A study of the fourth-order nonlinear Benjamin-Ono equation of

plasma physics

Ntladi Derick Maserumule (derick@aims.ac.za)

African Institute for Mathematical Sciences (AIMS)

Supervised by: Professor Chaudry Masood Khalique

North West University, Mafikeng Campus, South Africa

2 November 2023
Submitted in partial fulfillment of a structured masters degree at AIMS South Africa
Abstract
In this project we study the fourth-order nonlinear Benjamin-Ono (BO) equation that describes the
propagations of internal waves in deep water. This equation arises in several scientific fields, such as
plasma physics, nonlinear optics and fluid dynamics. We employ the second-order Boltzmann equation
as an illustrative example of the methods to be utilized in this dissertation. We shall perform Lie
symmetry analysis on the BO equation and explore one-parameter Lie-groups, group invariant solutions,
together with travelling wave solutions. Furthermore, we shall derive conservation laws using multiplier
method.
Keywords : Benjamin-Ono, Lie point symmetry, group invariant solution, conserved vectors

Declaration
I, the undersigned, hereby declare that the work contained in this research project is my original work, and
that any work done by others or by myself previously has been acknowledged and referenced accordingly.

Ntladi Derick Maserumule, 2 November 2023

i
Contents

Abstract i

1 Introduction 1

2 Preliminaries 2
2.1 One-parameter Lie group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2 Infinitesimal transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Group invariants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.4 Establishing of a symmetry group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.5 Prolongation formulae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.6 Lie algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.7 Kudryashov’s method to find solutions to PDEs . . . . . . . . . . . . . . . . . . . . . . 6
2.8 Conserved vectors via multiplier method . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Solutions and conserved vectors of the Boltzmann equation 9

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2 Computing symmetries of (3.1.1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.3 Conserved vectors for (3.1.1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

4 Exact solutions and conserved vectors of the fourth-order nonlinear Benjamin-Ono equa-
tion 21
4.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 Lie symmetries of (4.1.1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.3 Conservation laws of (4.1.1) via the multiplier method. . . . . . . . . . . . . . . . . . . 28
4.4 Concluding remarks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

5 Conlusion 30

References 33

ii
1. Introduction
Most physical phenomena are described by nonlinear partial differential equations (NLPDEs). So it
is important to study these equations and find their exact solutions. However, this is a difficult task
because there is no general theory.
Nevertheless many researchers have developed methods to find special solutions viz., Kudryashov method
(Mahmud et al., 2017), (G0 /G)-expansion method (Shakeel and Mohyud-Din, 2015), and symmetry
analysis (Hydon, 2005). The Lie symmetry method is a systematic approach mostly applied by scientists
to find closed-form exact solutions for differential equations developed in various research fields. This
method has been used for studies of many nonlinear differential equations and was seen considered
effective and precise in finding exact solutions even for difficult equations (Wang et al., 2022).
Conservation laws play an important role in solving differential equations (DEs) as they are helpful in
order reduction and finding solutions to DEs, determining the range of integrable solutions, improving
numerical methods and more (Hong et al., 2020). In the paper by Khalique and Mehmood (2021) it
was stated that a conserved vector for a particular symmetry is actually a symmetry of the system itself
and this finding was exemplified in an equation that emerged during the examination of Maxwellian tails
and in the context of angular momentum in central force problems. Multiple approaches to find the
conserved vectors for DEs exists, viz., Noether method, the Boyer method, the Ibragimov, the variational
symmetries method, and the Herman–Pole method (Kara and Mahomed, 2002). Unlike some of the
methods methods which are extremely limited for application e.g Noether method, there is direct method
that is more general for constructing conserved vectors (Manafian et al., 2021).
A symmetry of a conserved vector corresponds to a symmetry inherent in the system, and this outcome is
demonstrated in an equation encountered in the examination of Maxwellian tails, as well as in the context
of angular momentum when dealing with a central force problem. In this project we study a NLPDE
called Benjamin-Ono (BO) equation, a fourth-order nonlinear equation with qxxxx as a dispersion. The
BO equation was derived to form a mathematical representation of one-dimensional waves in deep water
(Roudenko et al., 2021).
In Chapter 2, we give preliminaries in the form of definitions and theorems. We also provide expressions
of mathematical terms or operators to be used in the later chapters.
In Chapter 3, we give an illustrative example of a Boltzmann equation which is a second-order nonlinear
equation. We find the Lie point symmetries, commutation relations and commutator table for the
symmetries found, transformation groups, invariant solutions, travelling wave solution and conservation
laws using multiplier method for the Boltzmann equation.
Finally, in Chapter 4 we study the BO equation. We apply Lie symmetry symmetry method to find
group-invariant solutions together with travelling wave solution. We use Kudryashov’s method to find
the travelling wave solution and the power series solution method to find one of group-invariant solutions.
Finally we derive conservation laws for the BO equation using the multiplier method.

1
2. Preliminaries
In this chapter we recall some definitions and theorems of Lie theory and also demonstrate the algorithm
to find symmetries of a given partial differential equation (PDE). The contents of this chapter can be
found in (Ovsiannikov, 2014; Bluman and Kumei, 2013; Olver, 1993; Ibragimov, 1995; Ibragimov)

2.1 One-parameter Lie group

Consider t, x as independent variables and q a dependent variable and contemplate a change in these
variables as:
Tr : t̄ = F (t, x, q, r), x̄ = G(t, x, q, r), q̄ = H(t, x, q, r), (2.1.1)
where r is a real constant that continuously ranges in values from a neighbourhood N 0 ⊂ N ⊂ R of
r = 0 and F , G and H are differentiable functions.

2.1.1 Definition. A set G of transformations (2.1.1) is said to be a continuous one-parameter Lie group
of transformations in t, x, q provided
(i) (Closure) For Tr1 , Tr2 ∈ G where r1 , r2 ∈ N 0 ⊂ N then Tr1 Tr2 = Tr3 ∈ G, r3 = Φ(r1 , r2 ) ∈ N
(ii) (Identity) T0 ∈ G if and only if r = 0 such that T0 Tr = Tr T0 = Tr
(iii) (Inverse) For Tr1 ∈ G, r1 ∈ N 0 ⊂ N , Tr−1
1
= Tr1 −1 ∈ G, r1 −1 ∈ N such that
Tr1 Tr1 −1 = Tr1 −1 Tr1 = T0 .

The property of associativity follows from (i). Also, if the identity transformation occurs at r = r0 6= 0
i.e., Tr0 is the identity, then a shift of r = r̄ + r0 will give T0 as above. The group property (i) can be
written as
t̄¯ ≡ F (t̄, x̄, q̄, r ) = F (t, x, q, Φ(r , r )),
2 1 2
¯ ≡ G(t̄, x̄, q̄, r2 ) = G(t, x, q, Φ(r1 , r2 )),
x̄
q̄¯ ≡ H(t̄, x̄, q̄, r2 ) = H(t, x, q, Φ(r1 , r2 )). (2.1.2)
Here Φ gives group composition law. The parameter r is said to be canonical provided Φ(r1 , r2 ) = r1 +r2 .

2.1.2 Theorem. For any φ(a, b), there exists the canonical parameter ã defined by
Z a
ds ∂ φ(s, b)
ã = , where w(s) = .
0 w(s) ∂b b=0

Let us now consider the PDE

P(t, x, q, qt , qx , qxx , . . . ) = 0. (2.1.3)

2.1.3 Definition. A one-parameter group G of transformations (2.1.1) is said to be a symmetry group

of (2.1.3) if equation (2.1.3) has the same form in variables t̄, x̄ and q̄, i.e.,
P(t̄, x̄, q̄, q̄t̄ , q̄x̄ , q̄x̄x̄ , . . . ) = 0 (2.1.4)
with P being the same function as in (2.1.3).

2
Section 2.2. Infinitesimal transformations Page 3

2.2 Infinitesimal transformations

In accordance with the Lie’s theory, the establishment of symmetry group G is equal to finding the
associated infinitesimal transformations

t̄ = t + r τ (t, x, q), x̄ = x + r ξ(t, x, q), q̄ = q + r η(t, x, q) (2.2.1)

acquired from (2.1.1) by expanding F , G and H in Taylor series in r about r = 0 and also using the
initial conditions

F |r=0 = t, G|r=0 = x, H|r=0 = q .

Thus, we have
∂F ∂G ∂H
τ (t, x, q) = , ξ(t, x, q) = , η(t, x, q) = . (2.2.2)
∂r r=0 ∂r r=0 ∂r r=0

We define generator X of infinitesimal transformations by writing (2.2.1) as

t̄ = (1 + r X )t, x̄ = (1 + r X )x, q̄ = (1 + r X )q ,

where
∂ ∂ ∂
X = τ (t, x, q) + ξ(t, x, q) + η(t, x, q) . (2.2.3)
∂t ∂x ∂q
The operator X is called the infinitesimal generator of the group G. If G is admitted by the PDE (2.1.3),
it is said that X is called an admitted operator of (2.1.3) or X is a Lie point symmetry of the PDE
(2.1.3).

2.3 Group invariants

2.3.1 Definition. A function I(t, x, q) is said to be an invariant of group of transformations (2.1.1)
provided

I(t̄, x̄, q̄) = I(t, x, q), (2.3.1)

for all t, x, q and r.

2.3.2 Theorem. A necessary and sufficient condition for I(t, x, q) to be an invariant is that
∂I ∂I ∂I
X I ≡ τ (t, x, q) + ξ(t, x, q) + η(t, x, q) = 0. (2.3.2)
∂t ∂x ∂q

Thus every one-parameter group of transformations (2.1.1) has two invariants that are functionally
independent. These can be taken as left-hand side of any first integrals

I1 (t, x, q) = K1 , I2 (t, x, q) = K2 ,

of the corresponding Lagrange system

dt dx dq
= = .
τ (t, x, q) ξ(t, x, q) η(t, x, q)
Section 2.4. Establishing of a symmetry group Page 4

2.3.3 Theorem. (Lie equations) For transformations (2.2.1) or its generator X , the associated one-
parameter group G is acquired by finding the solution to Lie equations

dt̄ dx̄ dq̄

= τ (t̄, x̄, q̄), = ξ(t̄, x̄, q̄), = η(t̄, x̄, q̄) (2.3.3)
dr dr dr
along with initial conditions

t̄ r=0 = t, x̄|r=0 = x, q̄|r=0 = q .

