Applied Mathematics and Computation 182 (2006) 412–424

www.elsevier.com/locate/amc

Peakons, kinks, compactons and solitary patterns solutions

for a family of Camassa–Holm equations by using
new hyperbolic schemes
Abdul-Majid Wazwaz
Department of Mathematics and Computer Science, 3700 W. 103rd Street, Saint Xavier University,
Chicago, IL 60655, United States

Abstract

In this paper, a family of Camassa–Holm equations with distinct parameters is investigated. New solitary wave solu-
tions that include peakons, kinks, compactons, solitary patterns solutions, and plane periodic solutions are formally
derived. New schemes that rest mainly on hyperbolic functions are employed to achieve our goal. The work highlights
the qualitative change in the physical structures of the obtained solutions.
Ó 2006 Elsevier Inc. All rights reserved.

Keywords: Camassa–Holm family of equations; Peakons; Solitons; Compactons; Kinks

1. Introduction

Camassa and Holm [1] derived a completely integrable wave equation (CH)
ut þ 2kux  uxxt þ auux ¼ 2ux uxx þ uuxxx ð1Þ
by retaining two terms that are usually neglected in the small amplitude, shallow water limit. The constant k is
related to the critical shallow water wave speed. Eq. (1) can be derived as an asymptotic model for long gravity
waves at the surface of shallow water [1–15]. The CH equation, being a model equation for water waves, has
its integrable bi-Hamiltonian structure.
For k = 0, a = 3, it has been shown [1–4] that the CH equation (1) has peaked solitary wave solutions of the
form
uðx; tÞ ¼ ceðjxctjÞ ; ð2Þ
where c is the wave speed. The name ‘‘peakons’’, that is, solitary waves with slope discontinuities, was used to
single them from general solitary wave solutions since they have a corner at the peak of height c. However, for
k 5 0, a 5 0, Qian and Tang [12] investigated the CH equation and obtained two peakons of the form

E-mail address: wazwaz@sxu.edu

0096-3003/$ - see front matter Ó 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.amc.2006.04.002
 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424 413

 rffiffiffi 
6k a 6kt 
uðx; tÞ ¼ exp  x ; a 6¼ 3 ð3Þ
3a 3 3  a
and
  rffiffiffi  
2k a 2kt 
uðx; tÞ ¼ 3 exp  x 2 : ð4Þ
1þa 3 1 þ a
The last peaked solitary wave (4) works for every a, a > 0.
Nonlinear partial differential equations with dispersion and dissipation effects, that arise in scientific appli-
cations, have been under huge size of investigations. Many powerful methods, such as Bäcklund transforma-
tion, inverse scattering method, Hirota bilinear forms, pseudo spectral method, the tanh–sech method, the
sine–cosine method [15–20], and many others were successfully used to investigate these types of equations.
Practically, there is no unified method that can be used to handle all types of nonlinearity.
It is the objective of this work to further complement previous studies to make a further progress in this
field. In this work we will study a family of Camassa–Holm equations of the form
ut  uxxt þ aux þ buux ¼ kux uxx þ uuxxx ð5Þ

and the nonlinearly dispersive integrable equation

3
ut  k 1 uxxt þ a1 ux þ uux ¼ 2ux þ b1 uxxx þ uuxxx ; ð6Þ
k1
where a, b, k, a1, b1, and k1 are constants, and u(x, t) is the unknown function depending on temporal variable
t and spatial variable x [1–6]. It is obvious that these two equations contain both linear dispersion terms uxxt
and uxxx, and the nonlinear dispersion term uuxxx. Eq. (6) appeared first in the works of Fuchssteiner [7], but
Camassa and Holm [1] rederived it for certain values of the coefficients. Moreover, this equation can be re-
garded as an integrable perturbation of the well-known BBM equation [5,6]
ut þ ux þ uux  uxxt ¼ 0: ð7Þ

