August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

International Journal of Computational Methods

Vol. 18, No. 1 (2021) 2050018 (23 pages)
c World Scientific Publishing Company
DOI: 10.1142/S0219876220500188

Approximate Solution of Homogeneous and Nonhomogeneous

5αth-Order Space-Time Fractional KdV Equations

Swapan Biswas∗ and Uttam Ghosh†,‡

∗PrabhuJagatbandhu College
Howrah, West Bengal, India
†Department of Applied Mathematics

University of Calcutta, Kolkata, India

‡uttam math@[Link]

Received 6 September 2019

Revised 21 February 2020
Accepted 28 February 2020
Published 30 June 2020

In this paper, the semi-inverse method is applied to derive the Lagrangian of the 5αth
Korteweg de Vries equation (KdV). Then the time and space differential operators of
the Lagrangian are replaced by corresponding fractional derivatives. The variation of
the functional of this Lagrangian is devoted to lead the fractional Euler Lagrangian via
Agrawal’s method, which gives the space-time fractional KdV equation. Jumarie deriva-
tive is used to obtain the space-time fractional KdV equations. The homotopy analysis
method (HAM) is applied to solve the derived space-time fractional KdV equation.
Then numerical solutions are compared with the known analytical solutions by tables
and figures.

Keywords: Homotopy analysis method; Caputo fractional derivatives; KdV equation and
approximate solution.

1. Introduction
In recent years, differential equation of fractional order has been applied in various
fields of basic science and engineering. Its application includes different fundamen-
tal branches such as mathematical biology [Carmichael et al. (2015); Kilbas et al.
(2006); Magin (2006)], electromagnetic theory [Engheta (1996, 1995, 1996)], fluid
mechanics [Schneider and Wyss (1989)], signal processing [Chen (2005)], engineer-
ing [Magin (2006)], theoretical physics [Das et al. (2018)], etc. At the same time, the
solution of the fractional differential equation plays an important role in fractional
calculus. Most of the fractional differential equations do not have exact analytical
solutions. Therefore, finding approximation or numerical solution of such equations
is of great challenge in mathematics. Adomian decomposition method [Adomian

‡ Corresponding author.

2050018-1
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

(1994, 1992); Adomian and Rach (1992)], variational iteration method [Adomian
(1992); He (1997, 1998); Inokuti et al. (1978); He (2000)], homotopy analysis method
(HAM) [Liao (1992, 2003); Abbasbandy (2006); He (2005); Motsa (2012)], hamilto-
nian approach Taghipour et al. (2014), homotopy pertrubation method [He (1999);
Guo et al. (2013)], travelling wave transform method [Guo et al. (2012)], descretiza-
tion method [Guo et al. (2018, 2019); Khader (2015); He and Ma (2010)], meshfree
method [Gu (2005); Gu et al. (2011); Garg and Pant (2018)], and the fractional
complex transformation method [Gepreel and Omran (2011)] are relatively new
approaches to provide approximation solution of linear and nonlinear fractional dif-
ferential equation. In 1992, Liao [1992] employed the basic ideas of the homotopy in
topology to propose method for non-linear problems, namely HAM. Furthermore,
HAM always provides us a family of solutions involving an auxiliary parameter h,
which adjust and control the convergence region and rate of convergence of the
series solution. Moreover, by means of the so-called h-curve, a valid region of h can
be studied to gain a convergent series solution.
In 1834, John Russell was the first to observe the solitary wave. He observed a
large protrusion of water slowly traveling on the Edinburgh–Gkasgow canal without
change in shape. The remarkable discovery motivated Russell to conduct physical
laboratory experiments and derived the formula
c2 = g(h + a),
where c is the speed of the wave, a is the maximum amplitude above the water
surface, h is the depth and g is the acceleration of gravity. In 1895, Diederik Johannes
Korteweg together with his Ph.D. student Gustav de Vries analytically derived a
nonlinear partial differential equation, well known now as the KdV equation of
shallow water. The space-time fractional KdV equations can be used to analyze
the effect of the higher-order wave dispersion. The advantage of using fractional
derivatives in partial differential equations while modeling the real world problems
is the nonlocal property, and this means that the next state of the system depends
not only on its present state, but also on all of its past states.
J. H. He [He (1997, 2004)] proposes Semi-inverse method which can be exerted
for analyzing different nonlinear differential equations. The semi-inverse method
is applied for constructing the variational principles from the field equations and
boundary conditions.
In this paper, we have applied the homotopy analysis method for solving the
5αth-order space-time fractional order KdV equations which arises from the prob-
lems in mathematical physics.
This paper is organized as follows: In Sec. 2, we describe some necessary defi-
nitions and mathematical preliminaries which is applied to the fractional calculus.
In Sec. 3, a short idea of homotopy analysis method is described. The space-time
fraction homogeneous and nonhomogeneous KdV equations are derived using He’s
semi-inverse method [He (1997, 2004)] in Sec. 4. In Secs. 5 and 6, as application, the
HAM is applied to construct approximation solution of the 5αth-order homogeneous

2050018-2
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

space-time fractional KdV equation [Goswami et al. (2018)]

Dtα u + uDxα u − uDx3α u + u5α

x = 0, −∞ < x < ∞, t > 0, 4 < 5α < 5 (1)

with initial condition u(x, 0) = Eα (xα ) and 5αth-order nonhomogeneous space-time

fractional KdV equation [Goswami et al. (2018)]

t2
Dtα u − uDxα u + u5α α α
x = cosα (x ) + 2t sinα (x ) + sinα (2xα ),
2
−∞ < x < ∞, t > 0, 4 < 5α < 5, (2)

with initial condition u(x, 0) = 0.

Also, our numerical results illustrated and demonstrated the accuracy of the
method. Finally, in Sec. 7, some conclusions are given.

2. Prelimineries and Notation

In this section, we provide some basic definitions which are listed as follows:

Definition 2.1 ([Kilbas et al. (2006)]). The left and right Riemann–Liouville
integrals with order α ∈ R+ of the given function f (t),t ∈ (a, b) are defined as
 t
−α 1
a Dt = (t − u)α−1 f (u)du,
Γ(α) a
 b
−α 1
and t Db = (u − t)α−1 f (u)du,
Γ(α) t

respectively, where Γ(.) is the Euler’s gamma function.

Definition 2.2 ([Kilbas et al. (2006)]). The left and right Riemann–Liouville
derivatives with order α ∈ R+ of the given function f (t),t ∈ (a, b) are defined as

α dm −(m−α)
a Dt f (t) = m a Dt f (t)
dt
 t
1 dm
= (t − u)m−α−1 f (u)du
Γ(m − α) dtm a

and

α dm −(m−α)
t Db f (t) = (−1)m t Db f (t)
dtm
 b
1 dm
= (−1)m (u − t)m−α−1 f (u)du,
Γ(m − α) dtm t
+
respectively, where m − 1 < α ≤ m ∈ Z .