2.4 Establishing of a symmetry group

We now present the algorithm to calculate Lie point symmetries for a PDE. First we present some
definitions.

2.5 Prolongation formulae

Let us contemplate a PDE
E(t, x, q, qt , qx , qtt , qxx , qtx , . . .) = 0. (2.5.1)
Here the variables t, x are independent and q is dependent. Consider
∂ ∂ ∂
X = τ (t, x, q) + ξ(t, x, q) + η(t, x, q)
∂t ∂x ∂q
as a generator of one-parameter group G. The first prolongation of X is defined by
∂ ∂
X [1] = X + ζt (t, x, q, qt , qx ) + ζx (t, x, q, qt , qx ) ,
∂qt ∂qx
where

ζt = Dt (η) − qt Dt (τ ) − qx Dt (ξ),
ζx = Dx (η) − qt Dx (τ ) − qx Dx (ξ)

and Dt , Dx are total derivatives given by

∂ ∂ ∂ ∂
Dt = + qt + qtx + qtt + ··· , (2.5.2)
∂t ∂q ∂qx ∂qt
∂ ∂ ∂ ∂
Dx = + qx + qxx + qtx + ··· . (2.5.3)
∂x ∂q ∂qx ∂qt
Similarly, the second prolongation of X is given by
∂ ∂ ∂ ∂ ∂
X [2] = X + ζt + ζx + ζtt + ζtx + ζxx ,
∂qt ∂qx ∂qtt ∂qtx ∂qxx
where

ζtt = Dt (ζt ) − qtt Dt (τ ) − qtx Dt (ξ),

ζtx = Dx (ζt ) − qtt Dx (τ ) − qtx Dx (ξ),
ζxx = Dx (ζx ) − qtx Dx (τ ) − qxx Dx (ξ).
Section 2.5. Prolongation formulae Page 5

Thus invoking the definitions of Dt and Dx , we have

ζt = ηt + qt ηq − qt τt − qt2 τq − qx ξt − qt qx ξq ,
ζx = ηx + qx ηq − qt τx − qt qx τq − qx ξx − qx2 ξq ,
ζtt = ηtt + 2qt ηtq + qtt ηq + (qt )2 ηqq − 2qtt τt − qt τtt − 2(qt )2 τtq
− 3qt qtt τq − (qt )3 τqq − 2qtx ξt − qx ξtt − 2qt qx ξtq − (qt )2 qx ξqq
− (qx qtt + 2qt qtx )ξq ,
ζtx = ηtx + qx ηtq + qt ηxq + qxt ηq + qt qx ηqq − qtx (τt + ξx ) − qt τtx − qtt τx
− qt qx (τtq + ξxq ) − qt2 τxq − (2qt qtx + qx qtt )τq − (qt )2 qx τqq − qx ξtx
− qxx ξt − (qx )2 ξtq − (2qx qtx + qt qxx )ξq − qt (qx )2 ξqq ,
ζxx = ηxx + 2qx ηxq + qxx ηq + (qx )2 ηqq − 2qxx ξx − qx ξxx − 2(qx )2 ξxq
− 3qx qxx ξq − (qx )3 ξqq − 2qtx τx − qt τxx
− 2qt qx τxq − (qt qxx + 2qx qtx )τq − qt (qx )2 τqq .

Likewise, ζxxx and ζxxxx can be computed and are given, respectively, by

ζxxx =Dx (ζxx ) − qtxx Dx (τ ) − qxxx Dx (ξ)

= − qt τqqq qx3 − 3τqq qx2 qtx − 3qt qx2 τxqq − 3τq qx qtxx − 3qt τqq qxx qx
− 6qx qtx τxq − 3qt qx τxxq − qt τq qxxx − 3τq qxx qtx − 3τx qtxx
− 3qt qxx τxq − 3τxx qtx − qt τxxx + ηqqq qx3 + 3qx2 ηxqq + 3ηqq qxx qx
+ 3qx ηxxq + ηq qxxx + 3qxx ηxq − ξqqq qx4 − 3qx3 ξxqq − 6ξqq qxx qx2
− 3qx2 ξxxq − 4ξq qxxx qx − 9qxx qx ξxq − qx ξxxx − 3ξq qxx
2

− 3qxxx ξx − 3qxx ξxx + ηxxx

and

ζxxxx = Dx (ζxxx ) − qtxxx Dx (τ ) − qxxxx Dx (ξ)

= − ξqqqq qx5 + ηqqqq qx4 − qt τqqqq qx4 − 4ξxqqq qx4 − 10qxx ξqqq qx3 − 4qtx τqqq qx3
+ 4ηxqqq qx3 − 4qt τxqqq qx3 − 6ξxxqq qx3 − 10qxxx ξqq qx2 − 6qtxx τqq qx2
+ 6qxx ηqqq qx2 − 6qxx qt τqqq qx2 − 24qxx ξxqq qx2 − 12qtx τxqq qx2 + 6ηxxqq qx2
− 6qt τxxqq qx2 − 4ξxxxq qx2 − 5qxxxx ξq qx − 4qtxxx τq qx + 4qxxx ηqq qx
2
− 15qxx ξqq qx − 4qxxx qt τqq qx − 12qxx qtx τqq qx − 16qxxx ξxq qx − 12qtxx τxq qx
+ 12qxx ηxqq qx − 12qxx qt τxqq qx − 18qxx ξxxq qx − 12qtx τxxq qx + 4ηxxxq qx
− 4qt τxxxq qx − ξxxxx qx + qxxxx ηq − 10qxx qxxx ξq − qxxxx qt τq − 4qxxx qtx τq
2 2
− 6qxx qtxx τq + 3qxx ηqq − 3qxx qt τqq − 4qxxxx ξx − 4qtxxx τx + 4qxxx ηxq
2
− 12qxx ξxq − 4qxxx qt τxq − 12qxx qtx τxq − 6qxxx ξxx − 6qtxx τxx + 6qxx ηxxq
− 6qxx qt τxxq − 4qxx ξxxx − 4qtx τxxx + ηxxxx − qt τxxxx .

It should be noted that the expression for ζxxxx will be needed in Chapter 4.
Section 2.6. Lie algebras Page 6

2.5.1 Group admitted by a PDE

The infinitesimal operator
∂ ∂ ∂
X = τ (t, x, q) + ξ(t, x, q) + η(t, x, q) (2.5.4)
∂t ∂x ∂q

is said to be a Lie symmetry of the PDE (2.5.1) provided

X [2] (E) = 0 (2.5.5)

whenever E = 0. This is written as

X [2] E = 0. (2.5.6)
E=0

Here the symbol |E=0 denotes evaluated on E = 0.

2.5.2 Definition. Equation (2.5.6) is called the determining equation of Lie point symmetries of (2.5.1).

2.6 Lie algebras

Let S1 and S2 be given as
∂ ∂ ∂
S1 = τ 1 (t, x, q) + ξ 1 (t, x, q) + η 1 (t, x, q) , (2.6.1)
∂t ∂x ∂q
and
∂ ∂ ∂
S2 = τ 2 (t, x, q) + ξ 2 (t, x, q) + η 2 (t, x, q) .
∂t ∂x ∂u

2.6.1 Definition. The commutator (or Lie bracket) of S1 , S2 , denoted as [S1 , S2 ], is defined as
[S1 , S2 ] = S1 (S2 ) − S2 (S1 ).

2.6.2 Definition. A Lie algebra is a vector space Ξ of operators which satisfies the condition that for
any S1 , S2 ∈ Ξ, the commutator [S1 , S2 ] ∈ Ξ.

2.6.3 Theorem. The set of all solutions of any determining equation forms a Lie algebra.

2.7 Kudryashov’s method to find solutions to PDEs

Let us recall the PDE (2.5.1), viz.,

E(t, x, q, qt , qx , qtt , qxx , qtx , . . .) = 0 (2.7.1)

with t, x as independent variables and q dependent variable.

Kudryashov’s method is a power tool which can be used to find exact solutions of PDEs. It can be
described as follows (Kudryashov, 2012): can be utilized to seek out the exact travelling wave solutions
of the equations.
Step 1.
Section 2.8. Conserved vectors via multiplier method Page 7

The PDE (2.7.1) is transformed to an ODE by utilizing the substitution ξ = ax + bt, q(t, x) = Q(ξ) in
the form
P(Q, Q0 , Q00 , Q000 , . . . ) = 0. (2.7.2)

d
Here, P is a function of u(ξ), prime denotes and V is the speed of travelling wave.
dξ
Step 2.
We assume the solution of (2.7.2) takes the form
L
X
Q(ξ) = Cj ψ j (ξ), (2.7.3)
j=0

were Cj (j = 0, 1, 2, ..., L) are constants that are to be found, and ψ solves the ODE

ψ 0 (ξ) = ψ 2 (ξ) − ψ(ξ) (2.7.4)

whose solution is
1
ψ(ξ) = , (2.7.5)
1 + Keξ
where K is a constant.
Step 3.
By using the balancing procedure we determine the positive integer L in (2.7.3).
Step 4.
We substitute (2.7.3) into (2.7.2) and invoke (2.7.4) to obtain a polynomial in ψ j . Equating the like
powers of ψ j to zero, we obtain a system of algebraic equations, which are solved using Maple or
Mathematica to give the values of constants Ci .

2.8 Conserved vectors via multiplier method

Consider mth order system of PDEs of independent variables x = (x1 , x2 ) and dependent variable q,
given by

E(x, q, q(1) , . . . , q(m) ) = 0. (2.8.1)

2.8.1 Definition. The Euler operator is defined by (Ibragimov, 1995)

δ ∂ X ∂
= + (−1)s Di1 . . . Dis . (2.8.2)
δq ∂q s≥1 ∂qi1 i2 ...is

2.8.2 Definition. The vector C = (C t , C x ), for C t , C x ∈ A, is a conserved vector of (2.8.1) if C t , C x

satisfy

Dt C t + Dx C x |(2.8.1) = 0. (2.8.3)

The equation (2.8.3) defines a local conservation law of (2.8.1) (?).