Several forms of the CH equation (5) have been investigated in the literature for several values of the con-
stants. For b = 3, k = 2, Eq. (5) reduces to the completely integrable wave Camassa–Holm (CH) equation
ut  uxxt þ aux þ 3uux ¼ 2ux uxx þ uuxxx ; ð8Þ

that describes the unidirectional propagation of shallow water waves over a flat bottom and possesses peakons
solutions if a = 0. As stated before, the CH equation is bi-Hamiltonian and has an infinite number of conser-
vation laws.
For b = 4, k = 3, Eq. (5) becomes the second completely integrable Degasperis–Procesi (DP) equation [9]
ut  uxxt þ aux þ 4uux ¼ 3ux uxx þ uuxxx ; ð9Þ
that also possesses peakons solutions if a = 0.
However, for a = 1, b = 1, k = 3, Eq. (5) is the Fornberg–Whitham (FW) equation
ut  uxxt þ ux þ uux ¼ 3ux uxx þ uuxxx ; ð10Þ

that appeared in the study of the qualitative behaviors of wave-breaking [2]. This equation admits a peaked
solution of the form

uðx; tÞ ¼ Aeð2jx3tjÞ :
1 4
ð11Þ

Moreover, for b = 3, k = 2, and a is replaced by 2a we obtain the Fuchssteiner–Fokas–Camassa–Holm

(FFCH) equation

ut  uxxt þ 2aux þ 3uux ¼ 2ux uxx þ uuxxx ð12Þ

414 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424

first arose in the work of Fuchssteiner and Fokas using a bi-Hamiltonian approach [2]. Eq. (12) has been rede-
rived by Camassa and Holm from physical considerations as a model for dispersive shallow water waves [2].
In solitary wave theory, travelling wave solutions of particular interest appear in many forms. Solitons,
which are localized travelling waves with infinite support asymptotically zero at large distances form the main
type of travelling wave solutions. Kink waves, which rise or descend from one asymptotic state to another,
form another type. A third type is the peakons that are peaked solitary wave solutions. In this case, the trav-
elling wave solutions are smooth except for a peak at a corner of its crest. Peakons are the points at which
spatial derivative changes sign [3] so that peakons have a finite jump in first derivative of the solution
u(x, t) [4]. Cuspons are other forms of solitons where solution exhibits cusps at their crests. Unlike peakons
where the derivatives at the peak differ only by a sign, the derivatives showed the jump of a cuspon diverges
[4]. A new type was developed recently, termed compactons which are solitons with compact support where
each compacton is a soliton confined to a finite core without exponential wings. Most recently, Rosenau
and Pikovsky [8] introduced a new type of travelling wave, termed kovaton, which is a robust compacton with
a flat top made of two compact kinks glued together. The discrete kovatons are hat-shaped, and kovatons
collide elastically as compactons.
The appearance of nonanalytic solitary wave solutions, including peakons, cuspons, compactons and kova-
tons, has increased the menagerie of solutions appearing in model equations, both integrable and nonintegra-
ble [5,6]. All nonlinear equations that give rise to nonanalytic solutions have nonlinear dispersion as a
distinguished feature.
We aim in this work to formally derive new sets of solitons solutions, in addition to the peakon solutions
obtained before, for the family of Camassa–Holm equations (22). Many strategies will be pursued to achieve
our goal. The work rests mainly on new ansatze that use one hyperbolic function or combine two hyperbolic
functions. The proposed schemes will be used to determine distinct solutions of distinct physical structures.
In what follows, we highlight the main features of the proposed methods briefly, because details can be
found in [15–20]. The power of the methods, that will be used, is its ease of use to determine shock or solitary
type of solutions.

2. The methods

In this section we present four ansatze, the sinh–cosh ansatze I and II, the rational tanh and coth ansatze,
and the exponential peakon algorithm to handle the family of Camassa–Holm equations (22). These schemes
can be used directly in a straightforward manner to determine the unknown parameters involved in each ansatz.

2.1. A sinh–cosh ansatz I

In this case we use a linear combination of cosh and sinh ansatz of the form
uðx; tÞ ¼ k þ acosh2 ½lðx  ctÞ þ bsinh2 ½lðx  ctÞ; ð13Þ
where k, l, a and b are parameters that will be determined. The proposed ansatz can be applied directly to the
given equation to determine these parameters.

2.1.1. A cosh ansatz I or sinh ansatz I

A cosh ansatz I reads

uðx; tÞ ¼ k þ acosh2 ½lðx  ctÞ; ð14Þ

where k, l and a are parameters that will be determined.

Similarly we introduce a sinh ansatz I of the form

uðx; tÞ ¼ k þ asinh2 ½lðx  ctÞ; ð15Þ

where k, l and a are parameters that will be determined. The proposed cosh or sinh ansatze I can be applied
directly to the given equation to determine these parameters.
 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424 415

2.2. A sinh–cosh ansatz II

In this case we use a linear combination of cosh and sinh ansatz of the form
uðx; tÞ ¼ k þ a cosh½lðx  ctÞ þ b sinh½lðx  ctÞ; ð16Þ
where k, l, a and b are parameters that will be determined. The proposed ansatz can be applied directly.

