Journal of Advanced Research in Applied Sciences and Engineering Technology 31, Issue 1 (2023) 132-143

Journal of Advanced Research in Applied

Sciences and Engineering Technology
Journal homepage:
https://semarakilmu.com.my/journals/index.php/applied_sciences_eng_tech/index
ISSN: 2462-1943

Approximate Analytical Solution for Time-Fractional Nonlinear Telegraph

Equations with Source Term
Abdul Rahman Farhan Sabdin1, Che Haziqah Che Hussin2,*, Graygorry Brayone Ekal1, Arif
Mandangan1, Jumat Sulaiman1

1
Faculty of Science and Natural Resources, Universiti Malaysia Sabah, Jalan UMS, 88400 Kota Kinabalu, Sabah, Malaysia
2 Preparatory Centre of Science and Technology, Universiti Malaysia Sabah, Jalan UMS, 88400, Kota Kinabalu, Sabah, Malaysia

ARTICLE INFO ABSTRACT

Article history: In this study, we considered time-fractional nonlinear telegraph equations (TFNLTEs).
Received 28 January 2023 For solving the TFNLTEs, we deployed a method known as the Multistep Modified
Received in revised form 23 May 2023 Reduced Differential Transform Method (MMRDTM). Prior to the multistep technique,
Accepted 29 May 2023 the nonlinear term in TFNLTEs is replaced with corresponding Adomian polynomials. It
Available online 14 June 2023
can be observed that the MMRDTM is much simpler and more straightforward. On top
of that, it works exceptionally where the obtained solutions are more accurately
approximated over time. To demonstrate the performance of the MMRDTM in terms of
its capabilities and accuracies, we provided two numerical examples of solving TFNLTEs
Keywords: by using MMRDTM and Modified Reduced Differential Transform Method (MRDTM). By
Adomian Polynomials; Nonlinear comparing the absolute errors of the obtained solutions by both methods, we
Telegraph Equations; Multistep; Source demonstrated that the solutions provided by the MMRDTM much closer to the exact
Term; Time-Fractional; Reduced solutions compared to the corresponding solutions yielded by the MRDTM. This justified
Differential Transform Method that the MMRDTM provides highly accurate and precise solutions for the TFNLTEs.

1. Introduction

Various physical phenomena in science and engineering such as diffusion process, rheology,
damping laws, electric transmission, viscoelasticity, and fluid mechanics can be effectively
enlightened and demonstrated by fractional partial differential equations (FPDEs). Unfortunately,
accurate analytical solutions for these kind of equations can only be approximated seldomly.
Alternatively, approximation and numerical techniques are employed to address such issues. Various
approaches have been proposed such as finite difference method [1], Adomian decomposition
method [2], Fourier method [3], variational iteration method [4], wavelet method [5], homotopy
analysis method [6], Tau method [7], and the fractional Sumudu decomposition method (FSDM) [8].
In recent years, the telegraph equations have attracted many researchers due to its usage in
physical, chemical, and biological sciences. It is a hyperbolic partial differential equation that is used
to describe radio frequencies, random walk theory, electromagnetic waves, voltage and current on
*
Corresponding author.
E-mail address: haziqah@ums.edu.my

https://doi.org/10.37934/araset.31.1.132143

132
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

transmission lines, and oceanic diffusion, among other things. It also possesses both diffusion and
wave motion characteristics. Besides, the telegraph equation is better at modelling reaction-diffusion
than the standard diffusion equation, and it is a preferable model for understanding several fluid flow
problems involving suspensions as compared to the heat equation [9]. On the other hand, the non-
standard telegraph equations are well defined by a time-fractional derivative of order α (non-
integer). Moreover, fractional telegraph equations are the generalisation of the standard telegraph
equations, which naturally come from fractal space-time. The time-fractional telegraph equation
(TFTE) is important in modeling the Brownian motion. Orsingher and Zhao [10] demonstrated that
TFTEs determine the telegraph processes with Brownian time and the law of iterated Brownian
motion. Apart from that, Nirmala and Balachandran [11] studied the application of the TFTEs to
calculate signal and power losses during transmission media in a communication system.
Furthermore, Vyawahare and Nataraj [12] modelled a neutron transport in a nuclear reactor using
the TFTE. Recently, Madhukar et al., [13] showed concrete proof for the damped-hyperbolic nature
of transient heat conduction in porcine blood and muscle tissue. Through that experiment, they
found that the TFTE mimics the wave-like nature of heat conduction and temperature profiles as well
as avoids the requirement for additional adjustable parameters.
Several authors suggested different ways to obtain solutions of the TFTEs. For instance, the
reproducing kernel theorem was used by Jiang and Lin [14] to find the solution to this equation. Next,
Kumar et al., [15] devise a local meshless method to obtain the solution for the TFTEs with linear
source term using radial basis functions. This is followed by a numerical experiment on various
complicated domains to demonstrate the method's effectiveness, and it shows a pleasant result.
Furthermore, Wang and Mei [16] applied the generalized finite difference scheme in time and the
Legendre spectral Galerkin technique in space to solve the TFTEs with a forcing term. Furthermore,
Kumar et al., [17] presented a finite difference scheme for the generalised TFTEs with forcing term
that is defined using generalised fractional derivative terms. Ray [18] formulated and implemented
a modification to the Fractional Reduced Differential Transform Method (FRDTM) for solving the
fractional Korteweg–De Vries equation (FKdVEs). This strategy used Adomian polynomials to
substitute the nonlinear term of the equation. Consequently, the nonlinear problems can have
solutions in shorter time with fewer calculated terms. Besides that, Hassani [19] proposed variable-
order space–time fractional telegraph equation using transcendental Bernstein series.
In 2018, Hussin et al., [20] proposed the Multistep Modified Reduced Differential Transform
Method (MMRDTM) and deployed this method for solving nonlinear Schrodinger equations (NLSE).
As a result, the MMRDTM performed outstandingly as the NLSEs are successfully approximated with
high accuracy and precision. Furthermore, the MMRDTM was also applied by Hussin et al., [21] to
handle Klein-Gordon equations. Once again, the MMRDTM performed outstandingly as
approximation of the equations are obtained with great precision and efficiency. Motivated by these
results, Hussin et al., [22] implemented the MMRDTM to determine the approximate analytical
solutions of the one-dimensional fractional NLSE. As expected, the results are obtained by the
MMRDTM with high accuracy to the exact solutions. Hussin et al., also use the method to obtain
solitary wave solutions [23-24].
In this paper, we proposed a multistep technique and a variation by using Adomian polynomials
to discover the solution to the one-dimensional time-fractional nonlinear telegraph equations
(TFNLTEs) with source term using the MMRDTM. To produce Adomian polynomials, we deployed
parametrization approaches instead of doing time-consuming high-derivative computations. The
following one-dimensional TFNLTEs [25] is considered