2050018-3
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

Definition 2.3 ([Kilbas et al. (2006)]). The left and right Caputo derivatives
with order α ∈ R+ of the given function f (t),t ∈ (a, b) are defined as

C α −(m−α) dm
a Dt f (t) = a Dt f (t)
dtm
 t
1
= (t − u)m−α−1 f m (u)du,
Γ(m − α) a

and

C α −(m−α) dm
t Db f (t) = t Db f (t)
dtm
 b
1
= (u − t)m−α−1 f m (u)du,
Γ(m − α) t

+
where m − 1 < α < m ∈ Z

The main advantage of the Caputo’s approach is that the Caputo derivative for a
constant (C) is zero, while the Riemann–Liouville fractional derivative for a constant
(t−a)−α
(C) is not zero but equal to RL α
a Dt C = C Γ(1−α) .

Definition 2.4 ([Guo et al. (2012)]). The Jumarie derivatives with order α ∈
R+ of the given function f (t),t ∈ (a, b) are defined [Jumarie (2009)] as f (t),t ∈ (a, b)
are defined as

J α dm −(m−α)
a Dt f (t) = D f (t)
dtm t
 t
1 dm
= (t − u)m−α−1 [f (u) − f (0)]du.
Γ(m − α) dtm 0

Definition 2.5. Fractional derivative of the product of two integrable functions

f (t) and g(t) of fractional order α > 0 is denoted and defined by
∞

α
a Dt [f (t)g(t)] = α
Cn f n (t)a Dtα−n g(t).
n=1

Mittag–Leffler functions: Mittag–Leffler functions are denoted by Eα (z) and is

defined by
∞
 z αk
Eα (z α ) = , α ∈ C, R(α) > 0.
Γ(αk + 1)
k=0

The function Eα (z α ) was defined and studied by Mittag–Leffler in the year 1903.
It is a direct generalization of the exponential function. For α = 1, we have the
exponential function ex . The fractional sine and cosine functions [Banerjee et al.

2050018-4
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

(2017)] defined by Mittag–Leffler function are as follows:

∞
Eα (itα ) + Eα (−itα )  t2kα
cosα (tα ) = = (−1)k ,
2 Γ (2kα + 1)
k=1
∞
Eα (it ) − Eα (−it ) 
α α
t(2k+1)α
sinα (tα ) = = (−1)k .
2i Γ (2kα + α + 1)
k=1

3. Basic Idea of the Homotopy Analysis Method

The HAM was introduced by S. J. Liao [Liao (1992, 2003)] in 1992. To describe the
basic idea of HAM, we consider the following fractional differential equation:
N F [u(x, t)] = 0, (3)
where N F is a nonlinear operator, x is a spatial independent variable, t denotes
the temporal independent variable and u(x, t) is an unknown function. By means
of generalizing the traditional homotopy method, Liao [1992] constructs the zeroth-
order deformation equation, i.e.
(1 − α(q))L[φ(x, t; q) − u0 (x, t)] = β(q)hH(x, t)N F [φ(x, t; q)], (4)
where q ∈ [0, 1] is the embedding parameter, h = 0 is an auxiliary parameter, L is
an auxiliary linear operator, φ(x, t; q) is an unknown function, u0 (x, t) is an initial
guess of u(x, t) and H(x, t) denotes a non-zero auxiliary function. We assume that
the analytic function α(q) and β(q) satisfying α(0) = β(0) = 0 , α(1) = β(1) = 1
∞ m
∞ m
with α(q) ∼ m=1 αm q , β(q) ∼ m=1 βm q are convergent for |q| ≤ 1, i.e.
∞ ∞
α(1) = Σk=1 αk = 1, β(1) = Σk=1 βk = 1, where αk and βk are constants. It is clear
that when q = 0 and q = 1,
φ(x, t; 0) = u0 (x, t) and φ(x, t; 1) = u(x, t). (5)
Thus, as the embedding parameter q increases from 0 to 1, then the solution φ(x, t; q)
varies from the initial guess u0 (x, t) to the solution u(x, t) of Eq. (3).
If the auxiliary linear operator L, the initial guess u0 (x, t), the auxiliary param-
eter h and the auxiliary function H(x, t) are properly chosen, then the homotopy
series
∞

φ(x, t; q) = u0 (x, t) + um (x, t)q m (6)
m=1

is convergent at q = 1. Then we have the homotopy series solution

∞

φ(x, t; q) = u0 (x, t) + um (x, t), (7)
m=1

where

1 ∂ m φ(x, t; q) 
um (x, t) =  . (8)
m! ∂q m q=0

2050018-5
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

In practice, the M th-order homotopy approximation

M

φ(x, t; 1) = u0 (x, t) + um (x, t),
m=1
(9)
M

u(x, t) = u0 (x, t) + um (x, t).
m=1

Let us define the vector

−
u→
n = {u0 (x, t), u1 (x, t), . . . , un (x, t)}. (10)

Differentiating the zeroth-order deformation equation (4) m-times with respect to

the embedding parameter q and then setting q = 0 and finally dividing them by
m!, we have the so-called mth-order deformation equation
m−1
 m

L[um (x, t) − αm−n un (x, t)] = hH(x, t) βn Dm−n N F [φ(x, t; q)], (11)
m n=1

where

∂ m−n N F [φ(x, t; q)] 
Rm (−
→ 1
u m−n ) = Dm−n N F [φ(x, t; q)] =  . (12)
(m − n)! ∂q m−n q=0

Applying L−1 on both sides of Eq. (11), we get

m−1
 m−1
 ti
um (x, t) = αm−n un (x, t) − αm−n uim−1 (0)
n=1 i=0
i!
m

+h βn L−1 [H(x, t)Rm (−
→
u m−n )]. (13)
n=1

When α(q) = β(q) = q, then Eq. (13) reduces to

m−1
 ti
um (x, t) = χm um−1 (x, t) − χm uim−1 (0)
i=0
i!

+ hL −1
[H(x, t)Rm (→
−
u m−1 )], (14)

where χm is the characteristic function defined by


0 m≤1
χm = . (15)
1 m>1

The mth-order deformation equation is linear and can be easily solved.