2.2.1. A cosh ansatz II or a sinh ansatz II

In this case we use a cosh ansatz II that reads
uðx; tÞ ¼ k þ a cosh½lðx  ctÞ ð17Þ

and a sinh ansatz II of the form

uðx; tÞ ¼ k þ a sinh½lðx  ctÞ; ð18Þ
where k, l, and a are parameters that will be determined.

2.3. A tanh ansatz or a coth ansatz

In this case we use a rational tanh ansatz of the form

1
uðx; tÞ ¼ ð19Þ
k þ a tanh½lðx  ctÞ
and a rational coth ansatz of the form
1
uðx; tÞ ¼ ; ð20Þ
k þ a coth½lðx  ctÞ
where k, l and a are parameters that will be determined.

2.4. The exponential peakon algorithm

In this algorithm we assume that u(x, t) has the solution in the form
uðx; tÞ ¼ k þ aeljxctj ; ð21Þ
where k, a and l are parameters that will be determined.

3. The CH family of equations

We first start our analysis by applying the proposed ansatze on the CH generalized form
ut  uxxt þ aux þ buux ¼ kux uxx þ uuxxx : ð22Þ
As stated before, we employ these schemes directly in a straightforward manner to determine the parameters k,
l and a.

3.1. Using a sinh–cosh ansatz I

We begin our analysis by using the sinh–cosh ansatz (13) into the generalized family of the Camassa–Holm
equation (22). We then collect the coefficients of the hyperbolic functions in the resulted equation to find
coeff: of cosh2 : bða2 þ b2 Þ  4l2 ðk þ 1Þða2 þ b2 Þ  8l2 ðab þ kbÞ þ 2bab;

coeff: of cosh0 : 2kl2 ða2 þ b2 Þ þ ð4cl2  4kl2 þ bk þ a  cÞða þ kÞ ð23Þ

þ 4al2 bð4k þ 1Þ  bb2  cb þ 4l2 b2  bab:

416 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424

Equating each coefficient to zero and solving we obtain

k ¼ k0 ; k0 is any arbitrary constant;
b ¼ b0 ; b0 is any arbitrary constant;
2½ðk þ 1Þðc  aÞ  bc ð24Þ
a ¼ b0  2k0 þ ;
rffiffiffiffiffiffiffiffiffiffiffi bk
1 b
l¼ ; k 6¼ 1:
2 kþ1
In view of the last results, we obtain the solitary patterns solutions
  " rffiffiffiffiffiffiffiffiffiffiffi #
2½ðk þ 1Þðc  aÞ  bc 2 1 b
uðx; tÞ ¼ k0 þ b0  2k0 þ cosh ðx  ctÞ
bk 2 kþ1
" rffiffiffiffiffiffiffiffiffiffiffi #
2 1 b
þ b0 sinh ðx  ctÞ ; k 6¼ 1: ð25Þ
2 kþ1

In what follows we list the specific solutions for the variety of equations using the specific values of k and b:
   
1 1
uðx; tÞ ¼ k0 þ ðb0  2k0  aÞcosh2 ðx  ctÞ þ b0 sinh2 ðx  ctÞ ;
2 2
     
2 1 1
uðx; tÞ ¼ k0 þ b0  2k0  a cosh2 ðx  ctÞ þ b0 sinh2 ðx  ctÞ ;
3 2 2
      ð26Þ
8 2 1 2 1
uðx; tÞ ¼ k0 þ b0  2k0  þ 2c cosh ðx  ctÞ þ b0 sinh ðx  ctÞ ;
3 4 4
   
2 1 2 1
uðx; tÞ ¼ k0 þ ðb0  2k0  2aÞcosh ðx  ctÞ þ b0 sinh ðx  ctÞ ;
2 2
for the CH, DP, FW, and FFCH equations, respectively.
It is interesting to point out that the previous solutions are valid only for b > 0, k > 0. However, for b < 0,
k > 0, we obtain the following compactons solutions:
  " rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
2½ðk þ 1Þðc  aÞ  bc 2 1 b
uðx; tÞ ¼ k0 þ b0  2k0 þ cos  ðx  ctÞ
bk 2 kþ1
" rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
2 1 b
 b0 sin  ðx  ctÞ : ð27Þ
2 kþ1