𝜕2𝛼 𝑢(𝑥,𝑡) 𝜕𝛼 𝑢(𝑥,𝑡) 𝜕2 𝑢(𝑥,𝑡)

+𝜃 =𝛾 + 𝑓(𝑥, 𝑡), (1)
𝜕𝑡 2𝛼 𝜕𝑡 𝛼 𝜕𝑥 2

133
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

where 𝑥 ∈ (0, 𝑙) and 𝑡 ∈ (0, 𝜏). The initial conditions are given by

𝑢(𝑥, 0) = 𝜑(𝑥),

𝑢𝑡 (𝑥, 0) = 𝜓(𝑥),

where 𝛼 is a parameter defining the fractional derivative’s order. Furthermore, the functions
𝑓, 𝜑, and 𝜓 are sufficiently smooth prescribed functions. Then, the rates 𝜃 and 𝛾 are arbitrary
nonnegative and positive constants, respectively. In the case where ½ < 𝛼 ≤ 1, the Caputo
fractional derivatives are used to describe the time-fractional derivative.

2. Formation of Fractional Multistep Modified Reduced Differential Transform Method

To demonstrate the fundamental concepts behind the use of fractional MMRDTM, let consider a
general nonlinear partial differential equation

𝐿𝑢(𝑥, 𝑡) + 𝑅𝑢(𝑥, 𝑡) + 𝑁𝑢(𝑥, 𝑡) = 𝑔(𝑥, 𝑡) , (2)

subjected to the initial condition,

𝑢(𝑥, 0) = 𝑓(𝑥), (3)

where 𝐿 ≡ 𝐷𝑡𝛼 , 𝑅, 𝑁𝑢(𝑥, 𝑡) and 𝑔(𝑥, 𝑡) are invertible linear operator, remaining part of the linear
operator, nonlinear term and inhomogeneous term respectively. Next, the following iteration
formula is formed according to the Reduced Differential Transformation Method (RDTM)

(𝑟 + 1)𝑈𝑟+1 (𝑥) = 𝑆𝑟 (𝑥) − 𝑅𝑈𝑟 (𝑥) − 𝑁𝑈𝑟 (𝑥), (4)

where 𝑈𝑟 (𝑥), 𝑆𝑟 (𝑥), 𝑁𝑈𝑟 (𝑥) and 𝑅𝑈𝑟 (𝑥) represent the transformation functions of 𝐿𝑢(𝑥, 𝑡),
𝑅𝑢(𝑥, 𝑡), 𝑁𝑢(𝑥, 𝑡) and 𝑔(𝑥, 𝑡) respectively. From the initial condition, we have

𝑈0 (𝑥) = 𝑓(𝑥). (5)

Then, the nonlinear term is written as follows

𝑁(𝑢, 𝑡) = ∑∞ 𝑛
𝑛=0 𝐴𝑛 (𝑈0 (𝑥), 𝑈1 (𝑥), … , 𝑈𝑛 (𝑥))𝑡 , (6)

where 𝐴𝑛 is the correspond Adomian’s polynomials. Recently, a novel technique for calculating the
Adomian polynomials was claimed in [26], such as

