Progr. Fract. Differ. Appl. 9, No.

3, 421-425 (2023) 421

Progress in Fractional Differentiation and Applications
An International Journal

http://dx.doi.org/10.18576/pfda/090307

Approximate Scheme for Fractional Differential Equation

of Order 1 < α < 2 Using Finite Difference Method
Dhrumit K. Trivedi and Nimisha Pathak∗
Department of Applied Mathematics, The Maharaja Sayajirao University of Baroda, Vadodara, Gujarat, India

Received: 2 May 2021, Revised: 18 Jun. 2021, Accepted: 1 Jul. 2021

Published online: 1 Jul. 2023

Abstract: In this article, we develop composite Simpson’s ( 13 )rd rule for fractional differential equation with Caputo derivative of
order α , 1<α < 2 using finite difference method. We use forward, central and backward difference formulae for approximating the
derivatives. We give comparison of numerical solution given by Simpson’s ( 13 )rd rule and Trapezoidal rule. Result shows that Simpson’s
rule gives better approximation. Furthermore, as an application of the scheme we solve test problems using Mathematical software.

Keywords: Fractional differential equation, composite Simpson’s ( 31 )rd rule, finite difference method, Matlab.

1 Introduction

The fractional calculus is a name for the theory of integrals and derivatives of arbitrary real or complex order which have
been focus of many studies due to their application in various areas of physics and engineering [1,2,3,4,5,6,7,8]. There
are efficient methods for exact and numerical solutions of fractional differential equations. Particularly, Exact solutions
of fractional differential equations can not be obtained easily. Some of these numerical methods include: Variational
Iteration Method [9], Extrapolation Method [10], General (n+1)-explicit finite difference formulas [11] and fractional
finite difference method (FFDM).
In this article, we consider linear fractional differential equation containing Caputo fractional derivative. We develop
a new technique by which fractional derivatives of u(t) are approximated by finite difference method. The composite
Simpson’s ( 31 )rd rule [12] is applied to approximate the integral part in Caputo definition of order 1 < α < 2, then FFDM
is used to approximate first, second and third order derivatives. Our method efficiency is demonstrated by comparing
approximate solutions for fractional differential equations with the exact solutions or with other approximate solutions
those are obtained by other solvers and methods. Moreover, MATLAB is used to solve these differential equations.

1.1 Preliminaries

In this section, we will introduce definitions of fractional derivative. Also, we will provide commonly used finite
difference formulae and composite Simpson’s rule for definite integral.

Definition 1: The Riemann – Liouville fractional derivative [13] is defined as

!n Z
1 d t
(Dα u)(t) = (t − x)n−α −1u(x)dx, t ∈ [a, b], (1)
Γ (n − α ) dt a

∗ Corresponding author e-mail: nimisha.pathak-appmath@msubaroda.ac.in

c 2023 NSP
Natural Sciences Publishing Cor.
422 D. Trivedi et al. : Approximate scheme for Linear fractional differential ...

where n is a positive integer satisfying n − 1 < α ≤ n, α ∈ R+ .

Definition 2: The Caputo fractional derivative [13] for n − 1 < α ≤ n, α ∈ R+ is defined as

Z t
α 1
(D u)(t) = (t − x)n−α −1 un (x)dx, t ∈ [a, b]. (2)
Γ (n − α ) a

1.1.1 Finite difference formulae

Finite difference formulae are commonly used to solve differential equations numerically. Suitable finite difference
formulae are used to replace derivative terms on a discrete domain. The accuracy is increased if the number of mesh
points are increased. In general, Taylor series is used to obtained finite difference formulae.
′′
For u (x), the finite difference formula is

′′ u(x + 2h) − 2u(x + h) + u(x)

u (x) = + O(h). (3)
h2
′′′
The forward finite difference formula for u (x) is

′′′ −u(x) + 3u(x + h) − 3u(x + 2h) + u(x + 3h)

u (x) = + O(h). (4)
h3
′′′
The central finite difference formula for u (x) is

′′′ −u(x − 2h) + 2u(x − h) − 2u(x + h) + u(x + 2h)

u (x) = + O(h2 ). (5)
2h3
′′′
The backward finite difference formula for u (x) is