3.1.1. Using a cosh ansatz I

We next use cosh ansatz I (14) into the generalized family of the Camassa–Holm equation (22). Proceeding
as before, collecting the coefficients of the hyperbolic functions in the resulted equation and equating it to zero,
and solving we find
k ¼ k0 ; k0 is any arbitrary constant;
2½ðk þ 1Þðc  aÞ  bc
a ¼ 2k0 þ ;
bk ð28Þ
rffiffiffiffiffiffiffiffiffiffiffi
1 b
l¼ ; k 6¼ 1:
2 kþ1
Consequently, we obtain the solitary patterns solutions
  " rffiffiffiffiffiffiffiffiffiffiffi #
2½ðk þ 1Þðc  aÞ  bc 2 1 b
uðx; tÞ ¼ k0 þ 2k0 þ cosh ðx  ctÞ ; k 6¼ 1: ð29Þ
bk 2 kþ1
 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424 417

This in turn gives the solitary patterns solutions

 
2 1
uðx; tÞ ¼ k0 þ ð2k0  aÞcosh ðx  ctÞ ;
2
   
2 2 1
uðx; tÞ ¼ k0 þ 2k0  a cosh ðx  ctÞ ;
3 2
    ð30Þ
8 2 1
uðx; tÞ ¼ k0 þ 2k0  þ 2c cosh ðx  ctÞ ;
3 4
 
2 1
uðx; tÞ ¼ k0 þ ð2k0  2aÞcosh ðx  ctÞ :
2
For the CH, DP, FW, and FFCH, respectively.
It is interesting to point out that the previous solutions are valid only for b > 0, k > 0. However, for b < 0,
k > 0, we obtain the following compactons solutions:
  " rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
2½ðk þ 1Þðc  aÞ  bc 2 1 b
uðx; tÞ ¼ k0 þ 2k0 þ cos  ðx  ctÞ : ð31Þ
bk 2 kþ1

3.1.2. Using a sinh ansatz I

We now use a sinh ansatz I (15) into the generalized family of the Camassa–Holm equation (22). Proceed-
ing as before, find
k ¼ k0 ; k0 is any arbitrary constant;
2½ðk þ 1Þðc  aÞ  bc
a ¼ 2k0  ;
bk ð32Þ
rffiffiffiffiffiffiffiffiffiffiffi
1 b
l¼ ; k 6¼ 1:
2 kþ1
Consequently, we obtain the solitary patterns solutions
  " rffiffiffiffiffiffiffiffiffiffiffi #
2½ðk þ 1Þðc  aÞ  bc 2 1 b
uðx; tÞ ¼ k0 þ 2k0  sinh ðx  ctÞ ; k 6¼ 1: ð33Þ
bk 2 kþ1

This in turn gives the solutions

 
1
uðx; tÞ ¼ k0 þ ð2k0 þ aÞsinh2 ðx  ctÞ ;
2
   
2 1
uðx; tÞ ¼ k0 þ 2k0 þ a sinh2 ðx  ctÞ ;
3 2
    ð34Þ
8 1
uðx; tÞ ¼ k0 þ 2k0 þ  2c sinh2 ðx  ctÞ ;
3 4
 
1
uðx; tÞ ¼ k0 þ ð2k0 þ 2aÞsinh2 ðx  ctÞ ;
2
for the CH, DP, FW, and FFCH equations, respectively.
It is interesting to point out that the previous solutions are valid only for b > 0, k > 0. However, for b < 0,
k > 0, we obtain the following compactons solutions:
  " rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
2½ðk þ 1Þðc  aÞ  bc 2 1 b
uðx; tÞ ¼ k0  2k0  sin  ðx  ctÞ : ð35Þ
bk 2 kþ1
418 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424

3.2. Using sinh–cosh ansatz II

In this section, we apply the sinh–cosh ansatz II (16). We apply this assumption into the generalized family
of the Camassa–Holm equation (22). Proceeding as before, collecting the coefficients of the hyperbolic func-
tions in the resulted equation, equating it to zero, and solving we find
a ¼ a0 ; a0 is any arbitrary constant;
b ¼ b0 ; b0 is any arbitrary constant;
½ðk þ 1Þðc  aÞ  bc ð36Þ
k¼ ;
bk
rffiffiffiffiffiffiffiffiffiffiffi
b
l¼ ; k 6¼ 1:
kþ1