Section 2.9. Summary Page 8

We now give the algorithm to find the conserved vectors for PDEs. A multiplier Q(x, q, q(1) , . . .) has
the property that

QE = Dt C t + Dx C x (2.8.4)

holds identically.
The multiplier Q is obtained by solving the determining equation

δ(QE)
= 0. (2.8.5)
δq
Once we find the multipliers, then we can use the homotopy operator to construct conserved vectors
(Olver, 1993).

2.9 Summary
In this chapter we presented a short introduction to Lie theory. First we gave some basics definitions
and stated some theorems and thereafter we delivered the algorithm to construct symmetries of PDEs.
We described Kudryashov’s method for finding exact solutions of PDEs and thereafter we presented the
multiplier method for deriving conserved vectors for PDEs. These results will be used in the latter part
of our project.
3. Solutions and conserved vectors of the
Boltzmann equation
3.1 Introduction
The one-dimensional (n + 1)th Boltzmann equation, with n being a positive integer, reads (Ernst, 1981)
n
∂ ∂ Γ(2n) 2
 
+1 − q= q .
∂t ∂x Γ(n)

Here q depends on (t, x). Krook and Wu (Krook and Wu, 1976) derived the Boltzmann equation.
This equation appears in development of Maxwellian tails in gas dynamics Euler et al. (1988); Kara
and Mahomed (2000) and furnishes a statistical account of a gas of interconnecting particles (Caflisch,
1980) in kinetic theory.

For n = 1, the (n + 1)th Boltzmann equation becomes

qtx + qx + q 2 = 0. (3.1.1)

In this chapter we examine (3.1.1) by computing its symmetries, performing symmetry reductions,
constructing group-invariant solutions and finally deriving its conserved vectors.

3.2 Computing symmetries of (3.1.1)

The vector field
∂ ∂ ∂
S = τ (t, x, q) + ξ(t, x, q) + η(t, x, q)
∂t ∂x ∂q
is a symmetry of the Boltzmann equation (3.1.1) if

S [2] (qtx + qx + q 2 ) = 0, (3.2.1)

(3.1.1)

where
∂ ∂ ∂
S [2] = η + ζx + ζtx
∂q ∂qx ∂qtx
is second prolongation of S. Equation (3.2.1) then becomes

η(2q) + ζx + ζtx |(3.1.1) = 0, (3.2.2)

The values of ζx and ζtx are now substituted in (3.2.2) and qtx is replaced by −(qx + q 2 ). We get

2qη + ηx + qx ηq − qt τx − qt qx τq − qx ξx − qx2 ξq + ηtx + qx ηtq + qt ηxq

−qx ηq − q 2 ηq + qt qx ηqq + qx (τt + ξx ) + q 2 (τt + ξx ) − qt τtx − qtt τx
−qt qx (τtx + ξxq ) − qt2 τxq + 2(qt qx + qt q 2 )τq − qx qtt τq − qt2 qx τqq
−qx ξtx − qxx ξt − qx2 ξtq + 2(qx2 + qx q 2 )ξq − qt qxx ξq − qt qx2 ξqq = 0.

9
Section 3.2. Computing symmetries of (3.1.1) Page 10

Now separating the above equation on derivatives of q gives

qxx : ξt = 0,
qt qxx : ξq = 0,
qt qx2 : ξqq = 0,
qt2 qx : τqq = 0,
qt2 : τxq = 0,
qx qtt : τq = 0,
qtt : τx = 0,
qx2 : −ξtq + ξq = 0,
qt qx : τq − τtx + ηqq − ξxq = 0,
qx : τt + ηtq − ξtx + q 2 ξq = 0,
qt : −τx − τtx + ηxq+ + 2q 2 τq = 0,
Rest : q 2 (τt − ηq + ξx ) + 2q(η) + ηx + ηtx = 0.

The above twelve equations can be simplified and so we get eight PDEs, viz.,

τq = 0, (3.2.3)
τx = 0, (3.2.4)
ξt = 0, (3.2.5)
ξq = 0, (3.2.6)
ηqq = 0, (3.2.7)
τt + ηtq = 0, (3.2.8)
ηxq = 0, (3.2.9)
2
q (τt − ηq + ξx ) + 2q(η) + ηx + ηtx = 0. (3.2.10)

Equations (3.2.3) and (3.2.4) yields

τ = a(t),
where a(t) depends on t only. From (3.2.5) and (3.2.6) we obtain

ξ = b(x),

where b(x) depends on x only. Integrating (3.2.7) with regards to q gives ηq = c(t, x) and integrating
again we have
η = c(t, x)q + d(t, x), (3.2.11)
where the functions c(t, x), d(t, x) depend on x, t. If we differentiate equation (3.2.11) with regards to
x and q we have

ηx = cx (t, x)q + dx (t, x), (3.2.12)

ηxq = cx (t, x). (3.2.13)

But from (3.2.9), namely ηxq = 0, we get cx (t, x) = 0 and we observe that c does not depend on x but
only on t. So
η = c(t)q + dx (t, x). (3.2.14)
Section 3.2. Computing symmetries of (3.1.1) Page 11

By differentiating (3.2.14) with regards to t and x we have

ηtx = dtx . (3.2.15)
From (3.2.15) we have
τt = −ηtq ,
a (t) = −c0 (t).
0

Substituting τ, ξ and η in (3.2.10) yields

q 2 (a0 (t) + c(t) + b0 (x)) + 2qd(t, x) + dx (t, x) + dt,x (t, x) = 0. (3.2.16)
Now separating (3.2.16) on powers of q, we obtain
q 2 : a0 (t) + c(t) + b0 (x) = 0,
q : d = 0,
q 0 : dx + dt,x = 0.
The above three equations can be simplified and reduced to
a0 (t) + c(t) + b0 (x) = 0. (3.2.17)
From (3.2.17)
c(t) = −a0 (t) − b0 (x). (3.2.18)
From (3.2.18) replace a0 (t) by c0 (t) to obtain,
c0 (t) − c(t) = b0 (x) (3.2.19)
The above equation implies that both sides should be constant, say A1 . Hence
b0 (x) =A1 , (3.2.20)
0
c (t) − c(t) =A1 . (3.2.21)
Equations (3.2.20) and (3.2.21) can easily be solved to obtain the solutions
b(x) = A1 x + A2 , (3.2.22)
t
c(t) = A3 e − A1 , (3.2.23)
where A2 and A3 are constants. Since a0 (t) = −c0 (t) this implies that a0 (t) = −A3 et , and thus
a(t) = A4 − A3 et , where A4 is a constant. Therefore
τ = A4 − A3 e t ,
ξ = A1 x + A2 ,
η = q(A3 et − A1 ).
Thus the infinitesimal generator becomes
∂ ∂ ∂
S = (A4 − A3 et ) + (A1 x + A2 ) + q(A3 et − A1 ) . (3.2.24)
∂t ∂x ∂q
Hence the symmetries of Boltzmann equation (3.1.1) are
∂ ∂ ∂ ∂ ∂ ∂
S1 = , S2 = , S3 = x − q , S4 = −et + qet . (3.2.25)
∂t ∂x ∂x ∂q ∂t ∂u
Section 3.2. Computing symmetries of (3.1.1) Page 12

3.2.1 Commutator table for symmetries of (3.1.1)

We compute the commutation relation for the four symmetries of (3.1.1). We start by calculating
[S1 , S2 ]. Now by definition

[S1 , S2 ] = S1 S2 − S2 S1
∂ ∂ ∂ ∂
   
= −
∂t ∂x ∂x ∂t
= 0.

Likewise
∂ ∂ ∂ ∂ ∂ ∂
    
[S1 , S3 ] = x −q − x −q
∂t ∂x ∂q ∂x ∂q ∂t
= 0.

Also
∂ ∂ ∂ ∂ ∂ ∂
    
[S2 , S3 ] = x −q − x −q
∂x ∂x ∂q ∂x ∂q ∂x
∂
=
∂x
= S2 .

The Lie bracket of S1 and S4 is

∂ ∂ ∂ ∂ ∂ ∂
   
[S1 , S4 ] = −et + qet − −e t
+ qet
∂t ∂t ∂q ∂t ∂q ∂t
∂ ∂
= −et + qet
∂t ∂q
= S4 ,

whereas
∂ ∂ ∂ ∂ ∂ ∂
   
[S2 , S4 ] = −et + qet − −e t
+ qet =0
∂x ∂t ∂q ∂t ∂q ∂x
Finally
∂ ∂ ∂ ∂ ∂ ∂ ∂ ∂
     
t
[S3 S4 ] = x −q −e + qet − −e t
+ qet x −q =0
∂x ∂q ∂t ∂q ∂t ∂q ∂x ∂q
Thus the commutator table of Lie algebras of the Boltzmann equation is given below. .