′′′ u(x) − 3u(x − h) + 3u(x − 2h) − u(x − 3h)

u (x) = + O(h). (6)
h3
However, there are number of methods for estimating definite integrals, we will use the following composite Simpson’s
rule [12].
Theorem 1: Let u ∈ C4 [a,b], n be even, h = b−c
n , and xi = a + ih, for i = 0, 1, 2, ....n. There exists a γ ∈(a,b) for which the
composite Simpson’s rule for n subintervals can be written as
n n
Z b 2 −1 2
h  b−a 4 4
u(x)dx = u(a) + 2 ∑ u(x2i ) + 4 ∑ u(x2i−1 ) + u(b) − h u (γ ). (7)
a 3 i=1 i=1 180

2 FFDM for the fractional derivative

In this section, we introduce method of approximating the Caputo fractional derivative. We used Caputo definition of
fractional derivatives because it has a benefit of dealing properly with the initial value problem which is given in terms of
their integer order and the field variables.
For a = 0, n = 1,t ≥ 0 and 1 < α < 2, the Caputo definition for fractional derivative equation (2) is defined as
Z t
1 ′′
(Dα u)(t) = (t − x)1−α u (x)dx. (8)
Γ (2 − α ) 0

In right hand side of equation (8), we apply integration by parts and get,
" Zt
#
α 1 2−α ′′ 2−α ′′′
(D u)(t) = (t) u (0) + (t − x) u (x)dx . (9)
(2 − α )Γ (2 − α ) 0

c 2023 NSP
Natural Sciences Publishing Cor.
Progr. Fract. Differ. Appl. 9, No. 3, 421-425 (2023) / www.naturalspublishing.com/Journals.asp 423

The Composite Simpson’s rule is used to approximate the last integral in (9), we get,
" n
Z t 2 −1
2−α ′′′ h 2−α ′′′ ′′′
(t − x) u (x)dx ≈ (t) u (0) + 2 ∑ (t − x2i )2−α u (x2i )
0 3 i=1
n
(10)
2 −1

2−α ′′′ 2−α ′′′
+ 4 ∑ (t − x2i−1) u (x2i−1 ) + (t − b) u (b) ,
i=1
b–a
where h = n and xi = a + ih, for i = 0, 1, 2, ...n.

Now, substituting value of equation (10) in equation (9), we get

" "
α 1 2−α ′′ h ′′′
(D u)(t) = (t) u (0) + (t)2−α u (0)
(2 − α )Γ (2 − α ) 3
n n
2 −1 2 −1
2−α ′′′ ′′′
+ 2 ∑ (t − x2i ) u (x2i ) + 4 ∑ (t − x2i−1)2−α u (x2i−1 ) (11)
i=1 i=1
##
2−α ′′′
+ (t − b) u (b) .

′′ ′′′
For small value of h, we approximate u (0) and u (0) using forward finite difference formula for 2rd order
′′′ ′′′
derivative, equation (3), and 3rd order derivative, equation (4) respectively. Also, u (x2i ) and u (x2i−1 ) can be
rd
approximated using central finite difference formula for 3 order derivative, equation (5). Finally, backward finite
′′′
difference formula for 3rd order derivative in equation (6) is used to approximate u (b).
′′ ′′′ ′′′ ′′′ ′′′
By substituting values of u (0), u (0), u (x2i ), u (x2i−1 ) and u (b) in equation (11), we obtain
" #
1 h
(Dα u)(t) =
 
P + Q + 2R + 4S + T . (12)
(2 − α )Γ (2 − α ) 3

where

u(2h) − 2u(h) + u(0)