Consequently, we obtain the travelling wave solutions

  "rffiffiffiffiffiffiffiffiffiffiffi # "rffiffiffiffiffiffiffiffiffiffiffi #
½ðk þ 1Þðc  aÞ  bc b b
uðx; tÞ ¼ þ a0 cosh ðx  ctÞ þ b0 sinh ðx  ctÞ : ð37Þ
bk kþ1 kþ1

This in turn gives the solutions

 a
uðx; tÞ ¼  þ a0 cosh½ðx  ctÞ þ b0 sinh½ðx  ctÞ;
2
 
1
uðx; tÞ ¼  a þ a0 cosh½ðx  ctÞ þ b0 sinh½ðx  ctÞ;
3
      ð38Þ
4 1 1
uðx; tÞ ¼ c  þ a0 cosh ðx  ctÞ þ b0 sinh ðx  ctÞ ;
3 2 2
uðx; tÞ ¼ ðaÞ þ a0 cosh½ðx  ctÞ þ b0 sinh½ðx  ctÞ;
for the CH, DP, FW, and FFCH equations, respectively.
It is normal to point out that the previous solutions are valid only for b > 0, k > 0. However, for b < 0,
k > 0, we obtain the following solutions:
  "rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi # "rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
ððk þ 1Þðc  aÞ  bcÞ b b
uðx; tÞ ¼ þ a0 cos  ðx  ctÞ þ b0 sin  ðx  ctÞ : ð39Þ
bk kþ1 kþ1

3.2.1. Using a cosh ansatz II

We next use cosh ansatz II (17) into the generalized family of the Camassa–Holm equation (22). Proceeding
as before, collecting the coefficients of the hyperbolic functions in the resulted equation, equating it to zero,
and solving we find
a ¼ a0 ; a0 is any arbitrary constant;
½ðk þ 1Þðc  aÞ  bc
k¼ ;
bk ð40Þ
rffiffiffiffiffiffiffiffiffiffiffi
b
l¼ ; k 6¼ 1:
kþ1
Consequently, we obtain the solutions
  "rffiffiffiffiffiffiffiffiffiffiffi #
½ðk þ 1Þðc  aÞ  bc b
uðx; tÞ ¼ þ a0 cosh ðx  ctÞ ; k 6¼ 1: ð41Þ
bk kþ1
 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424 419

Consequently, we obtain the solutions

 a
uðx; tÞ ¼  þ a0 cosh½ðx  ctÞ;
2
 
1
uðx; tÞ ¼  a þ a0 cosh½ðx  ctÞ;
3
    ð42Þ
4 1
uðx; tÞ ¼ c  þ a0 cosh ðx  ctÞ ;
3 2
uðx; tÞ ¼ ðaÞ þ a0 cosh½ðx  ctÞ;

for the CH, DP, FW, and FFCH equations, respectively.

It is worth noting that the previous solutions are valid only for b > 0, k > 0. However, for b < 0, k > 0, we
obtain the following periodic solutions:
  "rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
½ðk þ 1Þðc  aÞ  bc b
uðx; tÞ ¼ þ a0 cos  ðx  ctÞ : ð43Þ
bk kþ1

3.2.2. Using a sinh ansatz II

It is interesting to point out that by using the sinh ansatz II (18) we obtain the same values for the param-
eters k, a, and l. In other words we obtain the general solution for the family in the form

  "rffiffiffiffiffiffiffiffiffiffiffi #
½ðk þ 1Þðc  aÞ  bc b
uðx; tÞ ¼ þ a0 sinh ðx  ctÞ ; k 6¼ 1; ð44Þ
bk kþ1

valid for b > 0. However, for b < 0 we find the periodic solutions
  "rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
½ðk þ 1Þðc  aÞ  bc b
uðx; tÞ ¼ þ a0 sin  ðx  ctÞ : ð45Þ
bk kþ1

3.3. Using a tanh or coth ansatz

We finally use the rational tanh ansatz (19) into the generalized family of the Camassa–Holm equation (22).
Proceeding as before, we find

bk
k¼ ;
2½ðk þ 1Þðc  aÞ  bc
bk
a¼ ; ð46Þ
2½ðk þ 1Þðc  aÞ  bc
rffiffiffiffiffiffiffiffiffiffiffi
1 b
l¼ ; k 6¼ 1:
2 kþ1