[Si , Sj ] S1 S2 S3 S4
S1 0 0 0 S4
S2 0 0 S2 0
S3 0 −S2 0 0
−S4 −S4 0 0 0

Table 3.1: Commutator table

Section 3.2. Computing symmetries of (3.1.1) Page 13

3.2.2 Transformation groups of (3.1.1)

We utilize the Lie equations (LEs) and associated initial conditions

dt̄ dx̄ dq̄

= τ (t̄, x̄, q̄), t̄|a=0 = t, = ξ(t̄, x̄, q̄), x̄|a=0 = x, = η(t̄, x̄, q̄), q̄|a=0 = q
da da da
to find the transformation groups for each of the four symmetries. For each Si , let Gi be associated
group. Firstly we calculate the one-parameter group associated with the symmetry S1 , namely
∂
S1 = .
∂t
Using LEs, we get

dt̄ dx̄ dq̄

= 1, t̄|a=0 = t, = 0, x̄|a=0 = x, = 0, q̄|a=0 = q.
da da da
dt̄
From da = 1 we get dt̄ = da. Integrating both sides we obtain t̄ = a + c, with c being a constant.
Applying the initial condition we have c = t and so t̄ = t + a.
The other two Lie equations and initial conditions provide us with x̄ = x and q̄ = q. Thus, the one
parameter group G1 associated with the symmetry S1 is given by

G1 : t̄ = t + a, x̄ = x, q̄ = q.
∂
Likewise for the symmetry S2 = ∂x , we get

G2 : t̄ = t, x̄ = x + a, q̄ = q.
∂ ∂
Considering the symmetry S3 = x ∂x − q ∂q , we obtain

G3 : t̄ = t, x̄ = xea , q̄ = qe−a .
∂ ∂
Finally, the symmetry S4 − et ∂t + qet ∂q gives

G4 : t̄ = − ln a + e−t , x̄ = x, q̄ = q(1 + aet ).

Because each Gi is a symmetry group, we conclude that if q = f (t, x) is a solution of (3.1.1) then so
are

q1 = f (t − a, x),
q2 = f (t, x − a),
q3 = e−a f (t, e−a x),
1
   
−t
q4 = f − ln e − a , x .
1 − aet

3.2.3 Construction of group-invariant solutions of (3.1.1)

We now make use of the Lie point symmetries acquired in previous section to perform symmetry reduc-
tions and obtain group-invariant solutions for the Boltzmann equation (3.1.1).
Section 3.2. Computing symmetries of (3.1.1) Page 14

∂
Case 1. Firstly, we commence our analysis by examining the symmetry S1 = ∂t . The Lagrange system
associated with S1 is
dt dx dq
= = .
1 0 0
Clearly the two invariants are J1 = x and J2 = q. Thus the invariant solution is

q = φ(x).

The substitution of q into equation (3.1.1) yields the first-order nonlinear ordinary differential equation
(NLODE)
φ0 + φ2 = 0,
whose solution is
1
φ(x) = (3.2.26)
C +x
with C a constant. Hence the group-invariant solution of the Boltzmann equation (3.1.1) under S1 is
1
q(t, x) = . (3.2.27)
C +x
Case 2.
∂
For symmetry generator S2 = ∂x the associated Lagrange system is

dt dx dq
= = .
0 1 0
The two invariant are J1 = t and J2 = q. The invariant solution is q = h(t), hence we have

qx = 0, qtx = 0. (3.2.28)

Substituting (4.2.18) into (3.1.1) we get

h(t) = 0

and hence the solution of (3.1.1) under S2 is

q(t, x) = 0.

Case 3.
∂ ∂
For S3 = x ∂x − q ∂q the associated Lagrange system is

dt dx dq
= = ,
0 x −q
 
which gives J1 = t and J2 = qx as invariants. Hence we have q = x1 Φ(t), where Φ depends on t
only. The substitution of this expression for q into equation (3.1.1) gives the first-order NLODE

Φ0 (t) + Φ(t) − Φ2 (t) = 0.

Simplifying the above equation we get

dΦ dΦ
− = dt,
Φ−1 Φ
Section 3.2. Computing symmetries of (3.1.1) Page 15

which integrates to

ln(Φ − 1) − ln |Φ| = t + k,

where k is a constant. This gives

1
Φ(t) = ,
1 − Cet
where C = ek is a constant. Hence solution of (3.1.1) under S3 is
1
q(t, x) = .
x(1 − Cet )
Case 4.
∂ ∂
The associated Lagrange system for the point symmetry S4 = −et ∂t + qet ∂q is

dt dx dq
t
= = t.
−e 0 qe

This gives J1 = x and J2 = qet as two invariants. Consequently the solution is J2 = Ψ(J1 ), where Ψ
depends on x only. Hence

q(t, x) = e−t Ψ(x).

The substitution of this q in (3.1.1) gives

Ψ(x) = 0

and therefore the solution to (3.1.1) under S4 is

q(t, x) = 0.

Case 5.
∂ ∂
Let us now consider the symmetry S1 + aS4 = (1 − aet ) ∂t + aqet ∂q , with a an arbitrary constant and
obtain an exact solution of (3.1.1). The associated Lagrange system to S1 + aS4 is
dt dx ∂
= = aqet ,
1 − aet 0 ∂q

which provides invariants J1 = x, J2 = q(1 − aet ). Hence J2 = ψ(J1 ), where ψ depends on x only, is
the solution. Thus
ψ(x)
q(t, x) = .
1 − aet
The substitution of q in (3.1.1) gives ψ 0 (x) + ψ 2 (x) = 0, whose solution is
1
ψ(x) = , (3.2.29)
x+K
where K is a constant. Hence
1
q(t, x) = . (3.2.30)
(x + K)(1 − aet )
is the solution of (3.1.1) under S1 + aS4 .
Section 3.2. Computing symmetries of (3.1.1) Page 16

Case 6.
(Travelling wave solutions) Finally we consider the combination S of translation symmetries S1 and S2 ,
viz., S = S1 + cS2 , with c being a constant. Hence
∂ ∂
S= +c .
∂t ∂x
The corresponding Lagrange system is
dt dx dq
= = .
1 c 0
Solving the above we get the invariants, J1 = q and J2 = x − ct. So the invariant solution is

q = h(x − ct),

where h depends on its argument. We then have

qx = h0 (x − ct), qtx = −ch00 (x − ct)

and so (3.1.1) becomes

ch00 (z) − h0 (z) − h2 (z) = 0, z = x − ct. (3.2.31)
Now we employ the power series method to solve the NLODE (3.2.31). Suppose the solution to (3.2.31)
takes the form ∞ X
h(z) = an z n ,
n=0

with ai (i = 0, 1, 2, ...) being constants to be found. The first and second derivatives to be used are
given, respectively, as
∞ ∞
h0 = nan z n−1 or h0 =
X X
(n + 1)an+1 z n
n=1 n=0
and ∞ ∞
00 00
X X
n−2
h = n(n − 1)an z or h = (n + 2)(n + 1)an+2 z n .
n=2 n=0

Inserting the values of h, h0 and h00 into (3.2.31), we get

∞ ∞ ∞ ∞
! !
X X X X
n n n n
c (n + 2)(n + 1)an+2 z − (n + 1)an+1 z − an z an z
n=0 n=0 n=0 n=0
∞
X ∞
X
= 2ca2 + 6caz3 + c (n + 2)(n + 1)an+2 z n − a1 − 2a2 z − (n + 1)an+1 z n
n=2 n=0
∞ ∞
!
X X
− a20 − a21 z − ak an−k z n = 0. (3.2.32)
n=0 n=0

Comparing the coefficients of z in equation (3.2.32) for n = 0 and n = 1, we have

2ca2 − a1 − a20 = 0 and 6ca3 − 2a2 − a21 = 0.

By considering n ≥ 2, we get the recurrence relation

(n + 1)an+1 + ∞
P
k=0 ak an−k
an+2 = . (3.2.33)
c(n + 2)(n + 1)
Section 3.2. Computing symmetries of (3.1.1) Page 17

Invoking the recursion relation (3.2.33) one can easily obtain other terms. Hence the power series
solution of (3.2.31) can be written as
a1 + a20 2 2a2 + a21 3 (n + 1)an+1 + ∞
P
k=0 ak an−k n+2
h(z) = a0 + a1 z + z + z + z ,
2c 6c c(n + 2)(n + 1)
and so the power series solution of the Boltzmann equation (3.1.1) is
a1 + a20 2a2 + a21
q(t, x) = a0 + a1 (x − ct) + (x − ct)2 + (x − ct)3 + · · · .
2c 6c
P∞ n
Using power series for the above problem let h(z) = n=0 an z (z = x − ct)and when we differentiate
h,
∞ ∞
0 00
X X
n−1
h = nan z and h = n(n − 1)an z n−2 .
n=1 n=0
Substituting h, h0 and h00 into (3.1.1) gives,
∞ ∞ ∞
!2
X X X
n−2 n−1 n
−c n(n − 1)an−2 z + nan−1 z =− an z . (3.2.34)
n=2 n=1 n=0

The left hand side (LHS) of (3.2.34) and h2 are simplified and expanded below respectively,
∞
X
LHS = [−c(n + 2)(n + 1)an+2 z 2 + (n + 1)an+1 ]z n
n=0

h2 = (a20 + a0 a1 z + a0 a2 z 2 + a0 a3 z 3 + ...) + (a0 a1 z + a21 z 2 + a1 a2 z 3 + ...)

+ (a0 a2 z 2 + a1 a2 z 3 + ...) + (a0 a3 z 3 + ...) + ...,
= a20 + 2(a0 a1 )z + (2a0 a2 + a21 )z 2 + 2(a0 a3 + a1 a2 )z 3 + ...

Grouping each coefficient of z n to obtain the values of an ,

a20 + a1
a2 = ,
2c
2a0 a1 + 2a2
a3 = ,
6c
2a0 a2 + a21 + 3a3
a4 = ,
12c
2a0 a3 + 2a1 a2 + 4a4
a5 = ,
20c
2a0 a4 + 2a1 a3 + a22 + 5a5
a6 = ,
30c
..
.
Pn−1
k=1 ak−1 an−k−1 + (n − 1)an−1
an = .
n(n − 1)c
Where k = 1, 2, 3, ...