𝐴0 = 𝑁(𝑈0 (𝑥)),

1 𝜋
𝐴𝑛 (𝑈0 (𝑥), 𝑈1 (𝑥), … , 𝑈𝑛 (𝑥)) = 2𝜋 ∫−𝜋 𝑁(∑𝑛𝑘=0 𝑈𝑘 (𝑥)𝑒 𝑖𝑘𝑥 ) 𝑒 −𝑖𝑛𝜆 𝑑𝜆,

where 𝑛 ≥ 1. As demonstrated, this algorithm does not necessitate the time-consuming calculation
of high derivatives. The function 𝑢(𝑥, 𝑡) can be portrayed on the basis of the differential
transformation properties such as

134
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

𝑢(𝑥, 𝑡) = ∑∞ 𝛼𝑟
𝑟=0 𝑈𝑟 (𝑥)𝑡 . (7)

Definition 1. If function 𝑢(𝑥, 𝑡) is analytic and differentiated continuously with respect to space 𝑥
and time 𝑡 in the desired domain, then let

1
𝑈𝑟 (𝑥) = Γ(𝑟𝛼+1) { 𝐷𝑡𝛼𝑟 𝑢(𝑥, 𝑡)}𝑡=0 . (8)

The generalized fractional RDTM of the function 𝑢(𝑥, 𝑡) is given by

1
𝑢(𝑥, 𝑡) = ∑∞ 𝛼𝑘 𝛼𝑘
𝑘=0 (Γ(𝑘𝛼+1) { 𝐷𝑡 𝑢(𝑥, 𝑡)}𝑡=0 ) 𝑡 . (9)

By applying the Riemann-Liouville integral 𝐽𝛼 on both sides of Eq. (2), we obtain

𝑢(𝑥, 𝑡) = 𝐽𝛼 𝑔(𝑥, 𝑡) − 𝐽𝛼 𝑅𝑢(𝑥, 𝑡) − 𝐽𝛼 𝑁𝑢(𝑥, 𝑡) + Φ, (10)

where from the initial condition Φ = 𝑢(𝑥, 0) = 𝑓(𝑥). Thus, plugging Eq. (8) and Eq. (6), for 𝑢(𝑥, 𝑡)
and 𝑁(𝑢, 𝑡) respectively, in Eq. (9) yields

∑∞
𝑟=0 𝑈𝑟 (𝑥)𝑡
𝛼𝑟
= 𝑓(𝑥) + 𝐽𝛼 (∑∞ 𝛼𝑟 𝛼 ∞ 𝛼𝑟 𝛼 ∞ 𝛼𝑟
𝑟=0 𝐺𝑟 (𝑥)𝑡 ) − 𝐽 (𝑅(∑𝑟=0 𝑈𝑟 (𝑥)𝑡 )) − 𝐽 (∑𝑟=0 𝐴𝑟 (𝑥)𝑡 ) (11)

where 𝑔(𝑥, 𝑡) = ∑∞ 𝛼𝑟
𝑟=0 𝐺𝑟 (𝑥)𝑡 , and 𝐺𝑘 (𝑥) is the transformed function of 𝑔(𝑥, 𝑡). After performing
𝛼
Riemann-Liouville integral 𝐽 , we obtain [18],

𝑡 𝛼(𝑟+1) Γ(𝛼𝑟+1) 𝑡 𝛼(𝑟+1) Γ(𝛼𝑟+1)

∑∞
𝑟=0 𝑈𝑟 (𝑥)𝑡
𝛼𝑟
= 𝑓(𝑥) + (∑∞
𝑟=0 𝐺𝑟 (𝑥) ) − (𝑅 (∑∞
𝑟=0 𝑈𝑟 (𝑥) )) −
Γ(𝛼(𝑟+1)+1) Γ(𝛼(𝑟+1)+1)
𝑡 𝛼(𝑟+1) Γ(𝛼𝑟+1)
(∑∞
𝑟=0 𝐴𝑟 (𝑥) ). (12)
Γ(𝛼(𝑟+1)+1)

Finally, following recursive formula is derived by equating coefficients with similar powers of 𝑡

𝑈0 (𝑥) = 𝑓(𝑥),

and

Γ(𝛼𝑟+1) Γ(𝛼𝑟+1) Γ(𝛼𝑟+1)

𝑈𝑟+1 (𝑥) = 𝐺𝑟 (𝑥) Γ(𝛼(𝑟+1)+1) − 𝑅 (𝑈𝑟 (𝑥) Γ(𝛼(𝑟+1)+1)) − 𝐴𝑟 (𝑥) Γ(𝛼(𝑟+1)+1) , (13)

for 𝑟 ≥ 0. Using the known 𝑈0 (𝑥), all 𝑈1 (𝑥), 𝑈2 (𝑥), … , 𝑈𝑛 (𝑥) components can be determined using
Eq. (13). Then by replacing these 𝑈0 (𝑥), 𝑈1 (𝑥), 𝑈2 (𝑥), … , 𝑈𝑛 (𝑥) in Eq. (8), an approximate solution
can be acquired as follows [18]

𝑢̃𝑝 (𝑥, 𝑡) = ∑𝑝𝑟=0 𝑈𝑟 (𝑥)𝑡 𝛼𝑟 ,

where 𝑝 is the order of this approximate solution. Consequently, the following series solution is
obtained

135
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

𝑢(𝑥, 𝑡) = 𝑙𝑖𝑚 𝑢̃𝑝 (𝑥, 𝑡).