2050018-6
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

4. Deduction of Fractional Differential Equation

1. Deduction of 5αth-order space-time fractional order KdV equation
(STFKdV)
Consider the regular homogeneous KdV equation of fifth order [Goswami et al.
(2018)] as

ut + uux − uuxxx + uxxxxx = 0, t ∈ T and x ∈ R, (16)

where u(x, t) is a field variable, x ∈ R is a space cordinate in the propagation

direction of the field and t ∈ T is the time variable. Using a potential function
v(x, t) where u(x, t) = vx (x, t) gives the potential equation of (16) as

vxt (x, t) + u(x, t)vxx (x, t) − u(x, t)vxxxx (x, t) + vxxxxxx = 0. (17)

The subscripts denote the partial differentiation of the function with respect to
the parameter. The Lagrangian of the KdV equation (16) can be defined using the
semi-inverse method [He (1997, 2004)] as follows: The corresponding function of the
potential equation (17) is given as follows:
 
J(v) = (c1 vxt (x, t) + c2 u(x, t)vxx (x, t) + c3 u(x, t)vxxxx (x, t)
R T

+ c4 vxxxxxx(x, t))v(x, t)dxdt, (18)

where the unknown constants ci , (i = 1, 2, 3, 4) can be determined by taking the

variation of the function (18) with respect to v(x, t), assuming that u(x, t) is a
constant function, integrating by parts using vt |R = vx |R = vxx |R = vxxx |R = 0 .
Optimizing the variation of function J(v), i.e. δJ(v) = 0 gives

2c1 vxt (x, t) + 2c2 u(x, t)vxx (x, t) + 2c3 u(x, t)vxxxx (x, t)
+ 2c4 vxxxxxx (x, t) = 0. (19)

Comparing the resultant (19) and (17), the unknown constants become c1 = c2 =
−c3 = c4 = 12 .
The Lagrangian of the potential form of (16) is given as follows:
1 1
L(vt , vx , vxx , vxxx ) = − vx (x, t)vt (x, t) − u(x, t)vx2 (x, t)
2 2
1 2 1 2
− u(x, t)vxx (x, t) − vxxx (x, t). (20)
2 2
To generate the 5αth-order space-time fractional KdV equation, replace the deriva-
tive term in Eq. (20) by the corresponding fractional derivative term, which gives
1 1
Lα (Dtα v, Dxα v, Dx2α v, Dx3α v) = − Dtα v(x, t)Dxα v(x, t) − u(x, t)[Dxα v(x, t)]2
2 2
1 1
− u(x, t)[Dx2α v(x, t)]2 − [Dx3α v(x, t)]2 . (21)
2 2
2050018-7
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

Let the functional of the STFKdV equation is of the following form:

 
Jα (v) = Lα (Dtα v, Dxα v, Dx2α v, Dx3α v)(dxdt)α . (22)
R T

Optimizing this variation of the function Jα (v), i.e. δJα (v) = 0 gives the Euler–
Lagrangian equation in the form (deduction is given in Appendix A.2)
       
∂Lα ∂Lα ∂Lα ∂Lα
− Dtα − D α
x + D 2α
x − D 3α
x = 0. (23)
∂Dtα v ∂Dxα v ∂Dx2α v ∂Dx3α v
By simplificaion
1 α α 1
D D v(x, t) + Dxα Dtα v(x, t) + u(x, t)Dx2α v(x, t) − u(x, t)Dx4α v(x, t)
2 t x 2
+ Dx6α v(x, t) = 0. (24)
Using the commutativity of fractional differential operator Dtα and Dxα , we get
Dtα Dxα = Dxα Dtα [Banerjee et al. (2017)] and substituting for the potential function
Dxα v(x, t) = u(x, t) in Eq. (24), we get
Dtα u + uDxα u − uDx3α u + Dx5α u = 0
which is the space-time fractional order homogeneous KdV equation.

2. Deduction of 5αth-order space-time fractional nonhomogeneous KdV

equation
Consider the regular nonhomogeneous KdV equation of fifth order [Goswami et al.
(2018)] as
ut − uux + uxxxxx = f (x, t), t ∈ T, x ∈ R, (25)
2
where f (x, t) = cos(x) + 2t sin(x) + t2 sin(2x), u(x, t) is a field variable, x ∈ R is
a space coordinate in the propagation direction of the field and t ∈ T is the time
variable. Using a potential function v(x, t) where u(x, t) = vx (x, t), the potential
equation of (25) is given as follows:
vxt (x, t) − u(x, t)vxx (x, t) + vxxxxxx − f (x, t) = 0. (26)
The subscripts denote the partial differentiation of the function with respect to
the parameter. The Lagrangian of the nonhomogeneous KdV equation (25) can be
defined using the semi-inverse method [He (1997, 2004)]. The corresponding function
of the potential equation (26) is given as follows:
 
J(v) = (c1 vxt (x, t) + c2 u(x, t)vxx (x, t) + c3 vxxxxxx(x, t)
R T

+ c4 f (x, t))v(x, t)dxdt, (27)

where the unknown constants ci , (i = 1, 2, 3, 4) can be determined by taking the
variation of the function (27) with respect to v(x, t), assuming that u(x, t) is a

2050018-8
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

constant function, integrating by parts using vt |R = vx |R = vxx |R = vxxx |R = 0 .

Optimizing the variation of function J(v), i.e. δJ(v) = 0 gives
2c1 vxt (x, t) + 2c2 u(x, t)vxx (x, t) + 2c3 vxxxxxx(x, t) + c4 f (x, t) = 0. (28)
Comparing the resultant (28) and (26), the unknown constants become c1 = −c2 =
c3 = 12 , c4 = −1 The Lagrangian of the potential form of (25) is given as follows:
1 1 1 2
L(vt , vx , vxxx ) = − vx (x, t)vt (x, t) + u(x, t)vx2 (x, t) − vxxx (x, t)
2 2 2
− v(x, t)f (x, t). (29)
To generate the 5αth-order space-time fractional KdV equation, replace the deriva-
tive term in Eq. (29) by the corresponding fractional derivative term, which gives
1 1
Lα (Dtα v, Dxα v, Dx3α v) = − Dtα v(x, t)Dxα v(x, t) + u(x, t)[Dxα v(x, t)]2
2 2
1 3α
− [Dx v(x, t)]2 − v(x, t)g(x, t), (30)
2
2
where g(x, t) = cosα (xα ) + 2t sinα (xα ) + t2 sinα (2xα ).
Let the functional of the nonhomogeneous STFKdV equation is of the following
form:
 
Jα (v) = Lα (v, Dtα v, Dxα v, Dx3α v)(dxdt)α . (31)
R T

Optimizing this variation of the function Jα (v), i.e. δJα (v) = 0 gives the Euler–
Lagrangian equation in the form (deduction is given in Appendix A.2)
     
∂Lα ∂Lα ∂Lα ∂Lα
− Dtα − D α
x − D 3α
x = 0. (32)
∂v ∂ Dtα v ∂ Dxα v ∂Dx3α v
By simplification, we get
1 α α 1
Dt Dx v(x, t) + Dxα Dtα v(x, t) − u(x, t)Dx2α v(x, t) + Dx6α v(x, t) = g(x, t). (33)
2 2
Using the commutativity of the fractional differential operator Dtα and Dxα , we get
Dtα Dxα = Dxα Dtα [Banerjee et al. (2017)]. By substituting for the potential function
Dxα v(x, t) = u(x, t) and the function g(x, t) in Eq. (33), we get
t2
Dtα u − uDxα u + Dx5α u = cosα (xα ) + 2t sinα (xα ) +
sinα (2xα ),
2
which is the space-time fractional order nonhomogeneous KdV equation.