Consequently, we obtain the solutions

2½ðk þ 1Þðc  aÞ  bc

uðx; tÞ ¼  h qffiffiffiffiffiffi i ; k 6¼ 1: ð47Þ
bk 1  tanh 12 kþ1b
ðx  ctÞ

The last solution blows up for x ! 1 for +tanh, and blows up for x ! 1 for tanh.
420 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424

This in turn gives the solutions

a
uðx; tÞ ¼  
1  ;
1  tanh 2 ðx  ctÞ
2a
uðx; tÞ ¼  
 ;
3 1  tanh 12 ðx  ctÞ
2ð3c  4aÞ ð48Þ
uðx; tÞ ¼ 
 ;
3 1  tanh 14 ðx  ctÞ
2a
uðx; tÞ ¼  
 ;
1  tanh 12 ðx  ctÞ
for the CH, DP, Fw, and FFCH equations, respectively.
The previous solutions are valid only for b > 0, k > 0. However, for b < 0, k > 0, we obtain complex solu-
tions that are beyond the scope of this work.
Moreover, using the rational coth ansatz (20) gives the same values for the parameters a, k and l. Based on
this we obtain the general solution for the CH family given by
2½ðk þ 1Þðc  aÞ  bc
uðx; tÞ ¼  h qffiffiffiffiffiffi i ; k 6¼ 1: ð49Þ
bk 1  coth 12 kþ1
b
ðx  ctÞ

3.4. Using the exponential peakon algorithm

We finally study the CH family of equations

ut  uxxt þ aux þ buux ¼ kux uxx þ uuxxx : ð50Þ
Substituting the peakon algorithm (21) into (50), and proceeding as before we find
a ¼ a0 ; a0 is any arbitrary constant;
ðk þ 1Þðc  aÞ  bc
k¼ ;
ð51Þ
rffiffiffiffiffiffiffiffiffiffiffi bk
b
l¼ ; k 6¼ 1:
kþ1
This in turn gives the peakon solution
ðk þ 1Þðc  aÞ  bc pffiffiffiffi
b
ffi
uðx; tÞ ¼ þ a0 e kþ1jxctj ; k 6¼ 1: ð52Þ
bk
This means that we obtain the following peakons solutions:
a
uðx; tÞ ¼  þ a0 ejxctj ;
2
a
uðx; tÞ ¼  þ a0 ejxctj ;
3 ð53Þ
uðx; tÞ ¼ a þ a0 ejxctj ;
3c  4 1
uðx; tÞ ¼ þ a0 e2jxctj ;
3
for the CH, DP, FFCH and FW equations, respectively.

4. The nonlinearly dispersive equation

In this section, we proceed in a like manner to the analysis presented before to examine the nonlinearly dis-
persive equation
3
ut  k 1 uxxt þ a1 ux þ uux ¼ 2ux þ b1 uxxx þ uuxxx ; ð54Þ
k
 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424 421

that was first introduced by Fuchssteiner [7]. Because the analysis is identical to that followed before, only
results will be presented and details will be skipped.

4.1. Using a sinh–cosh ansatz I

Using the sinh–cosh ansatz (13) into the nonlinear equation (54) and proceeding as before we obtain
k ¼ k0 ; k0 is any arbitrary constant;
b ¼ b0 ; b0 is any arbitrary constant;
a ¼ b0  2k0  ak þ b; ð55Þ
1
l ¼ pffiffiffi ; k > 0:
2 k
Consequently, we obtain the solitary patterns solutions
   
2 1 2 1
uðx; tÞ ¼ k0 þ ðb0  2k0  ak þ bÞcosh pffiffiffi ðx  ctÞ þ b0 sinh pffiffiffi ðx  ctÞ ; k > 0: ð56Þ
2 k 2 k
However, for k < 0, we obtain the following compactons solutions:
   
1 1
uðx; tÞ ¼ k0 þ ðb0  2k0  ak þ bÞ cos2 pffiffiffiffiffiffiffi ðx  ctÞ  b0 sin2 pffiffiffiffiffiffiffi ðx  ctÞ : ð57Þ
2 k 2 k

4.1.1. Using a cosh ansatz I

Using the cosh ansatz I (14) into equation (54) we find
k ¼ k0 ; k0 is any arbitrary constant;
a ¼ 2k0  ak þ b;
ð58Þ
1
l ¼ pffiffiffi ; k > 0:
2 k
The solitary patterns solutions
 
1 2
uðx; tÞ ¼ k0 þ ð2k0  ak þ bÞcosh p ffiffi
ffi ðx  ctÞ ð59Þ
2 k
are readily obtained. For k < 0, we obtain the following compactons solutions:
 