𝑝→∞

Finally, follow the multi-step scheme as obtained in Algorithm 1.

Algorithm 1 Multi-step scheme for fractional approximate solution as follows

Step 1 : By using the nodes 𝑡𝑚 = 𝑘ℎ, divide the interval [ 0, 𝑇 ] into 𝑀 subintervals
[𝑡𝑚−1 , 𝑡𝑚 ], 𝑚 = 1,2, … , 𝑀, equally sized ℎ = 𝑇/𝑀.

Step 2 : The RDTM is applied over the interval [0, 𝑡1 ] to the initial value problem. Then, use the
initial conditions 𝑢(𝑥, 0) = 𝑓0 (𝑥), 𝑢1 (𝑥, 0) = 𝑓1 (𝑥) to get the following approximate
solution

𝑢1 (𝑥, 𝑡) = ∑ 𝑈𝑟,1 (𝑥)𝑡 𝛼𝑟

𝑟=0

where 𝑡 ∈ [0, 𝑡1 ].

Step 3 : Use the following initial conditions

𝑢𝑚 (𝑥, 𝑡𝑚−1 ) = 𝑢𝑚−1 (𝑥, 𝑡𝑚−1 )

(𝜕⁄𝜕𝑡)𝑢𝑚 (𝑥, 𝑡𝑚−1 ) = (𝜕⁄𝜕𝑡)𝑢𝑚−1 (𝑥, 𝑡𝑚−1 )

at each subinterval [𝑡𝑚−1 , 𝑡𝑚 ] for the case where 𝑚 ≥ 2. Then, the MMRDTM is used to solve
the initial value problem over the interval [𝑡𝑚−1 , 𝑡𝑚 ] where 𝑡𝑚−1 replaces t 0 .

Step 4 : The procedure is reiterated and proceeded in order to yield a sequence of approximate
𝑢𝑚 (𝑥, 𝑡), 𝑚 = 1,2, … , 𝑀, for the solutions 𝑢(𝑥, 𝑡) such as [27-29]

𝑢𝑘 (𝑥, 𝑡) = ∑ 𝑈𝑟,𝑚 (𝑥)(𝑡 − 𝑡𝑚−1 )𝛼𝑟 ,

𝑟=0

where 𝑡𝑖−1 ≤ 𝑡 ≤ 𝑡𝑖 .

Step 5 : In fact, the MsFRDTM is executing the subsequent solution as follows

𝑢1 (𝑥, 𝑡) , 𝑡 ∈ [0, 𝑡1 ]
(𝑥, , 𝑡 ∈ [𝑡1 , 𝑡2 ]
𝑢(𝑥, 𝑡) = { 𝑢2 𝑡)
⋮
𝑢𝑀 (𝑥, 𝑡) , 𝑡 ∈ [𝑡𝑀−1 , 𝑡𝑀 ].

136
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

3. Numerical Results and Discussion

To demonstrate the effectiveness and accuracy of the proposed MMRDTM for approximating
solutions of fractional nonlinear Telegraph equation (FNLTE), we provide the following numerical
examples.

Example 1 The one-dimensional FNLTE is considered as

𝛼
𝑢𝑡𝑡 + 2𝑢𝑡𝛼 = 𝑢𝑥𝑥 − 𝑢2 − 𝑒 2𝑥−4𝑡 + 𝑒 𝑥−2𝑡 , (14)

with the initial conditions

𝑢(𝑥, 0) = 𝑒 𝑥 ,

𝑢𝑡 (𝑥, 0) = −2𝑒 𝑥 .

The exact solution of this equation for 𝛼 = 1 is 𝑒 𝑥−2𝑡 .

Applying MMRDTM to Eq. (14) and using the basic properties of MMRDTM, one can obtain

Γ(1+𝑟𝛼) 𝜕2 Γ(1+(1+𝑟)𝛼) (−4)𝑟

𝑈𝑅+2,𝑖 (𝑥) = (Γ(1+(2+𝑟)𝛼)) (𝜕𝑥 2 (𝑈𝑟,𝑖 (𝑥)) − ∑𝑛𝑟=0 𝐴𝑟,𝑖 − 2 ( ) 𝑈𝑟+1,𝑖 + 𝑒 2𝑥 ( )−
Γ(1+𝑟𝛼) 𝑟!

(−2)𝑟
𝑒𝑥 ( )).
𝑟!

From the initial condition, we assign

𝑈0 (𝑥) = 𝑒 𝑥 .