∞ Pn−1 !
X
k=1 ak−1 an−k−1 + (n − 1)an−1
u = h(z) = .
n=0
n(n − 1)c
Section 3.3. Conserved vectors for (3.1.1) Page 18

3.3 Conserved vectors for (3.1.1)

We now construct conserved vectors of Boltzmann equation (3.1.1) using the multiplier method.
Firstly we look for zeroth order multipliers Q = Q(t, x, q), which means that Q depends only on t, x
and q. The equation that determines the zeroth order multiplier is given by
δ n  o
Q(t, x, q) qtx + qx + q 2 = 0 (3.3.1)
δq
δ
with δq being the Euler operator, which is given by

δ ∂ ∂ ∂
= − Dx + Dt Dx + ··· . (3.3.2)
δq ∂q ∂qx ∂qtx

Equation (3.3.1) leads to

Qq (qtx + qx + q 2 ) + Dx Dt (Q) − Dx (Q) + 2qQ = 0, (3.3.3)

Expanding equation (3.3.3) yields

Qq (qtx + qx + q 2 ) + (Qtx + qtx Qq + qt Qqx + Qtq qx

+ qt qx Qqq ) + 2qQ − Qx − qx Qq = 0,

which on splitting with regard to derivatives of q gives

qt qx : Qqq = 0, (3.3.4)
qtx : Qq = 0, (3.3.5)
qx : Qtq − Qq = 0, (3.3.6)
qt : Qqx = 0, (3.3.7)
2
rest : 2qQ + Qq q + Qtx − Qx = 0. (3.3.8)

We observe that equations (3.3.4), (3.3.6) and (3.3.7) are already satisfied by (3.3.5). Integrating
(3.3.5) gives
Q = A(t, x),
where A(t, x) depends on both t and x. Substituting this Q in (3.3.8) we get

2qA(t, x) + Atx (t, x) − Ax (t, x) = 0,

which splits on q and gives

q : A = 0,
rest : Atx − Ax = 0.

Since A = 0, we see that the zeroth order multiplier Q = 0.

We now look for the first order multiplier Q(t, x, q, qt , qx ), that is to say that Q also depends on the
first derivatives of q. In this case, the equation that determines the multipliers is
δ n  o
Q(t, x, q, qx , qt ) qtx + qx + q 2 = 0. (3.3.9)
δq
Section 3.3. Conserved vectors for (3.1.1) Page 19

Expanding (3.3.9) and thereafter splitting on derivatives of q yields

Qqt qx = 0,
qx Qqqt − q 2 Qqt qt − qx Qqt qt + Qxqt = 0,
2q 2 Qqt qx + 2qx Qqt qx − 2Qq + 2Qqt = 0,
Qtqx − 2Qqx + qt Qqqx − q 2 Qqx qx − qx Qqx qx = 0,
qx Qtq − qx2 Qqqx + Qtx − Qx + qx qt Qqq − q 2 qx Qqqx − 2qqx Qqx
− q 2 qt Qqqt − 2qqt Qqt − qt qx Qqqt + 2qQ + qt Qxq − qx Qxqx
− q 2 Qxqx + q 2 Qq − q 2 Qtqt − qx Qtqt = 0.

Solving the above equations, using Maple, we get the first-order multiplier to be

Q = c3 qe2t + c3 qt e2t + c2 qe3t + c2 qt e3t + c3 xqx e2t + c1 qx e2t , (3.3.10)

where, c1 , c2 and c3 are constants. Consequently we get three first order multipliers, viz.,

Q1 = qx e2t , Q2 = (q + qt ) e3t , Q3 = (q + qt + xqx ) e2t .

The conserved vectors can now determined by utilizing

n o
Dt C t + Dx C x = Q(t, x, q, qt , qx ) qtx + qx + q 2 ,

where C t denotes the conserved density, and C x is the spatial flux.

Case 1.
We first find conserved vector of (3.1.1) related to multiplier Q1 = qx e2t . Thus
 
Dt C t + Dx C x = qx e2t qtx + qx + q 2 , (3.3.11)

where C t = C t (t, x, q, qx ) and C x (t, x, q, qt ). Equation (3.3.11) then becomes

 
Ctt + qt Cqt + qtx Cqtx + Cxx + qx Cqx = qx e2t qtx + qx + q 2 .

This equation splits on second derivatives of q giving us

qtx : Cqtx = qx e2t , (3.3.12)

 
rest : Ctt + qt Cqt + Cxx + qx Cqx 2t
= qx e qx + q 2
. (3.3.13)

Equation (3.3.12) yields

1
C t = e2t qx2 + A(t, x, q), (3.3.14)
2
where A depends on t, x, q. We now substitute C t in (3.3.13) and obtain

At + qt Aq + Cxx + qx Cqx = e2t q 2 qx .

The above equation can split on qx to give

qx : Cqx = e2t q 2 , (3.3.15)

rest : At + qt Aq + Cxx = 0. (3.3.16)
Section 3.4. Summary Page 20

Equation (3.3.15) gives

1
C x = q 3 e2t + B(t, x), (3.3.17)
3
x
where B depends on t, x. Inserting C in (3.3.16) we obtain

At + qt Aq + Bx = 0,

which splits on qt to give

qt : Aq = 0, (3.3.18)
rest : At + Bx = 0. (3.3.19)

Equation (3.3.18) yields A = A(t, x) and hence

1
C t = e2t qx2 + A(t, x). (3.3.20)
2
We can take A = B = 0 since they contribute to the trivial part of the conserved vector. Therefore,
the conserved vector corresponding to the first multiplier Q1 = qx e2t is (C1t , C1x ) with components
1
C1t = e2t qx2 ,
2
x 1
C1 = e2t q 3 .
3
In the same way, the conserved vectors related to the other two multipliers Q2 and Q3 can be constructed
and are given below.
Case 2.
For multiplier Q2 = (q + qt ) e3t the associated conserved vector is (C2t , C2x ) with
1
C2t = e3t q 3 + e3t qqx ,
3
1
C2x = e3t qt2 − e3t q 2 .
2

Case 3.
Finally for third multiplier Q3 = (q + qt + xqx ) e2t we get the associated conserved vector (C3t , C3x )
given by
1 1
C3t = e2t q 3 + e2t xqx2 + e2t qqx ,
3 2
1 1 1
C3x = e2t xq 3 − e2t q 2 + e2t qt2 .
3 2 2

3.4 Summary
In this chapter we investigated the second-order Boltzmann equation (3.1.1) which arises in gas dy-
namics. Lie point symmetries were calculated and then utilized to carry out symmetry reductions
and thereafter group-invariant solutions were constructed. Also, the commutator table and the corre-
sponding one-parameter groups of transformations for these symmetries were constructed. Furthermore,
the conserved vectors for the Boltzmann equation (3.1.1) were established by invoking the multiplier
method.
4. Exact solutions and conserved vectors of the
fourth-order nonlinear Benjamin-Ono equation
4.1 Introduction.
The (1+1)-dimensional fourth-order Benjamin-Ono (BO) equation is given by Singh et al. (2020)

qtt + β(q 2 )xx + γqxxxx = 0 (4.1.1)

or

qtt + 2βqx2 + 2βqqxx + γqxxxx = 0.

It models the evolution of waves in deep water (Benjamin, 1967; Ono, 1975; Ambrose and Wilkening,
2010)
Fu et al. (2003) used the Jacobi elliptic function expansion (JEFE) to obtain the exact periodic solutions
of two kinds of nonlinear wave equations including the fourth order BO equation. It was observed that
the solutions in their limit obtained by the JEFE method are similar to the outcomes generated by
several other techniques, including the homogeneous balance method, the hyperbolic tangent expansion
approach, the nonlinear transformation method, the trial function method, and the sine-cosine method.
Zhen et al. (2005) solved for the travelling wave solutions of BO equation using improved projective
Riccati equation method. The types of solutions obtained from this study are families of Weierstrass
elliptic function solutions, families of soliton solutions and families of trigonometric function solutions.
Xu et al. (2010) investigated the new periodic solitary wave doubly periodic solutions for (4.1.1) using
the bilinear method and the homoclinic test approach. Here β and γ are non-zero constants.
In Liu and Li (2017) the authors applied both the long wave limit and bilinear Hirota linear method to
find the bright rogue waves and the dark rogue waves for the BO equation and analysed their typical
dynamics. With the application of the travelling wave method, scientists were able to obtain the exact
solutions for this equation. The BO equation transforms into the bilinear form through the dependent
variable transformation. The applications of the bilinear method together with homoclinic test approach
can be useful for deriving first order rogue waves, periodic solitary waves and doubly periodic solutions
for the BO equation.
In the paper by Alam and Li (2020) the authors studied cusp, bright breathers, dark breathers, kink,
bright rogue waves using the e−φ(ξ) -expansion method for the fourth order BO equation. The solutions
obtained might be useful for understanding physical phenomena in real world. There are several methods
that were developed by groups of physicist and mathematicians to obtain various forms of soliton
solutions for analysing nonlinear partial differential equations (NLPDEs). The new analytical solutions
performed were notable by exponential, trigonometric and hyperbolic functions.
In the paper (Alquran et al., 2022), bidirectional waves for the BO equation were studied. BO equation
is classified based on the degree of time-derivative qtt = G(q, qx , qxx , qxxx ) as the type of equation
that describes the motion of bidirectional or unidirectional waves propagating simultaneously with same
magnitude of wave speed. One of the methods implemented to solve the BO equation was the rational
sine-cosine method.

21
Section 4.2. Lie symmetries of (4.1.1) Page 22

In (Wang et al., 2022) the authors solved equation (4.1.1) and found four symmetries, but in fact, this
∂
equation has only three symmetries. The extra symmetry ∂u is not a point symmetry of fourth-order
NLPDE (4.1.1) because it did not satisfy X E E=0 = 0, where X 4 is the fourth prolongation and E is
4

the BO equation (4.1.1). The authors went on to find commutator table, group invariant solutions and
finally solved for conservation laws using Ibragimov method.