5. Approximate Solution of the 5αth-order Space-Time Fractional

KdV Equation
In this section, to demonstrate the effectiveness of our approach, we will apply
Homotopy analysis method to construct approximate solution of the space-time

2050018-9
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

fractional KdV equation [Goswami et al. (2018)].

Dtα u + uDxα u − uDx3α u + u5α
x = 0, −∞ < x < ∞, t > 0, 4 < 5α < 5 (34)
α
with initial condition u(x, 0) = Eα (x ).
To solve this problem by means of the homotopy analysis method, we choose an
auxiliary linear operator
∂ α φ(x, t; q)
L[φ(x, t; q)] = (35)
∂tα
with property L(c) = 0, where c is a constant.
we define the nonlinear operator Nφ using (34) in the form
∂ α φ(x, t; q) ∂ α φ(x, t; q) ∂ 3α φ(x, t; q) ∂ 5α φ(x, t; q)
Nφ = α
+ φ(x, t; q) α
− φ(x, t; q) + .
∂t ∂x ∂x3α ∂x5α
(36)
We construct the zeroth-order deformation equation
(1 − q)L[φ(x, t; q) − u0 (x, t)] = qhH(x, t)N [φ(x, t; q)]. (37)
Obviously, when q = 0 and q = 1
φ(x, t; 0) = u0 (x, t) = Eα (z α ), φ(x, t; 1) = u(x, t). (38)
Thus, we obtain the mth-order deformation equation
L[um (x, t) − χm um−1 (x, t)] = qhH(x, t)Rm (−
→
u m−1 ), (39)
where
m−1

Rm (→
−
u m−1 ) = Dtα um−1 + a ui Dxα um−1−i + Dx3α um−1 (40)
i=0

and χm is already defined in (15).

Now we have a great freedom to choose the initial guess u0 (x, t), the auxiliary
linear operator L, the nonzero auxiliary parameter h and the auxiliary function
H(x, t). Assume that all of them are properly chosen so that the solution φ(x, t; q)
of the zeroth-order deformation equation (37) exists for all q ∈ [0, 1], H(x, t) = 1
and the power series (6) is convergent at q = 1. Now the solution of the mth-order
deformation equation [Banerjee et al. (2017)] for m ≥ 1 becomes
u (x, t) = χ u
m (x, t) + hD−α R (−
m m−1
→u t ).
m m−1 (41)
Using Eqs. (15), (40) and (41) along with the initial condition (38), the approximate
solution of Eq. (34) takes the following form:
u0 (x, t) = Eα (xα ),
tα
u1 (x, t) = hEα (xα ) ,
Γ(α + 1)
tα t2α
u2 (x, t) = (h + h2 )Eα (xα ) + h2 Eα (xα ) ,
Γ(α + 1) Γ(2α + 1)

2050018-10
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

tα t2α
u3 (x, t) = (h + 2h2 + h3 )Eα (xα ) + (2h2 + 2h3 )Eα (xα )
Γ(α + 1) Γ(2α + 1)
t3α
+ h3 Eα (xα ) ,
Γ(3α + 1)
tα
u4 (x, t) = (h + 3h2 + 3h3 + h4 )Eα (xα ) + (3h2 + 7h3 + 3h4 )Eα (xα )
Γ(α + 1)
t2α t3α t4α
× + (3h3 + 3h4 )Eα (xα ) + h4 Eα (xα ) ,
Γ(2α + 1) Γ(3α + 1) Γ(4α + 1)
tα
u5 (x, t) = (h + 4h2 + 6h3 + 4h4 + h5 )Eα (xα )
Γ(α + 1)
t2α
+ (4h2 + 12h3 + 12h4 + 4h5 )Eα (xα )
Γ(2α + 1)
t3α t4α
+ (6h3 + 12h4 + 6h5 )Eα (xα ) + (4h4 + 4h5 )Eα (xα )
Γ(3α + 1) Γ(4α + 1)
t5α
+ h5 Eα (xα ) ,
Γ(5α + 1)
and so on.
In this case, the approximate solution to the homogeneous KdV equation takes
the following form:
M

uapp (x, t) = un . (42)
n=0

Considering α = 1 and u0 = ex , the author in Banerjee et al. [2017] calculated

the exact solution as u(x, t) = ex−t .
Equation (42) is a family of approximate solutions to the problem (34) in
terms of the convergence parameter h. To find the suitable value of h for the

Fig. 1. The h-curve of the 5th-order approximation solution when α = 1 .

2050018-11
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

(a) (b)

(c) (d)

Fig. 2. (a) Exact solution of homogeneous KdV equation u(x, t) when α = 1. (b) Approximation
solution of homogeneous KdV equation u(x, t) obtained by HAM when α = 1. (c) Approximation
solution of homogeneous KdV equation u(x, t) when α = 0.85. (d) Approximation solution of
homogeneous KdV equation u(x, t) when α = 0.9.

