þ;
1 ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffic2 � g1jaþ b2j
sn; u ¼ ð2c2 � 2g1Þ
�12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:19Þ(6) Corresponding to Fig. 9(4), (1.1) has a pair of smooth solitary wave solu-
tion with peak form and valley form, which have the following paramet-
ric representation:
u ¼ eþffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi6e2 þ 2
psech
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi3e2 þ 1
p1
� �;
1 ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffic2 � g1jaþ b2j
sn; u ¼ ð2c2 � 2g1Þ
�12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:20Þ(7) Corresponding to Fig. 9(6), (1.1) has one smooth solitary wave solution
with valley form, which has the following parametric representation:
698 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
u ¼ eþ 1
3
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi9e2 þ 3
p� 4
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi9e2 þ 3
p
3þ 2ð3e2 þ 1Þ12 ;
1 ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffic2 � g1jaþ b2j
sn; u ¼ ð2c2 � 2g1Þ
�12u; v ¼ cu� 1
2u2 � bun þ g1:
ð3:21Þ
Theorem 2. There are 21 parametric conditions of (1.1) such that (1.1) has 21
formulae of smooth solitary wave solutions, kink and anti-kink wave solutions
given by (3.1)–(3.21).
4. Periodic solutions determined by (1.8)
In this section, all the periodic wave solutions of (1.1) will be given. Denote
that sn(u, k), cn(u, k), tn(u, k) are the Jacobian elliptic functions with themodulus k, K(k) is the complete elliptic integral of the first kind and E(k) is
the complete elliptic integral of the second kind [17].
Case I. g1 = c2.Suppose that a + b2 > 0.
Denote that u11 < u12 < u13 < u14 are the roots of 2h + 0.5u4 � 2cu3 +
4au = 0, h is Hamiltonian value, n = x � ct.
Corresponding to Fig. 2(2)–(4) and (8)–(10), Eq. (1.1) has one family of
smooth periodic wave solutions, which have the following parametric
representation:
u ¼ u14 � u13w1 sn2ðX1s; k1Þ
1� w1 sn2ðX1s; k1Þ; s ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2jaþ b2jq n;
v ¼ cu� 1
2u2 � bun þ g1;
ð4:1Þ
where
g11 ¼2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðu14� u12Þðu13 � u11Þp ; k21 ¼
ðu13� u12Þðu14 � u11Þðu14� u12Þðu13 � u11Þ
; w1 ¼u14 � u11u13 � u11
;
X1 ¼ffiffiffi2
p
2g11:
ð4:2Þ
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 699
Suppose that a + b2 < 0.
Denote that u21 < u22 < u23 < u24 are the roots of 2h � 0.5u4 + 2cu3 � 4au =
0, h is Hamiltonian value, n = x � ct.
(1) Corresponding to Fig. 3(6) and (12), Eq. (1.1) has one family of smoothperiodic wave solutions.
(2) Corresponding to Fig. 3(1), (5), (7) and (11), Eq. (1.1) has two families of
smooth periodic wave solutions.
(3) Corresponding to Fig. 3(2)–(5) and (8)–(10), Eq. (1.1) has three families
of smooth periodic wave solutions.
These periodic wave solutions have the following parametric representation:
u ¼ u24 � w2u21 sn2ðX2s; k2Þ1þ w2 sn2ðX2s; k2Þ
; s ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2jaþ b2j
q n;
v ¼ cu� 1
2u2 � bun þ g1;
ð4:3Þ
where
g2 ¼2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðu24 � u22Þðu23 � u21Þp ; k22 ¼
ðu24 � u23Þðu22 � u21Þðu24 � u22Þðu23 � u21Þ
;
w2 ¼u24 � u23u23 � u21
; X2 ¼ffiffiffi2
p
2g2:
Case II. g1 > c2.Suppose that a + b2 > 0.
Denote that /11 < /12 < /13 < /14 are the roots of 2hþ 12/4 � 2b/3 þ /2 þ
2d/ ¼ 0, h is Hamiltonian value, n = x � ct.Corresponding to Fig. 5(3)–(5), Eq. (1.1) has one family of smooth periodic
wave solutions, which have the following parametric representation:
/ ¼ /14 � /13w3 sn2ðX3f; k3Þ
1� w3 sn2ðX3f; k3Þ; f ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig1 � c2
aþ b2
sn;
/ ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2ðg1 � c2Þ
p u; v ¼ cu� 1
2u2 � bun þ g1;
ð4:4Þ
where
g3 ¼2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð/14 � /12Þð/13 � /11Þp ; k23 ¼
ð/13 � /12Þð/14 � /11Þð/14 � /12Þð/13 � /11Þ
;
w3 ¼/14 � /11
/13 � /11
; X3 ¼ffiffiffi2
p
2g3:
700 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
Suppose that a + b2 < 0.
Denote that /21 < /22 < /23 < /24 are the roots of �2h� 12/4 þ 2b/3 � /2�
2d/ ¼ 0, h is Hamiltonian value, n = x � ct.
(1) Corresponding to Fig. 6(1) and (7), Eq. (1.1) has one family of smoothperiodic wave solutions.
(2) Corresponding to Fig. 6(2) and (6), Eq. (1.1) has two families of smooth
periodic wave solutions.
(3) Corresponding to Fig. 6(3)–(5), Eq. (1.1) has three families of smooth
periodic wave solutions.