1
uðx; tÞ ¼ k0 þ ð2k0  ak þ bÞ cos2 pffiffiffiffiffiffiffi ðx  ctÞ : ð60Þ
2 k

4.1.2. Using a sinh ansatz I

Applying a sinh ansatz I (15) into Eq. (54) yields
k ¼ k0 ; k0 is any arbitrary constant;
a ¼ 2k0 þ ak  b;
ð61Þ
1
l ¼ pffiffiffi ; k > 0:
2 k
The solitary patterns solutions
 
1
uðx; tÞ ¼ k0 þ ð2k0 þ ak  bÞsinh2 pffiffiffi ðx  ctÞ ; ð62Þ
2 k
follow immediately.
422 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424

For k < 0, we obtain the following compactons solutions:

 
1
uðx; tÞ ¼ k0  ð2k0 þ ak  bÞ sin2 pffiffiffiffiffiffiffi ðx  ctÞ : ð63Þ
2 k

4.2. Using a sinh–cosh ansatz II

Using the sinh–cosh ansatz II (21) into the nonlinear equation (54) and proceeding as before we obtain

b  ak
k¼ ;
2
b ¼ b0 ; b0 is any arbitrary constant;
ð64Þ
a ¼ a0 ; a0 is any arbitrary constant;
1
l ¼ pffiffiffi ; k > 0:
k
Consequently, we obtain the solitary patterns solutions
   
b  ak 1 1
uðx; tÞ ¼ þ a0 cosh pffiffiffi ðx  ctÞ þ b0 sinh pffiffiffi ðx  ctÞ ; k > 0: ð65Þ
2 k k

However, for k < 0, we obtain the complex solutions

   
b  ak 1 1
uðx; tÞ ¼ þ a0 cos pffiffiffiffiffiffiffi ðx  ctÞ þ ib0 sin pffiffiffiffiffiffiffi ðx  ctÞ ; i2 ¼ 1: ð66Þ
2 k k

4.2.1. Using a cosh ansatz II

Using the cosh ansatz I (17) into Eq. (54) to obtain the solitary patterns solutions
 
b  ak 1
uðx; tÞ ¼ þ a0 cosh pffiffiffi ðx  ctÞ ð67Þ
2 k

are readily obtained. For k < 0, we obtain the following compactons solutions:
 
b  ak 1
uðx; tÞ ¼ þ a0 cos pffiffiffiffiffiffiffi ðx  ctÞ : ð68Þ
2 k

4.2.2. Using a sinh ansatz II

Applying a sinh ansatz I (18) into Eq. (54) to get the solitary patterns solutions
 
b  ak 1
uðx; tÞ ¼ þ a0 sinh pffiffiffi ðx  ctÞ ; ð69Þ
2 k

follow immediately.
For k < 0, we obtain the following complex solutions:
 
b  ak 1
uðx; tÞ ¼ þ ia0 sin pffiffiffi ðx  ctÞ : ð70Þ
2 k
 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424 423

4.3. Using a tanh or coth ansatz

We now use the rational tanh ansatz (19) into Eq. (54). Proceeding as before, we find
1
k¼ ;
ak  b
1
a¼ ; ð71Þ
ak  b
1
l ¼ pffiffiffi ; k > 0:
2 k
Consequently, we obtain the solutions
b  ak
uðx; tÞ ¼  h i ; k > 0: ð72Þ
1  tanh 2p1 ffiffik ðx  ctÞ

The last solution blows up for x ! 1 for +tanh, and blows up for x ! 1 for tanh.
Moreover, using the rational coth ansatz (20) gives the same values for the parameters a, k and l. Based on
this we obtain the general solution for the CH family given by
b  ak
uðx; tÞ ¼  h i ; k > 0: ð73Þ
1  coth 2p1 ffiffik ðx  ctÞ

4.4. Using the exponential peakon algorithm

We finally study the peakon solutions for Eq. (54). Substituting the peakon algorithm (21) into (54), and
proceeding as before we find
a ¼ a0 ; a0 is any arbitrary constant;
b  ak
k¼ ; ð74Þ
2
1
l ¼ pffiffiffi :
k
This in turn gives the peakon solution
b  ak p1ffi jxctj
uðx; tÞ ¼ þ a0 e k ; k > 0: ð75Þ
2

5. Discussion

In this work, a family the Camassa–Holm equations was investigated. The others’ works were focused on
the peakons solutions, peaked solitary wave solutions, where spatial derivative changes sign at a peak at a cor-
ner of its crest.
In this work, a set of new ansatze, where hyperbolic functions are combined, is introduced to formally
derive abundant solutions with distinct physical structures. Solitons, solitary patterns solutions, periodic solu-
tions, compactons, and peakons solutions are obtained. The nonlinearly dispersive equation, that was intro-
duced first by Fuchssteiner [7], was also handled in a like manner. The obtained results complement the useful
works of others for this important physical model. The proposed schemes are useful to be used in identical
nonlinear dispersive models.