4.2 Lie symmetries of (4.1.1)

Equation (4.1.1) admits the vector field

∂ ∂ ∂
X = τ (t, x, q) + ξ(t, x, q) + η(t, x, q) (4.2.1)
∂t ∂x ∂q
on the condition that  
X [4] qtt + 2βqx2 + 2βqqxx + γqxxxx = 0, (4.2.2)
(4.1.1)

where X [4] is the fourth prolongation of (4.1.1) given by

∂ ∂ ∂ ∂ ∂ ∂ ∂
X [4] = τ +ξ +η + ζx + ζtt + ζxx + ζxxxx .
∂t ∂x ∂q ∂qx ∂qtt ∂qxx ∂qxxxx

After expanding (4.2.2), we obtain

η(2βqxx ) + ζx (4βqx ) + ζtt (1) + ζxx (2βq) + ζxxxx (γ) = 0. (4.2.3)

(4.1.1)

Substituting the values of ζx , ζtt , ζxx , and ζxxxx into equation (4.2.3), we have

γηqqqq qx4 − γτqqqq qx5 − γqt ξqqqq qx4 − 4γτxqqq qx4 − 4βτq qx3 − 2qβτqq qx3
− 4γqtx ξqqq qx3 − 10γqxx τqqq qx3 + 4γηxqqq qx3 − 4γqt ξxqqq qx3 − 6γτxxqq qx3
− 4βqt ξq qx2 + 2qβηqq qx2 − 2qβqt ξqq qx2 − 6γqtxx ξqq qx2 − 10γqxxx τqq qx2
+ 6γqxx ηqqq qx2 − 6γqxx qt ξqqq qx2 − 4βτx qx2 − 4qβτxq qx2 − 12γqtx ξxqq qx2
− 24γqxx τxqq qx2 + 6γηxxqq qx2 − 6γqt ξxxqq qx2 − 4γτxxxq qx2 − 4qβqtx ξq qx
− 4γqtxxx ξq qx − 6qβqxx τq qx − 5γqxxxx τq qx − qtt τq qx + 4γqxxx ηqq qx
2
− 4γqxxx qt ξqq qx − 12γqxx qtx ξqq qx − 15γqxx τqq qx − qt2 τqq qx + 4βηx qx
− 4βqt ξx qx + 4qβηxq qx − 4qβqt ξxq qx − 12γqtxx ξxq qx − 16γqxxx τxq qx
+ 12γqxx ηxqq qx − 12γqxx qt ξxqq qx − 2qβτxx qx − 12γqtx ξxxq qx − 18γqxx τxxq qx
+ 4γηxxxq qx − 4γqt ξxxxq qx − γτxxxx qx − 2qt τtq qx − τtt qx + 2βηqxx + 4βηq qx2
+ 2qβqxx ηq + γqxxxx ηq + qtt ηq − 2qβqxx qt ξq − γqxxxx qt ξq − 4γqxxx qtx ξq
2
− 6γqxx qtxx ξq − 3qt qtt ξq − 10γqxx qxxx τq − 2qt qtx τq + 3γqxx ηqq + qt2 ηqq
− qt3 ξqq − 3γqxx
2
qt ξqq − 4qβqtx ξx − 4γqtxxx ξx − 4qβqxx τx − 4γqxxxx τx
2
+ 4γqxxx ηxq − 4γqxxx qt ξxq − 12γqxx qtx ξxq − 12γqxx τxq + 2qβηxx − 2qβqt ξxx
− 6γqtxx ξxx − 6γqxxx τxx + 6γqxx ηxxq − 6γqxx qt ξxxq − 4γqtx ξxxx − 4γqxx τxxx
+ γηxxxx − γqt ξxxxx − 2qtt ξt − 2qtx τt + 2qt ηtq − 2qt2 ξtq + ηtt − qt ξtt = 0.
(4.1.1)
Section 4.2. Lie symmetries of (4.1.1) Page 23

Replacing qtt by −2βqx2 − 2βqqxx − γqxxxx in the above equation, we obtain

γηqqqq qx4 − γξqqqq qx5 − γqt τqqqq qx4 − 4γξxqqq qx4 − 4βξq qx3 − 2qβξqq qx3
− 4γqtx τqqq qx3 − 10γqxx ξqqq qx3 + 4γηxqqq qx3 − 4γqt τxqqq qx3 − 6γξxxqq qx3
+ 4βηq qx2 − 4βqt τq qx2 + 2qβηqq qx2 − 2qβqt τqq qx2 − 6γqtxx τqq qx2
− 10γqxxx ξqq qx2 + 6γqxx ηqqq qx2 − 6γqxx qt ξqqq
1
qx2 − 4βξx qx2 − 4qβξxq
2 2
qx
− 12γqtx τxqq qx2 − 24γqxx ξxqq qx2 + 6γηxxqq qx2 − 6γqt τxxqq qx2 − 4γξxxxq qx2
− 4qβqtx τq qx − 4γqtxxx ξq1 qx − 6qβqxx ξq qx − 5γqxxxx ξq qx − qtt ξq qx
1 2
+ 4γqxxx ηqq qx − 4γqxxx qt ξqq qx − 12γqxx qtx τqq qx − 15γqxx ξqq qx − qt2 ξqq qx
2
+ 4βηx qx − 4βqt τx qx + 4qβηxq qx − 4qβqt τxq qx − 12γqtxx τxq qx − 16γqxxx ξxq qx
+ 12γqxx ηxqq qx − 12γqxx qt τxqq qx − 2qβξxx qx − 12γqtx τxxq qx − 18γqxx ξxxq qx
+ 4γηxxxq qx − 4γqt τxxxq qx − γξxxxx qx − 2qt ξtq qx − ξtt qx + 2βηqxx
+ 2qβqxx ηq + γqxxxx ηq + qtt ηq − 2qβqxx qt τq − γqxxxx qt τq − 4γqxxx qtx τq
2
− 6γqxx qtxx τq − 3qt qtt τq − 10γqxx qxxx ξq − 2qt qtx ξq + 3γqxx ηqq + qt2 ηqq
2
− 3γqxx qt τqq − 4qβqtx τx − 4γqtxxx τx − 4qβqxx ξx − 4γqxxxx ξx + 4γqxxx ηxq
2 1
− 4γqxxx qt τxq − 12γqxx qtx τxq − 12γqxx ξxq + 2qβηxx − 2qβqt ξxx − 6γqtxx τxx
− 6γqxxx ξxx + 6γqxx ηxxq − 6γqxx qt τxxq − 4γqtx τxxx − 4γqxx ξxxx + γηxxxx
− qt3 τqq − γqt τxxxx − 2qtt τt − 2qtx ξt + 2qt ηtq − 2qt2 τtq + ηtt − qt τtt = 0.

We now split the above equation on derivatives of q and after simplifying we obtain the system of PDEs:

τu = 0, (4.2.4)
τx = 0, (4.2.5)
ξt = 0, (4.2.6)
ξq = 0, (4.2.7)
ηqq = 0, (4.2.8)
τt − 2ξx = 0, (4.2.9)
2ηxq − 3ξxx = 0, (4.2.10)
ηtt + 2βqηxx + γηxxxx = 0, (4.2.11)
2ηtq − τtt = 0, (4.2.12)
2βqτt − 2βqτx + 3γηxxq − 2γξxxx + βη − 2βqξx = 0, (4.2.13)
2βτt + βηq − 2βξx = 0, (4.2.14)
4βqηxq − 2βqξxx + 4γηxxxq + 4βηx − γξxxxx − τtt = 0. (4.2.15)

Equation (4.2.4) and (4.2.5) implies that

τ = f (t),
where f (x) depends on t only. Likewise, (4.2.6) and (4.2.7) gives

ξ = g(x),

where g(x) depends on t only. Integrating (4.2.7) twice with respect to q yields

η = h1 (t, x)q + h2 (t, x),

Section 4.2. Lie symmetries of (4.1.1) Page 24

where h1 (t, x) and h2 (t, x) depend on t, x. Substituting values of τ and ξ in (4.2.9) we see that

f 0 (t) = 2g 0 (x).

This implies that f 0 (t) = 2g 0 (x) = k1 , for some constant k1 . Hence

f 0 (t) = k1 ,
1
g 0 (x) = k1 ,
2
which gives
1
f (t) = k1 t + k2 ,
2
g(x) = k1 x + k3,

for constants k2 , k3 . Thus

τ = k1 t + k2 ,
1
ξ = k1 x + k3 .
2
Now (4.2.10) yields h1 x (t, x) = 0 and so
h1 = h1 (t),
whereby substituting the new value of η into (4.2.12) we get h1 t (t) = 0. Hence

h1 (t) = k4 ,

for k4 a constant and therefore

η = k4 q + h2 (t, x).
Substituting the values of τ , ξ and η in (4.2.13) we obtain

k1 q + k4 q + h2 (t, x) = 0

and splitting on q leads to

q : k1 + k4 = 0, (4.2.16)
0 2
q : h (t, x) = 0. (4.2.17)

Equation (4.2.16) implies that k4 = −k1 , thus

1
τ = k1 t + k2 , ξ = k1 x + k3 , η = −k1 q.
2
Note that the remaining equations (4.2.14) and (4.2.15) are satisfied by these values of τ , ξ, η and so
the infinitesimal symmetries of (4.1.1) are

∂ ∂ ∂ ∂ ∂
X1 = , X2 = , X3 = 2t + x − 2q .
∂t ∂x ∂t ∂x ∂q
Section 4.2. Lie symmetries of (4.1.1) Page 25

4.2.1 Group of transformations

In this subsection, we calculate the one-parameter Lie group of transformations for equation (4.1.1).
To determine the group of transformations, we solve the Lie equations in conjunction with the initial
conditions:
dt̄ dx̄ dq̄
= τ (t̄, x̄, q̄), t̄|a=0 = t; = ξ(t̄, x̄, q̄), x̄|a=0 = x; = η(t̄, x̄, q̄), q̄|a=0 = q.
da da da
We calculate the one-parameter group Tai which is produced by each infinitesimal generator, X sub
i, where Xi , i = 1, 2, 3, 4.. Upon solving the system outlined above for each Xi , we derive the one-
parameter transformation group as

Ta1 : t̄ = t + a1 , x̄ = x, q̄ = q,
Ta2 : t̄ = t, x̄ = x + a2 , q̄ = q,
Ta3 : t̄ = te2a3 , x̄ = xea3 , q̄ = qe−2a3 .

Since each group Tai is a symmetry group, we can deduce that if q = σ(t, x) is a solution of equation
(4.1.1), then so are the functions
 
q1 = σ(t − a1 , x), q2 = σ(t, x − a2 ), q3 = e−2a3 σ te−2a3 , xe−a3 .