Divide the interval [0, 𝑇] into 𝑀 subintervals by using the nodes 𝑡𝑖 = 𝑖ℎ of equal step size ℎ = 𝑇/𝑀
such that [𝑡𝑖−1 , 𝑡𝑖 ], 𝑖 = 1,2, … ,10. After that, the RDTM is used to solve the initial value problem over
the interval [0, 𝑡1 ]. Then, use the initial conditions 𝑢(𝑥, 0) = 𝑓0 (𝑥) and 𝑢1 (𝑥, 0) = 𝑓1 (𝑥) to obtain
the following approximate solution

𝑢1 (𝑥, 𝑡) = ∑ 𝑈𝑟,1 (𝑥)𝑡 𝛼𝑟 ,

𝑟=0

where 0 ≤ 𝑡 ≤ 𝑡1 . For the case where 𝑖 ≥ 2, the initial conditions 𝑢𝑖 (𝑥, 𝑡𝑖−1 ) = 𝑢𝑖−1 (𝑥, 𝑡𝑖−1 ) and
(𝜕⁄𝜕𝑡)𝑢𝑖 (𝑥, 𝑡𝑖−1 ) is applied at each subinterval [𝑡𝑖−1 , 𝑡𝑖 ]. Then, the RDTM is used to solve the initial
value problem over the interval [𝑡𝑖−1 , 𝑡𝑖 ], where 𝑡0 is substituted by 𝑡𝑖−1 . This procedure is repeated
continuously to create a series of approximate solutions 𝑢𝑖 (𝑥, 𝑡), where 𝑖 = 1,2, … ,10, for the
solution 𝑢(𝑥, 𝑡) such as

𝑢𝑖 (𝑥, 𝑡) = ∑ 𝑈𝑟,𝑖 (𝑥)(𝑡 − 𝑡𝑖−1 )𝛼𝑟 ,

𝑟=0

137
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

where 𝑡 ∈ [𝑡𝑖−1 , 𝑡𝑖 ]. In fact, the MMRDTM yields the following solution

𝑢1 (𝑥, 𝑡) 𝑡 ∈ [0, 𝑡1 ]
𝑢 (𝑥, 𝑡) 𝑡 ∈ [𝑡1 , 𝑡2 ] .
𝑢(𝑥, 𝑡) = { 2
⋮ ⋮
𝑢10 (𝑥, 𝑡) 𝑡 ∈ [𝑡9 , 𝑡10 ]

Figure 1(a) is the exact solution compared to the approximate solutions by the proposed MMRDTM
and MRDTM in Figure 1(b) and Figure 1(c) respectively with 𝛼 = 0.6. The outcomes of the
approximate solution for 𝛼 = 0.8 are shown in Figure 2(a) for the MMRDTM and Figure 2(b) for the
MRDTM respectively. From the obtained results, obviously the approximations produced by the
MMRDTM very close to the exact solutions compared to the MRDTM.

(a) Exact solution (b) MMRDTM (c)MRDTM

Fig. 1. Comparison graphs of semi-analytical methods for 𝛼 = 0.6 with exact solution for Example 1

(a) MMRDTM (b) MRDTM

Fig. 2. Comparison graphs of semi-analytical methods with 𝛼 = 0.8 for Example 1

Table 1 and Table 2 show the error analyses of the solutions yielded by the MMRDTM and MRDTM
respectively for Example 1 with different values of 𝛼.

138
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

Table 1
Error comparison between MMRDTM and MRDTM for Example 1 with 𝛼 = 0.6
Time, 𝑡 Exact Solution Absolute Error MMRDTM Absolute Error MRDTM
0.1 1.349858808 0 0.282102680
0.2 1.105170918 0 0.2377261703
0.3 0.9048374180 5.000000x10-10 0.2476273192
0.4 0.7408182207 3.000000x10-10 0.4234119704
0.5 0.6065306597 5.000000x10-10 0.8803608197
0.6 0.4965853038 7.000000x10-10 1.750204400
0.7 0.4065696597 3.100000x10-9 3.182896167
0.8 0.3328710837 2.000000x10-9 5.346344380
0.9 2.245564710 2.600000x10-9 8.425688724
1.0 0.2231301601 2.59000x10-8 12.62251238

Table 2
Error comparison between MMRDTM and RDTM for Example 1 with 𝛼 = 0.8
Time, 𝑡 Exact Solution Absolute Error MMRDTM Absolute Error MRDTM
0.1 1.349858808 0 0.130571465
0.2 1.105170918 0 0.1114617051
0.3 0.9048374180 1.0000000x10-10 0.0694857406
0.4 0.7408182207 2.000000x10-10 0.0317849732
0.5 0.6065306597 1.000000x10-10 0.0178225805
0.6 0.4965853038 8.000000x10-10 0.0507412754
0.7 0.4065696597 1.000000x10-10 0.1622436105
0.8 0.3328710837 7.000000x10-10 0.3956705629
0.9 0.2725317930 2.000000x10-9 0.8083715096
1.0 0.2231301601 1.000000x10-10 1.473732968

Example 2 The one-dimensional fractional nonlinear Telegraph equation is taken into consideration

𝛼
𝑢𝑡𝑡 − 𝑢𝑥𝑥 + 2𝑢𝑡𝛼 + 𝑢2 = 𝑒 −2𝑡 cosh2 𝑥 − 2𝑒 −𝑡 cosh 𝑥, (17)

subject to the initial condition

𝑢(𝑥, 0) = cosh 𝑥,

𝑢𝑡 (𝑥, 0) = − cosh 𝑥.

The exact solution of this equation for α = 1 is 𝑒 −𝑡 cosh 𝑥.