These periodic wave solutions have the following parametric representation:
/ ¼ /24 � w4/21 sn2ðX2f; k4Þ
1þ w4 sn2ðX2f; k4Þ; f ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig1 � c2
jaþ b2j
sn;
/ ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2g1 � 2c2
p u; v ¼ cu� 1
2u2 � bun þ g1;
ð4:5Þ
where
g4 ¼2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð/24 � /22Þð/23 � /21Þp ; k24 ¼
ð/24 � /23Þð/22 � /21Þð/24 � /22Þð/23 � /21Þ
;
w4 ¼/24 � /23
/23 � /21
; X4 ¼ffiffiffi2
p
2g4:
Case III. g1 < c2.Suppose that a + b2 > 0.
Denote that u11 < u12 < u13 < u14 are the roots of 2hþ 12u4 � 2eu3 � u2þ
2gu ¼ 0, h is Hamiltonian value, n = x � ct.
Corresponding to Fig. 8(3)–(5), Eq. (1.1) has one family of smooth periodic
wave solutions, which have the following parametric representation:
u ¼ u14 � u13w5 sn2ðX51; k5Þ
1� w5 sn2ðX51; k5Þ; 1 ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffic2 � g1aþ b2
sn;
u ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2ðc2 � g1Þ
p u; v ¼ cu� 1
2u2 � bun þ g1;
ð4:6Þ
where
g5 ¼2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðu14 � u12Þðu13 � u11Þp ; k25 ¼
ðu13 � u12Þðu14 � u11Þðu14 � u12Þðu13 � u11Þ
;
w5 ¼u14 � u11
u13 � u11
; X5 ¼ffiffiffi2
p
2g5:
J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702 701
Suppose that a + b2 < 0.
Denote that u21 < u22 < u23 < u24 are the roots of �2h� 12u4 þ 2e/3 þ u2�
2gu ¼ 0, h is Hamiltonian value, n = x � ct.
(1) Corresponding to Fig. 9(1) and (7), Eq. (1.1) has one family of smoothperiodic wave solutions.
(2) Corresponding to Fig. 9(2) and (6), Eq. (1.1) has two families of smooth
periodic wave solutions.
(3) Corresponding to Fig. 9(3)–(5), Eq. (1.1) has three families of smooth
periodic wave solutions.
These periodic wave solutions have the following parametric representation:
u ¼ u24 � w6u21 sn2ðX61; k6Þ
1þ w6 sn2ðX61; k6Þ; 1 ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffic2 � g1jaþ b2j
sn;
u ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2ðg1 � c2Þ
p u; v ¼ cu� 1
2u2 � bun þ g1;
ð4:7Þ
where
g6 ¼2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðu24 � u22Þðu23 � u21Þp ; k26 ¼
ðu24 � u23Þðu22 � u21Þðu24 � u22Þðu23 � u21Þ
;
w6 ¼u24 � u23
u23 � u21
; X6 ¼ffiffiffi2
p
2g6:
From the above discussion, we have
Theorem 3. There are six different kind of periodic wave solutions given by
(4.1)–(4.6) for (1.1). When the modulus ki ! 0 or ki ! 1, the periodic wave
solutions approach to solitary wave solutions or kink (anti-kink) wave solutions.
5. Summary and conclusion
In summary, we have derived all solitary, kink (or anti-kink) and periodic
wave solutions. Obviously, this method is very valid to find explicit solutions
for the integrable system and also applied to other nonlinear evolution equa-tions in mathematical physics.
Acknowledgment
This work was supported by the National Natural Science Foundation of
China (Grant no. 10472091 and 10332030) and Natural Science Foundation
702 J. Shen et al. / Appl. Math. Comput. 171 (2005) 677–702
of Shaanxi Province. J. Shen is supported partially by the Doctorate Creation
Foundation and Graduate Starting Seed Found of NWPU (Grant no. CX
200423).
References
[1] S.N. Chow, J.K. Hale, Method of Bifurcation Theory, Springer-Verlag, New York, 1981.
[2] L. Perko, Differential Equations and Dynamical Systems, Springer-Verlag, New York, 1991.
[3] J. Guckenheimer, P.J. Holmes, Nonlinear Oscillations, Dynamical Systems and Bifurcations
of Vector Fields, Springer-Verlag, New York, 1983.
[4] D. Luo et al., Bifurcation Theory and Methods Dynamical Systems, World Scientific
Publishing, London, 1997.
[5] J.B. Li, J.W. Shen, Chaos, Solitons and Fractals 20 (2004) 827.
[6] J.W. Shen et al., Appl. Math. Comput. 161 (2) (2005) 365.
[7] J.B. Li, Z.R. Liu, Appl. Math. Modell. 25 (2000) 41.
[8] D.J. Kaup, Prog. Theor. Phys. 54 (2) (1975) 396.
[9] G.B Whitham, Pro. R. Soc. A 299 (1967) 6.
[10] L.J. Broer, Appl. Sci. Res. 31 (5) (1975) 377.
[11] B.A. Kupershmidt, Commun. Math. Phys. 99 (1) (1985) 51.
[12] E.G. Fan, H.Q. Zhang, Appl. Math. Mech. 19 (8) (1998) 667 (in Chinese).
[13] M.J Ablowitz, Soliton, Nolinear Evolution Equations and Inverse Scatting, Cambridge
University Press, New York, 1991.
[14] F.D. Xie et al., Phys. Lett. A 285 (2001) 76.
[15] Y. Chen et al., Chaos, Solitons and Fractals 22 (2004) 675.
[16] B.O. Peirce, Table of Integral, Shanghai Academic, 1953 (in Chinese).
[17] P.F. Byrd, M.D. Friedman, Handbook of Elliptic Integrals for Engineer and Scientists,
Springer-Verlag, New York, 1971.