References

[1] R. Camassa, D. Holm, An integrable shallow water equation with peaked solitons, Phys. Rev. Lett 71 (11) (1993) 1661–1664.
[2] R. Ivanov, On the integrability of a class of nonlinear dispersive wave equations, J. Nonlinear Math. Phys. 1294 (2005) 462–468.
424 A.-M. Wazwaz / Applied Mathematics and Computation 182 (2006) 412–424

[3] M.S. Alber, C. Miller, On peakon solutions of the shallow water equation, Appl. Math. Lett. 14 (1) (2001) 93–98.
[4] H.R. Dullin, G. Gottwald, D. Holm, Camassa–Holm, Korteweg–de Vries-5 and other asymptotically equivalent equations for
shallow water waves, Fluid Dyn. Res. 33 (2003) 73–95.
[5] Y.A. Li, P.J. Olver, Convergence of solitary-wave solutions in a perturbed bi-Hamiltonian dynamical system: I. Compactons and
peakons, Discr. Contin. Dyn. Syst. 3 (3) (1997) 419–432.
[6] Y.A. Li, P.J. Olver, Convergence of solitary-wave solutions in a perturbed bi-Hamiltonian dynamical system: II. Complex analytic
behavior and convergence to non-analytic solutions, Discr. Contin. Dyn. Syst. 4 (1) (1998) 159–191.
[7] B. Fuchssteiner, Some tricks from the symmetry-toolbox for nonlinear equations: generalization of the Camassa–Holm equation,
Physica D 95 (3–4) (1996) 229–243.
[8] P. Rosenau, A. Pikovsky, Phase compactons in chains of dispersively coupled oscillations, Phys. Rev. Lett. 94 (174102) (2005) 1–4.
[9] A. Degasperis, M. Procesi, Asymptotic Integrability Symmetry and Perturbation Theory, World Scientific Publishing, 2002, pp. 23–
37.
[10] Z. Liu, R. Wang, Z. Jing, Peaked wave solutions of Camassa–Holm equation, Chaos Solitons Fractals 19 (2004) 77–92.
[11] Z. Liu, T. Qian, Peakons of the Camassa–Holm equation, Appl. Math. Modell. 26 (2002) 473–480.
[12] T. Qian, M. Tang, Peakons and periodic cusp waves in a generalized Camassa–Holm equation, Chaos Solitons Fractals 12 (2001)
1347–1360.
[13] L. Tian, X. Song, New peaked solitary wave solutions of the generalized Camassa–Holm equation, Chaos Solitons Fractals 19 (2004)
621–637.
[14] A.M. Wazwaz, New compact and noncompact solutions for variants of a modified Camassa–Holm equation, Appl. Math. Comput.
163 (3) (2005) 1165–1179.
[15] A.M. Wazwaz, A class of nonlinear fourth order variant of a generalized Camassa–Holm equation with compact and noncompat
solutions, Appl. Math. Comput. 165 (2005) 485–501.
[16] A.M. Wazwaz, Partial Differential Equations: Methods and Applications, Balkema Publishers, The Netherlands, 2002.
[17] A .M. Wazwaz, An analytic study of compactons structures in a class of nonlinear dispersive equations, Math. Comput. Simul. 63 (1)
(2003) 35–44.
[18] A.M. Wazwaz, A study on nonlinear dispersive partial differential equations of compact and noncompact solutions, Appl. Math.
Comput. 135 (2–3) (2003) 399–409.
[19] A.M. Wazwaz, A construction of compact and noncompact solutions of nonlinear dispersive equations of even order, Appl. Math.
Comput. 135 (2–3) (2003) 324–411.
[20] A.M. Wazwaz, Compactons in a class of nonlinear dispersive equations, Math. Comput. Modell. 37 (3–4) (2003) 333–341.