Table 1. Absolute errors for problem (34) with different values of h, t and x when α = 1.
t x h = −1.5 h = −1.25 h = −1 h = −0.75 h = −0.5
0.01 −6 0.000000829820 0.000000027204 0 0.000000022312 0.000000751577
0.02 −4.8 0.000005890090 0.000000201791 0 0.000000136135 0.000004840419
0.3 −3.6 0.000031292997 0.000001116750 0 0.000000620773 0.000023371285
0.04 −2.40 0.000147506089 0.000005468033 0.000000000001 0.000002506369 0.000100265790
0.05 −1.2 0.000650715076 0.000024995735 0.000000000006 0.000009445388 0.000403098027
0.06 0 0.002751316542 0.000109279922 0.000000000064 0.000034001909 0.001555067463
0.07 1.20 0.011292724885 0.000462915360 0.000000000537 0.000118325475 0.005829822858
0.08 2.40, 0.045340460157 0.001914961486 0.000000003968 0.000400715978 0.021399359476
0.09 3.6 0.178959366151 0.007775784237 0.000000026670 0.001325570712 0.077284519825
0.10 4.80 0.696758304272 0.031102610860 0.000000166383 0.004291273436 0.275528186107
0.11 6 2.682438514569 0.122866765987 0.000000977250 0.013601110696 0.971945761086

2050018-12
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

(a) (b)

Fig. 3. (a) Approximation solution of homogeneous KdV equation (34) obtained by HAM at
t = 2, h = −1, −6 ≤ x ≤ 6 and different values of the fractional order α = 1, 0.9, 0.8 and 0.7. (b)
Approximation solution of homogeneous KdV equation (34) obtained by HAM at x = 0.21, h =
−1, 0 ≤ t ≤ 3.5 and different values of the fractional order α = 1, 0.9, 0.8 and 0.7.

approximate solution (42), we plot the so-called h-curve of ut (1, 0) using the
5th-term approximation as shown in Fig. 1. According to the h-curve, the valid
region of h corresponds to the line segment parallel to the x-axis. We see that the
convergence region for this problem is −1.4 ≤ h ≤ −0.6. In Table 1, we calculate
absolute error for different values of h, x and t. Finally, we see that h = −1 is the
most suitable value of h.

Table 2. Comparison of the exact solution and approximation solution by HAM for α = 1 and
h = −1.
x Exact solution Approximation solution Absolute error En
x = −6 0.002454088180 0.002454088180 0
x = −2 0.133988674669 0.133988674669 0
t = 0.01
x =2 7.315533762310 7.315533762310 0
x =6 399.414609927110 399.414609927109 0.000000000001
x = −6 0.002429669595 0.002429669595 0
x = −2 0.132655465080 0.132655465080 0
t = 0.02
x =2 7.242742985161 7.242742985160 0.000000000001
x =6 395.440368155324 395.440368155288 0.000000000036
x = −6 0.002405493979 0.002405493979 0
x = −2 0.131335521148 0.131335521148 0
t = 0.03
x =2 7.170676488347 7.170676488339 0.000000000007
x =6 391.505670749888 391.505670749482 0.000000000407
x = −6 0.002381558914 0.002381558914 0
x = −2 0.130028710878 0.130028710878 0
t = 0.04
x =2 7.099327065157 7.099327065115 0.000000000042
x =6 387.610124237783 387.610124235501 0.000000002282
x = −6 0.002357862006 0.002357862006 0
x = −2 0.128734903588 0.128734903585 0.000000000003
t = 0.05
x =2 7.028687580589 7.028687580430 0.000000000159
x =6 383.753339061112 383.753339052419 0.000000008693

2050018-13
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

6. Approximate Solution of Nonhomogeneous Space-Time

Fractional KdV Equation
In this section, we shall apply HAM to construct approximate solution of nonho-
mogeneous space-time fractional KdV equation [Goswami et al. (2018)]
t2
Dtα u − uDxα u + u5α α α
x = cosα (x ) + 2t sinα (x ) + sinα (2xα ),
2
− ∞ < x < ∞, t > 0, 4 < 5α < 5 (43)

with initial condition u(x, 0) = 0. To solve this problem by means of the HAM, we
choose auxiliary linear operators
∂ α φ(x, t; q)
L[φ(x, t; q)] = and
∂tα
∂ α ψ(x, t; q)
L[ψ(x, t; q)] = (44)
∂tα
with property L[c] = 0 where c is a constant. we define the nonlinear operators Nφ
using (43) in the form
∂ α φ(x, t; q) ∂ α φ(x, t; q) ∂ 5α φ(x, t; q)
Nφ = − φ(x, t; q) + − cosα (xα )
∂tα ∂xα ∂x5α
t2
− 2t sinα (xα ) − sinα (2xα ). (45)
2
We construct the zeroth-order deformation equation

(1 − q)L[φ(x, t; q) − u0 (x, t)] = qhH(x, t)N [φ(x, t; q)]. (46)

Obviously, when q = 0 and q = 1

φ(x, t; 0) = u0 (x, t), (47)

φ(x, t; 1) = u(x, t). (48)

Thus, we obtained the mth-order deformation equations

L[um (x, t) − χm um−1 (x, t)] = qhH(x, t)Rm (−

→
u m−1 ), (49)

where
m−1

Rm (−
→
u m−1 ) = Dtα vm−1 − ui Dxα um−1−i + Dx5α vm−1 − (1 − χm )(cosα (xα )
i=0

t2
+ 2t sinα (xα ) + sinα (2xα )), (50)
2
and χm is already defined in (15).
Now we have great freedom to choose the initial guess u0 (x, t), the auxiliary
linear operator L, the nonzero auxiliary parameter h and the auxiliary function

2050018-14
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

H(x, t). Assume that the solutions φ(x, t; q) of the zeroth-order deformation equa-
tions (46) exist for all q ∈ [0, 1], H(x, t) = 1 and the power series
∞

φ(x, t; q) = u0 (x, t) + um (x, t)q m (51)
n=1

is convergent at q = 1. Now the solution of the mth-order deformation equations

(49) for m ≥ 1 becomes
um (x, t) = χm um−1 (x, t) + hDt−α Rm (−
→
u m−1 ). (52)
Using Eqs. (15) and (50) along with the initial condition (47), Eq. (52) becomes
u0 (x, t) = 0,

α tα tα+1 tα+2
u1 (x, t) = −h cosα (x ) + 2 sinα (xα ) + sinα (2xα ) ,
Γ(α + 1) Γ(α + 2) Γ(α + 3)
tα tα+1
u2 (x, t) = −(h + h2 ) cosα (xα ) + 2 sinα (xα )
Γ(α + 1) Γ(α + 2)
tα+2 t2α
+ sinα (2xα ) − h2 − sinα (xα )
Γ(α + 3) Γ(2α + 1)
t2α+1 t2α+2
+ 2 cosα (xα ) + 25 cosα (2xα ) ,
Γ(2α + 2) Γ(2α + 3)
tα tα+1
u3 (x, t) = −(h + 2h2 + h3 ) cosα (xα ) + 2 sinα (xα ) + sinα (2xα )
Γ(α + 1) Γ(α + 2)
tα+2 t2α
× − (2h2 + 2h3 ) − sinα (xα ) + 2 cosα (xα )
Γ(α + 3) Γ(2α + 1)
t2α+1 t2α+2
× + 25 cosα (2xα )
Γ(2α + 2) Γ(2α + 3)
t3α
− h3 − cosα (xα ) cosα (xα )Γ(2α + 1)
Γ(α + 1)2 Γ(3α + 1)
t3α+2
+ 4 sinα (x) cosα (xα )Γ(2α + 3)
Γ(α + 2)2 Γ(3α + 3)
t3α+4
+ 2 sinα (2xα ) cosα (2xα )Γ(2α + 5)
Γ(α + 3)2 Γ(3α + 5)
+ (2 cosα (xα ) cosα (xα ) − 2 sinα (xα ) sinα (xα ))Γ(2α + 2)
t3α+1
× + (4 sinα (xα ) cosα (2xα )
Γ(α + 1)Γ(α + 2)Γ(3α + 2)
t3α+3
+ 2 sinα (2xα ) cosα (xα ))Γ(2α + 4)
Γ(α + 2)Γ(α + 3)Γ(3α + 4)