P = (t)2−α ,
h2
−u(0) + 3u(h) − 3u(2h) + u(3h)
Q = (t)2−α ,
h3
n −1
2
u(x2i − 2h) + 2u(x2i − h) − 2u(x2i + h) + u(x2i + 2h)
R= ∑ (t − x2i)2−α 2h3
,
i=1
n
2
−u(x2i−1 − 2h) + 2u(x2i−1 − h) − 2u(x2i−1 + h) + u(x2i−1 + 2h)
S = ∑ (t − x2i−1 )2−α ,
i=1 2h3

u(b) − 3u(b − h) + 3u(b − 2h) − u(b − 3h)

T = (t − b)2−α .
h3
Now, we can find Dα u(t) approximately using equation (12) which depends on the value of h. If we decrease value of
h, then approximate value of Dα u(t) will be more accurate.

3 Applications

The following examples will be solved using in Matlab for fractional differential equations of order 1 < α < 2 and
compare with the exact solutions. We will use equation (12) for the approximation of Dα u(t). Example 3.1 [14] Consider

c 2023 NSP
Natural Sciences Publishing Cor.
424 D. Trivedi et al. : Approximate scheme for Linear fractional differential ...

Table 1: Numerical solution of example 3.1 with h = 0.1

T Exact Approximate Abs. Error Approximate Abs. Error
(Trapezoidal Rule) (Simpson’s (1/3)rd Rule)
0.1 0.01 0.00999999999999998 2.26E-17 0.0100000000000000 0.00E+00
0.2 0.04 0.03999999999999980 2.08E-16 0.0399999999999999 1.11E-16
0.3 0.09 0.08999999999999940 5.97E-16 0.0900000000000000 0.00E+00
0.4 0.16 0.15999999999999900 1.03E-15 0.1599999999999990 1.03E-15
0.5 0.25 0.24999999999999800 2.00E-15 0.2500000000000000 0.00E+00
0.6 0.36 0.35999999999999700 3.00E-15 0.3600000000000000 0.00E+00
0.7 0.49 0.48999999999999700 2.94E-15 0.4900000000000000 0.00E+00
0.8 0.64 0.63999999999999600 4.11E-15 0.6400000000000000 0.00E+00
0.9 0.81 0.80999999999999400 6.11E-15 0.8100000000000000 0.00E+00
1 1 0.99999999999999300 6.99E-15 1.0000000000000000 0.00E+00

Table 2: Numerical solution of example 3.2 with h = 0.1

T Exact Approximate Abs. Error Approximate Abs. Error
(Trapezoidal Rule) (Simpson’s (1/3)rd Rule)
0.1 0.01 0.01000000000000320 3.20E-15 0.00999999999999845 1.55E-15
0.2 0.04 0.04000000000000930 9.29E-15 0.04000000000000150 1.49E-15
0.3 0.09 0.09000000000001480 1.48E-14 0.09000000000000420 4.20E-15
0.4 0.16 0.16000000000001900 1.90E-14 0.15999999999998800 1.20E-14
0.5 0.25 0.250000000000023000 2.30E-14 0.25000000000000000 0.00E+00
0.6 0.36 0.36000000000002800 2.80E-14 0.36000000000000000 0.00E+00
0.7 0.49 0.49000000000003300 3.31E-14 0.48999999999998500 1.49E-14
0.8 0.64 0.64000000000003800 3.79E-14 0.64000000000000000 1.11E-16
0.9 0.81 0.81000000000004300 4.30E-14 0.80999999999996400 3.61E-14
1 1 1.00000000000004000 4.00E-14 1.00000000000000000 0.00E+00

the following fractional differential equation:

r
t
Dα u(t) = t α u(t) + 4 − t α +2, (13)
π
′
with given conditions u(0) = 0, u (0) = 0.

For α = 1.5, the exact solution is u(t) = t 2 . The results are obtained using (12) for h = 0.1 are given in Table 1. We
compare results with solution given in [15] using trapezoidal method and also with exact solutions.