Firstly, apply MMRDTM to Eq. (17) and use the basic properties of MMRDTM; we will get

Γ(1+𝑟𝛼) 𝜕2 Γ(1+(1+𝑟)𝛼) (−2)𝑟

𝑈𝑟+2,𝑖 (𝑥) = (Γ(1+(2+𝑟)𝛼)) (𝜕𝑥 2 (𝑈𝑟,𝑖 (𝑥)) − ∑𝑛𝑟=0 𝐴𝑟,𝑖 − 2 ( ) 𝑈𝑟+1,𝑖 + ( ) cosh2 𝑥 −
Γ(1+𝑟𝛼) 𝑟!
(−2)𝑟
( ) cosh 𝑥) .
𝑟!

From the initial condition, we write

𝑈0 (𝑥) = 𝑒 𝑥 .

𝑇
Divide the interval [0, 𝑇] into 𝑀 subintervals by using the nodes 𝑡𝑖 = 𝑖ℎ of equal step size ℎ = 𝑀 such
that [𝑡𝑖−1 , 𝑡𝑖 ], 𝑖 = 1,2, … ,10. After that, the RDTM is used to solve the initial value problem over the

139
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

interval [0, 𝑡1 ]. Then, use the initial conditions 𝑢(𝑥, 0) = 𝑓0 (𝑥) and 𝑢1 (𝑥, 0) = 𝑓1 (𝑥) to obtain the
following approximate solution

𝑢1 (𝑥, 𝑡) = ∑𝑅𝑟=0 𝑈𝑟,1 (𝑥)𝑡 𝛼𝑟 ,

where 0 ≤ 𝑡 ≤ 𝑡1 . For 𝑖 ≥ 2, we use the initial conditions 𝑢𝑖 (𝑥, 𝑡𝑖−1 ) = 𝑢𝑖−1 (𝑥, 𝑡𝑖−1 ) and
(𝜕⁄𝜕𝑡)𝑢𝑖 (𝑥, 𝑡𝑖−1 ) at each subinterval [𝑡𝑖−1 , 𝑡𝑖 ]. Then RDTM is used to solve the initial value problem
over the interval [𝑡𝑖−1 , 𝑡𝑖 ], where 𝑡𝑖−1 replaces 𝑡0 . This procedure is repeatedly continued to create
a sequence of approximate solutions 𝑢𝑖 (𝑥, 𝑡), where 𝑖 = 1,2, … ,10, for the solution 𝑢(𝑥, 𝑡) such as

𝑢𝑖 (𝑥, 𝑡) = ∑𝑅𝑟=0 𝑈𝑟,𝑖 (𝑥)(𝑡 − 𝑡𝑖−1 )𝛼𝑟

where 𝑡 ∈ [𝑡𝑖−1 , 𝑡𝑖 ]. In fact, the MMRDTM takes the following solution

𝑢1 (𝑥, 𝑡) 𝑡 ∈ [0, 𝑡1 ]
𝑢 (𝑥, 𝑡) 𝑡 ∈ [𝑡1 , 𝑡2 ]
𝑢(𝑥, 𝑡) = { 2
⋮ ⋮
𝑢𝐾 (𝑥, 𝑡) 𝑡 ∈ [𝑡9 , 𝑡10 ].

Figure 3(a) is the exact solution compare to the approximate solutions by the proposed MMRDTM
and MRDTM in Figure 3(b) and Figure 3(c) respectively with 𝛼 = 0.5 for the Example 2. The outcomes
of the approximate solution for 𝛼 = 0.8 are shown in Figure 4(a) for the MMRDTM and Figure 4(b)
for the MRDTM respectively. From the obtained results, once again the approximations produced by
the MMRDTM look very close to the exact solutions compared to the MRDTM.

(a) Exact solution (b)MMRDTM (c)MRDTM

Fig. 3. Comparison graphs of semi-analytical methods for 𝛼 = 0.5 with exact solution for Example 2

Table 3 and Table 4 respectively tabulate the error analyses of the solutions yeilded by the
MMRDTM and MRDTM for Example 2 with different values of 𝛼. Just like the results obtained for the
Example 1, the applied MMRDTM with 𝛼 = 0.5 and 𝛼 = 0.8 is evidently approximates the exact
solution with high accuracy compared to the MRDTM for Example 2.

140
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

(a) (b)
Fig. 4. Comparison graphs of semi-analytical methods with 𝛼 = 0.8 for Example 2

Table 3
Error comparison between MMRDTM and MRDTM for Example 2 with 𝛼 = 0.5
Time, 𝑡 Exact Solution Absolute Error MMRDTM Absolute Error RDTM
0.1 1.020318167 0 0.8428996819
0.2 0.9232220555 1.000000000×10-10 0.7546500574
0.3 0.8353658610 3.041381265×10-10 0.6889462647
0.4 0.7558702887 1.004987562×10-10 0.6262301629
0.5 0.6839397204 4.000000000×10-10 0.5561971833
0.6 0.6188542508 4.000000000×10-10 0.4712445733
0.7 0.5599624824 8.000000000×10-10 0.3648586583
0.8 0.5066750067 8.000000000×10-10 0.2310455469
0.9 0.4584585049 4.200000000×10-9 0.0640865569
1.0 0.4148304099 6.900000000×10-9 -0.141581026