2050018-15
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

+ (2 cosα (xα ) cosα (2xα ) − sinα (2xα ) sinα (xα ))Γ(2α + 3)

t3α+2 t3α
× − h3 − cosα (xα )
Γ(α + 1)Γ(α + 3)Γ(3α + 3) Γ(3α + 1)
t3α+1 t3α+2
− 2 sinα (xα ) − 210 sinα (2xα ) ,
Γ(3α + 2) Γ(3α + 3)
and so on.
In this case, the approximate solution to the nonhomogeneous fractional KdV
equation takes the following form:
M

uapp (x, t) = un . (53)
n=0

Considering α = 1 and u0 = 0, Goswami et al. [2018] calculated the exact

solution as u(x, t) = t cos(x).
Equation (53) is a family of approximate solutions to the problem (2) in terms
of the convergence parameter h. To find the suitable value of h for the approximate
solution (53), we plot the so-called h-curve of ut (1, 0) using 3rd term approximation
as shown in Fig. 4. According to the h-curve, the valid region of h corresponds to
the line segment parallel to the x-axis. We see that the convergence region for this
problem is −1.5 ≤ h ≤ 0. In Table 3, we calculate the absolute error for different
values of h, x and t. Finally, we see that h = −1 is the most suitable value of h.
To test the accuracy and efficiency of the HAM, Table 2 shows the comparison
of the exact solution with approximate solution and absolute error |eexact − eapprox |
for homogeneous 5αth-order STFKdV equation (34) for 0.01 ≤ t ≤ 0.05 and −6 ≤
x ≤ 6. It can be observed that the approximate solution for α = 1 of Eq. (34) is very
near to the exact solution with high accuracy at the fifth term approximation. Since
the wave propagates towards right, Figs. 2(a)–2(d) illustrate the wave obtained from

Fig. 4. The h-curve of the 3rd-order approximation solution when α = 1.

2050018-16
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

Table 3. Absolute errors for problem (2) with different values of h, t and x when α = 1.
t x h = −1.5 h = −1.25 h = −1 h = 0.75 h = 0.5
0.01 −5.999 0.001231279468 0.000167207518 0.000013710118 0.000134431229 0.001182435018
0.02 −4.799 0.000665366913 0.000276623845 0.000199619069 0.000197400665 0.000033016715
0.03 −3.599 0.002919717250 0.000169021101 0.000203070267 0.000647088189 0.003613564003
0.04 −2.399 0.004821501689 0.001074862452 0.000494969401 0.000116243814 0.003026893027
0.05 −1.199 0.000439138771 0.001235812303 0.001209477462 0.001624655560 0.003745867909
0.06 0.001 0.003749027424 0.001303017466 0.001200030100 0.001480121075 0.007681400986
0.07 1.201 0.008457222031 0.003340661495 0.002338288240 0.002211579494 0.000277987515
0.08 2.401 0.002749829152 0.001708459575 0.002156229206 0.003353649142 0.010060888785
0.09 3.601 0.013126342496 0.002763884044 0.001271070862 0.000414669894 0.008038551915
0.10 4.801 0.010249000640 0.006104844526 0.004942470088 0.005698813394 0.007310810506
0.11 6.001 0.008031667791 0.001761018357 0.002764902585 0.004317834598 0.015757664104

the HAM solution of the homogeneous 5αth-order STFKdV equation (34) different
values of α. It is clear from the figures that the wave amplitude remain almost
same for α = 1, 0.85 and 0.9. Therefore, the solution graph of the homogeneous
5αth-order STFKdV equation (34) preserves its shape throughout the motion. The
solitary behavior is shown in two-dimensional figure (see Fig. 3(a)) at t = 2, h =
−1, −6 ≤ x ≤ 6 for different values of the fractional order α = 1, 0.9, 0.8 and 0.7. It
is clear from the figure that the system is affected by the order of derivative. Again,
Fig. 3(b) presents the solution u(x, t) at x = 0.21, 0 < t < 3 for different values
of the order of fractional order derivative parameter α. Obviously, amplitude and
speed of the wave decrease with decreasing of the parameter α for both the cases.

Table 4. Comparison of the exact solution and approximation solution by HAM for α = 1 and
h = −1.
Time x Exact Solution Approximation solution Absolute error (En )
x = −6 0.009601702867 0.009615364265 0.000013661399
x = −2 −0.004161468365 −0.004206461201 0.000044992836
t = 0.01
x=2 −0.004161468365 −0.004116036953 0.000045431412
x=6 0.009601702867 0.009587063807 0.000014639060
x = −6 0.019203405733 0.019256756138 0.000053350405
x = −2 −0.008322936731 −0.008500915925 0.000177979194
t = 0.02
x=2 −0.008322936731 −0.008141269505 0.000181667226
x=6 0.019203405733 0.019142094299 0.000061311434
x = −6 0.028805108600 0.028922128392 0.000117019793
x = −2 −0.012484405096 −0.012880106673 0.000395701577
t = 0.03
x=2 −0.012484405096 −0.012075650918 0.000408754178
x=6 0.028805108600 0.028660749352 0.000144359247
x = −6 0.038406811466 0.038609339604 0.000202528138
x = −2 −0.016645873462 −0.017340354210 0.000694480748
t = 0.04
x=2 −0.016645873462 −0.015919018094 0.000726855368
x=6 0.038406811466 0.038138363806 0.000268447660
x = −6 0.048008514333 0.048316210501 0.000307696168
x = −2 −0.020807341827 −0.021877479173 0.001070137346
t = 0.05
x=2 −0.020807341827 −0.019671170070 0.001136171758
x=6 0.048008514333 0.047569894529 0.000438619803

2050018-17
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

For physical problem, the roughness of the surface is measured by the fractional
parameter α. Thus, the fractional order derivative act as the delay parameter.
Table 4 shows the comparison of the exact solution with approximate solution
and absolute error |eexact −eapprox | for nonhomogeneous 5αth-order STFKdV equa-
tion (43) for 0.01 ≤ t ≤ 0.05 and −6 ≤ x ≤ 6. It can be seen that the values of the
approximate solution for α = 1 of the different grid points obtained by the HAM
are close to the values of the exact solution with high accuracy at the third-term
approximation. Figures 5(a)–5(d) illustrate the wave obtained from the HAM solu-
tion of the nonhomogeneous 5αth-order STFKdV equation (43). From the images
at the top of each figure, it is evident that the wave preserves its shape throughout
the motion. Due to fractional wave equation, Figs. 5(b)–5(d) for α = 1, 0.85 and 0.9,
respectively, show some discontinuity at the crest position of the 3D wave. These
effects arise due to nonlinear interaction among wave components and the presence
of fractional effect in Eq. (43).