4.2.2 Symmetry reductions and invariant solutions of (4.1.1)

We now perform symmetry reductions and find group invariant solutions for equation (4.1.1).
Subalgebra 1.
∂
For the symmetry X2 = ∂x the associated Lagrange equations are

dt dx dq
= = ,
0 1 0
which gives the invariants J1 = t and J2 = q. The invariant solution is q = h(t), hence we have

qtt = h00 (t), qxx = 0, qxxxx = 0. (4.2.18)

Substituting (4.2.18) into (4.1.1) gives

h00 (t) = 0.

The solution of this ODE is h(t) = C1 t + C2 and hence the solution of (4.1.1) under X2 is

q(t, x) = C1 t + C2 .

Subalgebra 2. We carry out symmetry reductions for equation (4.1.1) by employing the combination
X = X1 + cX2 with c a constant. Note that when c = 0, we get our symmetry X1 . The symmetry X
has two invariants

J1 = x − ct, J2 = q. (4.2.19)

Thus the invariant solution is

q = Φ(s), s = x − ct. (4.2.20)
Section 4.2. Lie symmetries of (4.1.1) Page 26

Substituting equation (4.2.20) into (4.1.1), we obtain the following fourth-order NLODE:

2βΦΦ00 + c2 Φ00 + 2βΦ02 + γΦ0000 = 0. (4.2.21)

Solution of (4.2.21) by Kudryashov’s method

We now solve equation (4.2.21) using Kudryashov’s approach (Kudryashov, 2012). We start by assuming
that the solution of equation (4.2.21) is of the form
M
X
Φ(s) = Ai Qi (s), (4.2.22)
i=0

where Q(s) satisfies the Riccati equation

Q0 (s) = Q2 (s) − Q(s), (4.2.23)

whose solution is given by

1
Q(s) = . (4.2.24)
1 + exp(s)
From the balancing procedure (Kudryashov, 2012) we get M = 2 for (4.2.21). Thus, the solution can
be written as
Φ(s) = A0 + A1 Q(s) + A2 Q2 (s). (4.2.25)
Substituting equation (4.2.25) into (4.2.21) and using (4.2.23), we obtain

6A2 c2 Q4 + 2A1 c2 Q3 − 10A2 c2 Q3 − 3A1 c2 Q2 + 4A2 c2 Q2 + A1 c2 Q + 20A22 βQ6

− 36A22 βQ5 + 24A1 A2 βQ5 + 6A21 βQ4 + 16A22 βQ4 + 12A0 A2 βQ4 − 42A1 A2 βQ4
− 10A21 βQ3 + 4A0 A1 βQ3 − 20A0 A2 βQ3 + 18A1 A2 βQ3 + 4A21 βQ2 − 6A0 A1 βQ2
+ 8A0 A2 βQ2 + 2A0 A1 βQ + 120A2 γQ6 + 24A1 γQ5 − 336A2 γQ5 − 60A1 γQ4
+ 330A2 γQ4 + 50A1 γQ3 − 130A2 γQ3 − 15A1 γQ2 + 16A2 γQ2 + A1 γQ = 0.

Splitting the above equation on powers of Q(s), we obtain

Q6 (s) : 2A0 A1 β + A1 c2 + A1 γ = 0,
Q5 (s) : 4A21 β − 6A0 A1 β + 8A0 A2 β − 3A1 c2 + 4A2 c2 − 15A1 γ + 16A2 γ = 0,
Q4 (s) : 2A0 A1 β − 5A21 β − 10A0 A2 β + 9A1 A2 β + A1 c2 − 10A2 c2 + 25A1 γ − 65A2 γ = 0,
Q3 (s) : 6A21 β + 16A22 β + 12A0 A2 β − 42A1 A2 β + 6A2 c2 − 60A1 γ + 330A2 γ = 0,
Q2 (s) : 24A1 A2 β − 36A22 β − 336A2 γ + 24A1 γ = 0,
Q(s) : A22 β + 6A2 γ = 0.

Solving the above algebraic equations we obtain

(c2 + γ) 6γ 6γ
A0 = − , A1 = , A2 = − . (4.2.26)
2β β β
Therefore, the general solution of (4.1.1) corresponding to the constants given by (4.2.26) is given by

6γ 6γ (c2 + γ)
q(t, x) = s
− 2 − , (4.2.27)
β (e + 1) β (es + 1) 2β
Section 4.2. Lie symmetries of (4.1.1) Page 27

where s = x − ct.
∂ ∂ ∂
Subalgebra 3. Considering X3 = 2t ∂t + x ∂x − 2q ∂q . The characteristic equations for symmetry X3
are
dt dx dq
= = .
2t x −2q
Solving the above equation we get the similarity solution and variable as
1 x
q(t, x) = ω(z), z=√ . (4.2.28)
t t
Substituting equation (4.2.28) into equation (4.1.1) we get the NLODE

4γω 0000 + 8βω 00 ω + 8ω + 8βω 02 + z 2 ω 00 + 7zω 0 = 0. (4.2.29)

Solution of (4.2.29) by the power series solution method. In this section, we employ the power
series solution method (Khalique and Simbanefayi, 2021; Liu et al., 2009) to address the differential
equation (DEQ) (4.2.29). The utilization of power series methods has demonstrated their effectiveness
in tackling intricate differential equations, spanning a broad spectrum from challenging semi-linear to
intricate nonlinear ones. Therefore, the primary objective of this section is to derive the analytical
solution for equation (4.1.1) by means of the power series technique, complemented by algebraic and
differential transformations. Assume the formal solution to be of the form
∞
X
ω(z) = dn z n , (4.2.30)
n=0

where di (i ≥ 0) are constants to be determined. Substituting equation (4.2.30) into equation (4.2.29),
we have
∞
X
96γd4 + 480γd5 z + 4γ (n + 1)(n + 2)(n + 3)(n + 4)dn+4 z n + 16βd0 d2
n=2
∞ X
" n #
X
+ 48βd1 d3 z + 8β (n − k + 1)(n − k + 2)dk dn−k+2 z n + 8d0 + 8d1 z
n=2 k=0
∞ ∞ X
" n #
X X
n
+8 dn z + 8βd21 + 32βd22 z + 8β (k + 1)(n − k + 1)dk+1 dn−k+1 z n
n=2 n=2 k=0
∞
X ∞
X
+ n(n − 1)dn z n + 7d1 z + 7 ndn z n = 0. (4.2.31)
n=2 n=2

When comparing the coefficients of z in equation (4.2.31), one gets

12γd4 + βd21 + 2βd0 d2 + d0 = 0, for n = 0, (4.2.32)

and
480γd5 + 32βd22 + 8d1 + 7d1 + 48βd1 d3 = 0, for n = 1. (4.2.33)
In view of equation (4.2.31) for n ≥ 2 the recursion relation is
n
"
1 X
dn+4 =− 8β (n − k + 1)(n − k + 2)dk dn−k+2
4γ(n + 1)(n + 2)(n + 3)(n + 4) k=0
n
#
X
+8dn + 8β (k + 1)(n − k + 1)dk+1 dn−k+1 + n(n − 1)dn + 7ndn . (4.2.34)
k=0
Section 4.3. Conservation laws of (4.1.1) via the multiplier method. Page 28

Utilizing equations (4.2.32), (4.2.33) and (4.2.34), it is possible to determine all the coefficients dn (n ≥
2) of the power series (4.2.30). For arbitrarily selected constant values d0 , d1 , d2 , d3 , d4 , and d5 the
remaining terms can also be determined in a unique manner through successive application of equations
(4.2.32), (4.2.33) and (4.2.34). Thus the power series solution of equation (4.2.29) can be written as
∞
X
ω(z) = d0 + d1 z + d2 z 2 + d3 z 3 + d4 z 4 + d5 z 5 + dn+4 z n+4
n=2
(d0 + βd21 + 2βd0 d2 ) 4 (32βd22 + 48βd1 d3 + 15d1 ) 5
= d0 + d1 z + d2 z 2 + d3 z 3 − z − z
12γ 480γ
∞ n
"
X 1 X
− 8β (n − k + 1)(n − k + 2)×
n=2
4γ(n + 1)(n + 2)(n + 3)(n + 4) k=0
n
X
dk dn−k+2 + 8dn + 8β (k + 1)(n − k + 1)dk+1 dn−k+1 + n(n − 1)dn
k=0
+7ndn ] z n+4 ,

where the extended power series of the BO equation (4.1.1) is

(d0 + βd21 + 2βd0 d2 ) 4 (32βd22 + 48βd1 d3 + 15d1 ) 5

q(t, x) = d0 + d1 z + d2 z 2 + d3 z 3 − z − z
12γ 480γ
∞ n
"
X 1 X
− 8β (n − k + 1)(n − k + 2)
n=2
4γ(n + 1)(n + 2)(n + 3)(n + 4) k=0
n
X
× dk dn−k+2 + 8dn + 8β (k + 1)(n − k + 1)dk+1 dn−k+1 + n(n − 1)dn
k=0
+7ndn ] z n+4 .
x
Here di (i = 0, 1, 2, . . . 5) are arbitrary constants and z = √ . In physical applications we consider the
t
approximate form

(d0 + βd21 + 2βd0 d2 ) 4

q(t, x) = d0 + d1 z + d2 z 2 + d3 z 3 − z
12γ
(32βd22 + 48βd1 d3 + 15d1 ) 5
− z + ··· . (4.2.35)
480γ

4.3 Conservation laws of (4.1.1) via the multiplier method.

The multiplier approach is admired for its ability to handle conserved quantities of DEs, whether they
possess variational principles or not. We employ the multiplier technique to derive conserved vectors for
the BO (4.1.1). We examine zeroth-order multipliers, viz., Λ = Λ(t, x, q), and employ these multipliers
to establish the conservation laws for equation (4.1.1). It is worth noting that the zeroth-order multipliers
Λ for this equation are determined based on the governing equation

δ n  o
Λ qtt + 2βqx2 + 2βqqxx + γqxxxx = 0. (4.3.1)
δq
Section 4.4. Concluding remarks. Page 29

Expansion of (4.3.1) and following a procedure similar to the algorithm used in Lie symmetry analysis,
we derive six determining equations which upon simplification are

Λq = 0, Λtt = 0, Λxx = 0.