Table 4
Error comparison between MMRDTM and RDTM for Example 2 with 𝛼 = 0.8
Time, 𝑡 Exact Solution Absolute Error MMRDTM Absolute Error RDTM
0.1 1.020318167 0 0.0554479257
0.2 0.9232220555 0 0.0606845490
0.3 0.8353658610 0 0.0536799417
0.4 0.7558702887 0 0.0413943693
0.5 0.6839397204 0 0.0270231605
0.6 0.6188542508 0 0.0125213638
0.7 0.5599624824 0 0.0005954892
0.8 0.5066750067 1.0×10-10 0.0108908191
0.9 0.4584585049 1.0×10-10 0.0168058678
1.0 0.4148304099 4.0×10-10 0.0165290344

4. Conclusions

We presented a new application of the MMRDTM for solving fractional nonlinear telegraph
equations (FNLTEs) with source term. We modified the method by substituting the nonlinear term
with its Adomian polynomials in the multi-step approach. Consequently, the approximated solutions
are obtained with high accuracy as demonstrated in Example 1 and 2. In a nutshell, MMRDTM
outperformed the MRDTM to acquire approximate solution for the FNLTEs in terms of practicality,
consistencency and accuracy. All demonstrated calculations in this paper were obtained by using
Maple 2022 software.

141
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

Acknowledgement
This study is funded by the research grant from Universiti Malaysia Sabah (SLB2110).

References
[1] Kilbas, Anatoliĭ Aleksandrovich, Hari M. Srivastava, and Juan J. Trujillo. Theory and applications of fractional
differential equations. Vol. 204. elsevier, 2006.
[2] Cui, Mingrong. "Compact finite difference method for the fractional diffusion equation." Journal of Computational
Physics 228, no. 20 (2009): 7792-7804. https://doi.org/10.1016/j.jcp.2009.07.021
[3] Chen, Chang-Ming, Fawang Liu, Ian Turner, and Vo Anh. "A Fourier method for the fractional diffusion equation
describing sub-diffusion." Journal of Computational Physics 227, no. 2 (2007): 886-897.
https://doi.org/10.1016/j.jcp.2007.05.012
[4] Momani, Shaher. "Analytic and approximate solutions of the space-and time-fractional telegraph
equations." Applied Mathematics and Computation 170, no. 2 (2005): 1126-1134.
https://doi.org/10.1016/j.amc.2005.01.009
[5] Irandoust-pakchin, Safar, Mehdi Dehghan, Somayeh Abdi-mazraeh, and Mehrdad Lakestani. "Numerical solution
for a class of fractional convection–diffusion equations using the flatlet oblique multiwavelets." Journal of Vibration
and Control 20, no. 6 (2014): 913-924. https://doi.org/10.1177/1077546312470473
[6] Dehghan, Mehdi, S. A. Yousefi, and A. Lotfi. "The use of He's variational iteration method for solving the telegraph
and fractional telegraph equations." International Journal for Numerical Methods in Biomedical Engineering 27, no.
2 (2011): 219-231. https://doi.org/10.1002/cnm.1293
[7] Saadatmandi, Abbas, and Mehdi Dehghan. "A tau approach for solution of the space fractional diffusion
equation." Computers & Mathematics with Applications 62, no. 3 (2011): 1135-1142.
https://doi.org/10.1016/j.camwa.2011.04.014
[8] Jassim, Hassan Kamil, and Habeeb Kadmim. "Fractional Sumudu decomposition method for solving PDEs of
fractional order." Journal of Applied and Computational Mechanics (2020).
[9] Eckstein, Eugene C., Jerome A. Goldstein, and Mark Leggas. "The mathematics of suspensions: Kac walks and
asymptotic analyticity." In Proceedings of the Fourth Mississippi State Conference on Difference Equations and
Computational Simulations, vol. 3, pp. 39-50. 1999.
[10] Orsingher, Enzo, and Xuelei Zhao. "Iterated processes and their applications to higher order differential
equations." Acta Mathematica Sinica 15, no. 2 (1999): 173-180. https://doi.org/10.1007/BF02650660
[11] Joice Nirmala, R., and K. Balachandran. "Analysis of solutions of time fractional telegraph equation." Journal of the
Korean Society for Industrial and Applied Mathematics 18, no. 3 (2014): 209-224.
https://doi.org/10.12941/jksiam.2014.18.209
[12] Vyawahare, Vishwesh A., and P. S. V. Nataraj. "Fractional-order modeling of neutron transport in a nuclear
reactor." Applied Mathematical Modelling 37, no. 23 (2013): 9747-9767.
https://doi.org/10.1016/j.apm.2013.05.023
[13] Madhukar, Amit, Yeonsoo Park, Woojae Kim, Hans Julian Sunaryanto, Richard Berlin, Leonardo P. Chamorro, Joseph
Bentsman, and Martin Ostoja-Starzewski. "Heat conduction in porcine muscle and blood: experiments and time-
fractional telegraph equation model." Journal of the Royal Society Interface 16, no. 160 (2019): 20190726.
https://doi.org/10.1098/rsif.2019.0726
[14] Jiang, Wei, and Yingzhen Lin. "Representation of exact solution for the time-fractional telegraph equation in the
reproducing kernel space." Communications in Nonlinear Science and Numerical Simulation 16, no. 9 (2011): 3639-
3645. https://doi.org/10.1016/j.cnsns.2010.12.019
[15] Kumar, Alpesh, Akanksha Bhardwaj, and Shruti Dubey. "A local meshless method to approximate the time-
fractional telegraph equation." Engineering with Computers 37 (2021): 3473-3488.
https://doi.org/10.1007/s00366-020-01006-x
[16] Wang, Ying, and Liquan Mei. "Generalized finite difference/spectral Galerkin approximations for the time-fractional
telegraph equation." Advances in Difference Equations 2017, no. 1 (2017): 1-16. https://doi.org/10.1186/s13662-
017-1348-2
[17] Kumar, Kamlesh, Rajesh K. Pandey, and Swati Yadav. "Finite difference scheme for a fractional telegraph equation
with generalized fractional derivative terms." Physica A: Statistical Mechanics and its Applications 535 (2019):
122271. https://doi.org/10.1016/j.physa.2019.122271
[18] Saha Ray, S. "Numerical solutions and solitary wave solutions of fractional KDV equations using modified fractional
reduced differential transform method." Computational Mathematics and Mathematical Physics 53, no. 12 (2013):
1870-1881. https://doi.org/10.1134/S0965542513120142