(a) (b)

(c) (d)

Fig. 5. (a) Exact solution of nonhomogeneous KdV equation u(x, t) when α = 1. (b) Approximation
solution of nonhomogeneous KdV equation u(x, t) obtained by HAM when α = 1 (c) Approxima-
tion solution of nonhomogeneous KdV equation u(x, t) when α = 0.85. (d) Approximation solution
of nonhomogeneous KdV equation u(x, t) when α = 0.9.

2050018-18
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

(a) (b)

Fig. 6. (a) Approximation solution of nonhomogeneous KdV equation u(x, t) obtained by HAM
when α = 0.7. (b) Approximation solution of nonhomogeneous KdV equation u(x, t) obtained by
HAM when α = 0.75.

The solution behavior as the fractional order derivative parameter α changes

is shown in Fig. 6. As α increases from 0.7 to 0.75, the parabolic shape at chest
and trough position of the approximate solution decreases. In Fig. 7(a), we see that
the chest position of the wave are moving downstream and trough position is going
upstream as α increase. Figure 7(b) indicates that the solution u(x, t) is affected
by the fractional order of the wave equation. Amplitude of the wave increases as
parameter α increases. Therefore, the speed and amplitude of the wave decrease due
to decreasing fractional order α. Thus, in both cases, the fractional order derivative
plays as the role of delay parameter α.

(a) (b)

Fig. 7. (a) Approximation solution of nonhomogeneous KdV equation (43) obtained by HAM at
t = 0.0021, h = −1, −6 ≤ x ≤ 6 and different values of the fractional order α = 1, 0.9, 0.8 and
0.7. (b) Approximation solution of nonhomogeneous KdV equation (43) obtained by HAM at
x = 0.21, h = −1, 0 ≤ t ≤ 5 and different values of the fractional order α = 1, 0.9, 0.8 and 0.7.

2050018-19
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

7. Conclusion
In this paper, the nonlinear space-time fractional homogeneous and non-
homogeneous KdV equation of 5αth-order has been deducted from the correspond-
ing classical differential equation with the help of Semi-inverse and Agrawal’s vari-
ational technique. As a second step, the HAM is applied to find the approximate
solutions of the generated space-time fractional differential equations. The solution
contains auxiliary parameter which adjust and control the convergence region of
solution series. A valid region of the auxiliary parameter is the horizontal line seg-
ment which was pointed by Liao. The validity of this method is shown by applying
it for space-time fractional homogeneous and nonhomogeneous KdV equations of
fifth order. Further, this paper shows that HAM is very powerful and an efficient
tool for solving the nonlinear fractional differential equation. It is clear from the
numerical results that the order of fractional derivative plays the role like the delay
parameter.

Appendix A.
A.1. Integration by Parts via Jumarie derivative
The fractional integration by parts is given by the rule [Jumarie (2012)]
 
f (x)Dxα g(x)(dx)α = Γ(α + 1)[f (x)g(x)]Ω − g(x)Dxα f (x)(dx)α . (A.1)
Ω Ω

A.2. Euler–Lagrange equation via Jumarie derivative

Let us consider the following variational functional:
 
Jα (v) = Lα (v, Dtα v, Dxα v, Dx3α v)(dxdt)α . (A.2)
R T

For extremum, taking δ operator to the both sides of (A.2) and using (A.1) and
δv|T = δv|R = δ(Dxα v) = δ(Dx2α v) = 0, we get
  
∂F ∂Lα ∂Lα ∂Lα
δJα (v) = δv + δ(Dtα v) + δ(Dxα v) + δ(Dx2α v)
R T ∂v ∂Dtα v α
∂Dx v ∂Dx2α v

∂Lα
+ δ(Dx v) (dxdt)α
3α
∂Dx3α v
  
∂F ∂Lα α ∂Lα α ∂Lα
= δv + D (δv) + Dx (δv) + D2α (δv)
R T ∂v ∂Dtα v t α
∂Dx v ∂Dx2α v x

∂Lα
+ D (δv) (dxdt)α
3α
∂Dx3α v x

2050018-20
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

      
∂F α ∂Lα α ∂Lα
= δv − Dt (δv) − Dx (δv)
R T ∂v ∂Dtα v ∂Dxα v
    
2α ∂Lα 3α ∂Lα
+ Dx (δv) − Dx (δv) (dxdt)α
∂Dx2α v ∂Dx3α v
        
∂F α ∂Lα α ∂Lα 2α ∂Lα
= − Dt − Dx + Dx
R T ∂v ∂Dtα v ∂Dxα v ∂Dx2α v
 
∂Lα
− Dx3α δv(dxdt)α .
∂Dx3α v

Optimizing this variation Jα (v), i.e. δJα (v) = 0 gives the Euler–Lagrange equation
for STFKdV equation in the form
       
∂F α ∂Lα α ∂Lα 2α ∂Lα 3α ∂Lα
− Dt − Dx + Dx − Dx = 0.
∂v ∂Dtα v ∂Dxα v ∂Dx2α v ∂Dx3α v