Solving the above system we get

Λ = (C1 x + C2 ) t + C3 x + C4 , (4.3.2)

here C1 , C2 , C3 and C4 represent constants. To obtain the necessary conserved quantities, we identify
them using the divergence identity provided as follows
 
Dt T t + Dx T x = Λ qtt + 2βqx2 + 2βqqxx + γqxxxx , (4.3.3)

where T t is the conserved density and T x is the spatial flux Olver (1993). Thus, the conserved vectors
associated with the four multipliers are presented below.
Case 1. For the first multiplier Λ1 = tx, the corresponding conserved vector is

T1t = txqt − xq,

T1x = 2 β txqqx − βtq 2 + γtxqxxx − γtqxx .

Case 2. For the second multiplier Λ2 = t, the corresponding conserved vector is

T2t = tqt − q,
T2x = 2βtqqx + γtqxxx .

Case 3. For the multiplier Λ3 = x, we obtain the conserved vector

T3t = xqt ,
T3x = 2βxqqx − βq 2 + γxqxxx − γqxx .

Case 4. The corresponding conserved vector for the multiplier Λ4 = 1 is

T4t = qt ,
T4x = 2βqqx + γqxxx .

4.4 Concluding remarks.

In this study, we examined the fourth-order nonlinear Benjamin-Ono equation. Our approach involved the
utilization of Lie symmetry analysis of differential equations, in conjunction with Kudryashov’s method,
the simplest equation technique, and the power series method. These methods allowed us to derive
solutions for the equation in the form of rational, exponential, and hyperbolic functions. Additionally,
we determined the conserved vectors of the Benjamin-Ono equation using the multiplier method.
5. Conlusion
In this thesis, we have studied two NLPDEs, the Boltzmann and BO equations, using Lie symmetry
analysis and presented exact solutions for these PDEs. In the third chapter, we provided an illustra-
tive example by performing Lie symmetry analysis for the Boltzmann equation and also obtained the
conserved vectors for this equation.
In the fourth chapter, which is an important part of this thesis because it is the foundation upon which
this thesis is based, we began by solving for Lie point symmetries of the fourth-order BO equation. In this
process, we obtained three symmetries: time translation, space translation, and scale transformation.
We then found the group invariant solutions of the space translation, scale transformation together with
travelling wave solution, the time translation was left out because it is covered under travelling wave
solution when the coefficient of space translation is zero. Higher-order differential equations are difficult
to integrate, we got fourth-order ODEs from the travelling wave solution and scale transformation which
are difficult to produce exact solutions. But it was possible to obtain an exact travelling wave solution
with the application of Kudryashov’s method. We also applied a power series to obtain an approximate
invariant solution for the scale transformation. Finally we have obtained the four zero-th order multipliers
and used them to obtain the conserved vectors for the BO equation, where all those vectors are local.

Future work
In future I will use the Ibragimov’s method to find the conservation laws.

30
Acknowledgements
First and foremost, I would like to express my gratitude to the God Almighty for paving the way to the
success of this thesis and for providing me with invaluable support from well-wishers throughout this
dissertation. I also wish to extend my appreciation to AIMS and its contributors for their unwavering
support for this endeavour. Additionally, I must acknowledge my supervisor, Professor Chaudry Masood
Khalique of North West University, whose dedication and support towards his students have been truly
commendable. I am also indebted to all the tutors who collaborated with me during the entire duration
of my research project. Finally, I want to convey my heartfelt thanks to my family for their love and
support, especially during the challenging moments when I was fully engaged in this research project.

31
References
Alam, M. N. and Li, X. Symbolic methods to construct a cusp, breathers, kink, rogue waves and
some soliton waves solutions of nonlinear partial differential equations. Computational Methods for
Differential Equations, 8(3):597–609, 2020.

Alquran, M., Ali, M., and Jadallah, H. New topological and non-topological unidirectional-wave solutions
for the modified-mixed KdV equation and bidirectional-waves solutions for the Benjamin Ono equation
using recent techniques. Journal of Ocean Engineering and Science, 7(2):163–169, 2022.

Ambrose, D. M. and Wilkening, J. Computation of time-periodic solutions of the Benjamin-Ono equa-

tion. Journal of nonlinear science, 20(3):277–308, 2010.

Benjamin, T. B. Internal waves of permanent form in fluids of great depth. Journal of Fluid Mechanics,
29(3):559–592, 1967.

Bluman, G. W. and Kumei, S. Symmetries and differential equations, volume 81. Springer Science &
Business Media, 2013.

Bluman, G. W., Cheviakov, A. F., and Anco, S. Construction of conservation laws: how the direct
method generalizes Noether’s theorem. In Proceedings of 4th Workshop “Group Analysis of Differential
Equations & Integrability, volume 1, pages 1–23, 2009.

Caflisch, R. E. The fluid dynamic limit of the nonlinear Boltzmann equation. Communications on Pure
and Applied Mathematics, 33(5):651–666, 1980.

Cheviakov, A. F. Computation of fluxes of conservation laws. Journal of Engineering Mathematics, 66:

153–173, 2010.

Ernst, M. H. Nonlinear model-Boltzmann equations and exact solutions. Physics Reports, 78(1):1–171,
1981.

Euler, N., Leach, P., Mahomed, F., and Steeb, W. H. Symmetry vector fields and similarity solutions
of a nonlinear field equation describing the relaxation to a Maxwell distribution. International journal
of theoretical physics, 27:717–723, 1988.

Fu, Z., Liu, S., Liu, S., and Zhao, Q. The JEFE method and periodic solutions of two kinds of nonlinear
wave equations. Communications in Nonlinear Science and Numerical Simulation, 8(2):67–75, 2003.

Hong, J. B. Z., Han, C. T., Rasid, N. S. M., et al. Lie symmetry analysis on Benjamin-Ono equation.
In Journal of Physics: Conference Series, volume 1593, page 012003. IOP Publishing, 2020.

Hydon, P. E. Symmetry analysis of initial-value problems. Journal of mathematical analysis and appli-
cations, 309(1):103–116, 2005.

Ibragimov, N. H. CRC handbook of Lie group analysis of differential equations, volume 3. CRC press,
1995.

Ibragimov, N. K. Elementary Lie group analysis and ordinary differential equations.

Kara, A. H. and Mahomed, F. M. Relationship between symmetries and conservation laws. International
Journal of Theoretical Physics, 39:23–40, 2000.

32
REFERENCES Page 33

Kara, A. H. and Mahomed, F. M. A basis of conservation laws for partial differential equations. Journal
of Nonlinear Mathematical Physics, 9(sup2):60–72, 2002.
Khalique, C. M. and Mehmood, A. On the solutions and conserved vectors for the two-dimensional
second extended Calogero-Bogoyavlenskii-Schiff equation. Results in Physics, 25:104–194, 2021.
Khalique, C. M. and Simbanefayi, I. Conserved quantities, optimal system and explicit solutions of a (1+
1)-dimensional generalised coupled mkdv-type system. Journal of Advanced Research, 29:159–166,
2021.
Krook, M. and Wu, T. T. Formation of Maxwellian tails. Physical Review Letters, 36(19):1107, 1976.
Kudryashov, N. A. One method for finding exact solutions of nonlinear differential equations. Commu-
nications in Nonlinear Science and Numerical Simulation, 17(6):2248–2253, 2012.
Liu, H., Li, J., and Zhang, Q. Lie symmetry analysis and exact explicit solutions for general Burgers’
equation. Journal of Computational and Applied Mathematics, 228(1):1–9, 2009.
Liu, Y.-K. and Li, B. Dynamics of rogue waves on multisoliton background in the Benjamin Ono
equation. Pramana, 88:1–6, 2017.
Mahmud, F., Samsuzzoha, M., and Akbar, M. A. The generalized Kudryashov method to obtain exact
traveling wave solutions of the PHI-four equation and the Fisher equation. Results in Physics, 7:
4296–4302, 2017.
Manafian, J., Alimirzaluo, E., and Nadjafikhah, M. Conservation laws and exact solutions of the (3+ 1)
(3 + 1)-dimensional Jimbo–Miwa equation. Advances in Difference Equations, 2021(1):1–13, 2021.
Olver, P. J. Applications of Lie groups to differential equations, volume 107. Springer Science & Business
Media, 1993.
Ono, H. Algebraic solitary waves in stratified fluids. Journal of the Physical Society of Japan, 39(4):
1082–1091, 1975.
Ovsiannikov, L. V. Group analysis of differential equations. Academic press, 2014.
Roudenko, S., Wang, Z., and Yang, K. Dynamics of solutions in the generalized Benjamin-Ono equation:
a numerical study. Journal of Computational Physics, 445:110–570, 2021.
Shakeel, M. and Mohyud-Din, S. T. New (G0 /G)-expansion method and its application to the Zakharov-
Kuznetsov–Benjamin-Bona-Mahony (ZK–BBM) equation. Journal of the Association of Arab Uni-
versities for Basic and Applied Sciences, 18:66–81, 2015.
Singh, S., Sakkaravarthi, K., Murugesan, K., and Sakthivel, R. Benjamin-ono equation: Rogue waves,
generalized breathers, soliton bending, fission, and fusion. The European Physical Journal Plus, 135
(10):1–17, 2020.
Wang, Z., Sun, L., Hua, R., Zhang, L., and Wang, H. Lie symmetry analysis, particular solutions and
conservation laws of Benjiamin Ono equation. Symmetry, 14(7):1315, 2022.
Xu, Z.-h., Xian, D.-q., and Chen, H.-l. New periodic solitary-wave solutions for the Benjiamin Ono
equation. Applied Mathematics and Computation, 215(12):4439–4442, 2010.
Zhen, W., De-Sheng, L., Hui-Fang, L., and Hong-Qing, Z. A method for constructing exact solutions
and application to Benjamin Ono equation. Chinese Physics, 14(11):2158, 2005.