142
Journal of Advanced Research in Applied Sciences and Engineering Technology
Volume 31, Issue 1 (2023) 132-143

[19] Hassani, Hossein, Zakieh Avazzadeh, and JA Tenreiro Machado. "Numerical approach for solving variable-order
space–time fractional telegraph equation using transcendental Bernstein series." Engineering with Computers 36
(2020): 867-878. https://doi.org/10.1007/s00366-019-00736-x
[20] Hussin, Che Haziqah Che, Adem Kilicman, and Amirah Azmi. "Analytical Solutions Of Nonlinear Schrodinger
Equations Using Multistep Modified Reduced Differential Transform Method." Compusoft 7, no. 11 (2018): 2939-
2944.
[21] Che, Hussin Che Haziqah, Ahmad Izani Ismail, Adem Kilicman, and Amirah Azmi. "Analytical solutions of nonlinear
Klein-Gordon equations using multistep modified reduced differential transform method." Thermal Science 23, no.
Suppl. 1 (2019): 317-326. https://doi.org/10.2298/TSCI181015045C
[22] Hussin, Che Haziqah Che, Amirah Azmi, Ahmad Izani Md Ismail, Adem Kilicman, and Ishak Hashim. "Approximate
Analytical Solutions of Bright Optical Soliton for Nonlinear Schrödinger Equation of Power Law
Nonlinearity." Baghdad Sci J 18, no. 1 (2021): 836-845. https://doi.org/10.21123/bsj.2021.18.1(Suppl.).0836
[23] Hussin, Che Haziqah Che, Amirah Azmi, and Adem Kilicman. "Solitary Wave Solutions with Compact Support for
The Nonlinear Dispersive K (m, n) Equations by Using Approximate Analytical Method." Journal of Advanced
Research in Fluid Mechanics and Thermal Sciences 88, no. 1 (2021): 24-34.
https://doi.org/10.37934/arfmts.88.1.2434
[24] Hussin, Che Haziqah Che, Arif Mandangan, Amirah Azmi, and Adem Kilicman. "Solitary Wave Solutions for Forced
Nonlinear Korteweg-de Vries Equation by Using Approximate Analytical Method." Journal of Advanced Research in
Fluid Mechanics and Thermal Sciences 99, no. 2 (2022): 197-206. https://doi.org/10.37934/arfmts.99.2.197206
[25] Saadatmandi, Abbas, and Mohadeseh Mohabbati. "Numerical solution of fractional telegraph equation via the tau
method." Math. Rep 17, no. 67 (2015): 2.
[26] Kataria, K. K., and P. Vellaisamy. "Simple parametrization methods for generating Adomian
polynomials." Applicable analysis and discrete Mathematics 10, no. 1 (2016): 168-185.
https://doi.org/10.2298/AADM160123001K
[27] Al-Smadi, Mohammed, Asad Freihat, Hammad Khalil, Shaher Momani, and Rahmat Ali Khan. "Numerical multistep
approach for solving fractional partial differential equations." International Journal of Computational Methods 14,
no. 03 (2017): 1750029. https://doi.org/10.1142/S0219876217500293
[28] Ebenezer, Bonyah, Asad Freihet, Muhammad Altaf Khan, Asfandiyar Khan, and Saeed Islam. "Application of multi-
step differential transform method to solve system of nonlinear fractional differential algebraic equations." Journal
of applied environmental and biological sciences 6, no. 5 (2016): 83-95.
[29] Momani, Shaher, O. Abu Arqub, Asad Freihat, and Mohammed Al-Smadi. "Analytical approximations for Fokker-
Planck equations of fractional order in multistep schemes." Applied and computational mathematics 15, no. 3
(2016): 319-330.

143