Example 3.2 [14] For the fractional differential equation:

r
2 1.5 t
D u(t) + D u(t) + u(t) = 4 − t 2 + 2, (14)
π

with given condition u(0) = 0, the exact solution is u(t) = t 2 . The results are for h = 0.1 are given in Table 2.

4 Conclusion

In this work, we derived composite Simpson’s ( 31 )rd rule for fractional differential equation with Caputo derivative of order
α , 1<α < 2 using finite difference method. We used forward, central and backward difference formulae for approximating
the derivatives. Finally, numerical examples are simulated to demonstrate efficiency of given method. We observed that
the obtained results using Matlab software gave very less absolute error compared with the exact solution. We gave
comparison of numerical solutions given by Simpson’s ( 13 )rd rule and Trapezoidal rule. Result shows that Simpson’s rule
gives better approximation.

c 2023 NSP
Natural Sciences Publishing Cor.
Progr. Fract. Differ. Appl. 9, No. 3, 421-425 (2023) / www.naturalspublishing.com/Journals.asp 425

References
[1] B. J. West, M. Bologna and P. Grigolini, Physics of fractal operators, Springer, Berlin, Heidelberg and New York, 2003.
[2] F. Mainardi, Fractional calculus and waves in linear viscoelasticity: an introduction to mathematical models, World Scientific
Publishing Company, Singapore, New Jersey, London and Hong Kong, 2010.
[3] A. A. Kilbas, H. M. Srivastava and J. J. Trujillo, Theory and applications of fractional differential equations, North-Holland
Mathematical Studies, Vol. 204, Elsevier (North-Holland) Science Publishers, Amsterdam, London and New York, 2006.
[4] M. D. Ortigueira, Fractional calculus for scientists and engineers, Springer, Berlin, Heidelberg and New York, 2011.
[5] D. Baleanu, K. Diethelm, E. Scalas and J. J. Trujillo, Fractional calculus: models and numerical methods, Series on Complexity,
Nonlinearity and Chaos, Vol. 3, World Scientific Publishing Company, Singapore, New Jersey, London and Hong Kong, 2012.
[6] R. L. Magin, Fractional calculus in bioengineering, Begerll House, West Redding, Connecticut, 2006.
[7] H. Sheng, Y. Q. Chen and T. S. Qiu,Fractional processes and fractional-order signal processing: techniques and applications,
Springer, Berlin, Heidelberg and New York, 2012.
[8] O. P. Agrawal, A general formulation and solution scheme for fractional optimal control problems, Nonlin. Dyn. 38, 323-337
(2004).
[9] W. Guo-Cheng and E. W. M. Lee, Fractional variational iteration method and its application, Phy. Lett. A 374(25), 2506-2509
(2010).
[10] K. Diethelm and G. Walz, Numerical solution of fractional order differential equations by extrapolation, Numer. Alg. 16(2), 231-
253 (1997).
[11] R. B. Albadarneh, N. T. Shawagfeh and Z. S. Abo-Hammour, General (n+1)-explicit finite difference formulas with proof, Appl.
Math. Sci. 6(20), 995-1009 (2012).
[12] R. L. Burden and J. D. Faires, Numer. Anal. 3rd ed., PWS Publishers, USA, 1986.
[13] I.Podlubny, Fractional differential equation, Mathematics in Science and Engineering, 198, Academic Press, San Diego, CA, 1999.
[14] D. Rostamy, M. Alipour, H. Jafari and D. Baleanu, Solving multi-term orders fractional differential equations by operational
matrices of BPs with convergence analysis, Romanian Report. Phy. 65(2), 334–349 (2013).
[15] R. B. Albadarneh, I. M. Batiha and M. Zurigat, Numerical solutions for linear fractional differential equations of order 1 < α < 2
using finite difference method (FFDM), Int. J. Math. Comput. Sci. 16(1), 103–111 (2016).

c 2023 NSP
Natural Sciences Publishing Cor.