References
Abbasbandy, S. [2006] “The application of homotopy analysis method to nonlinear equa-
tions arising in heat transfer,” Phys. Lett. A. 360(1), 109–113.
Adomian, G. and Rach, R. [1992] “Noise terms in decomposition series solution,” Comput.
Math. Appl. 24(11), 61.
Adomian, G. [1992] “A review of the decomposition method and some recent results for
nonlinear equation,” Math. Comput. Model. 13(7), 17.
Adomian, G. [1994] Solving Frontier Problems of Physics: The Decomposition Method
(Kluwer, Boston, MA).
Banerjee, J., Ghosh, U., Sarkar, S. and Das, S. [2017] “A study of fractional Schrdinger
equation composed of Jumarie fractional derivative,” Pramana J. Phys. 88(4), 70.
Carmichael, B. et al. [2015] “The fractional viscoelastic response of human breast tissue
cells,” Phys. Biol. 12(4), 046001.
Chen, Y. [2005] “Fractional order signal processing in biology/biomedical signal analysis,”
in Fractional Order Calculus Day at Utah State University, April 2005.
Das, T., Ghosh, U., Sarkar, S. and Das, S. [2018] “Time independent fractional Schrodinger
equation for generalized Mie-type potential in higher dimension framed with Jumarie
type fractional derivative,” J. Math. Phys. 20, 022111.
Engheta, N. [1995] “A note on fractional calculus and the image method for dielectric
spheres,” J. Electromagn. Waves Appl. 1179–1188.
Engheta, N. [1996] “On fractional calculus and fractional multipoles in electromagnetism,”
IEEE Transactions on Antennas and Propagation, AP-444, April 1996, pp. 554–566.
Engheta, N. [1996] “Electrostatic fractional image methods for perfectly conducting wedges
and cones,” IEEE Transactions on Antennas and Propagation, AP-44, 12 December
1996, pp. 1565–1574.
Engheta, N. [1996] “Use of fractional integration to propose some fractional solutions
for the scalar helmholtz equation,” in Progress in Electromagnetics Research (PIER)
Monograph Series, eds., Vol. 12 ed. Kong, J. A., Cambridge, MA (EMW Publishing,
Cambridge), pp. 107–132.
Garg, S. and Pant, M. [2018] “Meshfree methods: A comprehensive review of applications,”
Int. J. Comput. Methods 15(4), 1830001.

2050018-21
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

S. Biswas & U. Ghosh

Gepreel, K. A. and Omran, S. [2011] “Exact solutions for nonlinear partial fractional
differential equations,” Chin. Phys. B 21, 110204.
Goswami, A., Singh, J. and Kumar, D. [2018] “Numerical simulation of fifth order KdV
equations occurring in magneto-acoustic waves,” Ain Shams Eng. 9(4), 2265–2273.
Gu, Y., Zhuang, P. and Liu, Q. [2011] “An advanced meshless method for time fractional
diffusion equation,” Int. J. Comput. Methods 8(4), 653–665.
Gu, Y. T. [2005] “Meshfree methods and their comparisons,” Int. J. Comput. Methods
2(4), 477–515.
Guo, S. et al. [2012] “The improved fractional sub-equation method and its applications to
the space-time fractional differential equations in fluid mechanics,” Phys. Lett. 376(4),
407–411.
Guo, S., Mei, L. and Li, Y. [2013] “Fractional variational homotopy perturbation iteration
method and its application to a fractional diffusion equation” Appl. Math. Comput.
219(11), 5909–5917.
Guo, S., Mei, L., Zhang, Z. and Jiang, Y. [2018] “Finite difference/spectral-Galerkin
method for a two-dimensional distributed-order time-space fractional reaction-diffusion
equation,” Appl. Math. Lett. 85, 157–163.
Guo, S. et al. [2019] “An efficient finite difference/Hermite-Galerkin spectral method
for time-fractional coupled sine-Gordon equations on multidimensional unbounded
domains and its application in numerical simulations of vector solitons,” Comput. Phys.
Commun. 237, 110–128.
He, L. E. I. and Ma, G. [2010] “Development of 3D numerical manifold method,” Int. J.
Comput. Methods 7(1), 107–129.
He, J. H. [1997] “A new approach to nonlinear partial differential equations,” Commun.
Nonlinear Sci. Numer. Simul. 2(4), 230–235.
He, J.-H. [1997] “Semi-inverse method of establishing generalized variational principles
for fluid mechanics with emphasis on turbo-machinery aerodynamics,” Int. J. Turbo
Jet-Engines 14(1), 2328.
He, J. H. [1998] “A variational approach to nonlinear problems and its application,” Mech.
Appl. 20(1), 30–31.
He, J. H. [1999] “Homotopy perturbation technique,” Comput. Meth. Appl. Mech. Eng.
178, 257–262.
He, J. H. [2000] “Variational iteration method for autonomous ordinary differential sys-
tems,” Appl. Math. Comput. 118(2–3), 115–123.
He, J.-H. [2004] “Variational principles for some nonlinear partial differential equations
with variable coefficients,” Chaos Solitons Fractals 19, 847851.
He, J. H. [2005] “Homotopy perturbation method for bifurcation of nonlinear problems,”
Int. J. Non-linear Sci. Num. Sim. 6(2), 207–208.
Inokuti, M. et al. [1978] “General use of the Lagrange multiplier in nonlinear mathemati-
cal physics,” in Variational Method in the Mechanics of Solids, ed. Nemat-Nassed, S.
(Pergamon, Press), pp. 156–162.
Jumarie, G. [2009] “Table of some basic fractional calculus formulae derived from a mod-
ified Riemann–Liouville derivative for non-differentiable functions,” Appl. Math. Lett.
22, 378–385.
Jumarie, G. [2012] “An approach to differential geometry of fractional order via modified
Riemann–Liouville derivative,” Acta Mathematica Sinica (English Series) 28(9), 1741–
1768.
Khader, M. M. [2015] “Fractional Chebyshev finite difference method for solving the
fractional-order delay BVPs,” Int. J. Comput. Methods 12(6), 1550033.

2050018-22
August 15, 2020 1:11 WSPC/0219-8762 196-IJCM 2050018

Approximate Solution of 5αth-Order Space-Time Fractional KdV Equations

Kilbas, A. A., Srivastava, H. M. and Trujillo, J. J. [2006] Theory and Applications of Frac-
tional Differential Equations (North-Holland Mathematical Studies), Vol. 204 (Elsevier,
Amsterdam).
Liao, S. J. [1992] The proposed homotopy analysis technique for the solution of nonlinear
problems, Ph.D. thesis Shanghai Jiao Tong University.
Liao, S. J. (2003) Beyond Perturbation: Introduction to the Homotopy Analysis Method
(Chapman Hall/CRC press, Boca Raton).
Magin, R. L. [2006] Fractional Calculus in Bioengineering (Begell House Publisher, Inc.,
Connecticut).
Motsa, S. S. [2012] “Application of the new spectral homotopy analysis method (sham)
in the non-linear heat conduction and convective fin problem with variable thermal
conductivity,” Int. J. Comput. Methods 9(3), 1250039.
Schneider, W. and Wyss, W. [1989] “Fractional diffusion and wave equations,” J. Math.
Phys. 30, 134–144.
Taghipour, R., Akhlaghi, T. and Nikkar, A. [2014] “Explicit solution of the large amplitude
transverse vibrations of a flexible string under constant tension,” Latin Am. J. Solids
Struct. 11(3), 545–555.

2050018-23