International Journal of Scientific and Research Publications, Volume 9, Issue 3, March 2019 367

ISSN 2250-3153

On The Numerical Solution of Picard Iteration Method

for Fractional Integro - Differential Equation
*1
Ilejimi D.O, 2Okai J.O and 3Raheem R.L
*1 Department of Mathematical Sciences, Federal University of Agriculture Abeokuta
2Department of Mathematical Sciences, Abubakar Tafawa Balewa University, Bauchi, Nigeria
3African Institute for Mathematical Sciences, Ghana.

DOI: 10.29322/IJSRP.9.03.2019.p8757
http://dx.doi.org/10.29322/IJSRP.9.03.2019.p8757

Abstract: In this paper, the concept of Successive Approximation method also called the Picard Iteration Method
(PIM) for solving Fractional integro-differential equations is introduced. The fractional derivative is considered in
the Caputo sense [6]. The proposed method reduces the equation into a standard integro integral equation of the
second kind. Some test problems are considered to demonstrate the accuracy and the convergence of the
presented method.
Numerical results show’s that this approach is easy and accurate when applied to fractional integro-differential
equations.
Keywords: Picard Iteration method; Volterra fractional integro-differential equation; Fredholm fractional integro-
differential equation.

I. Introduction
Fractional calculus has a long history from 30 September 1695, when the derivative of order 𝛼𝛼 = 1/2 has been
described by Leibniz [7]. The theory of derivatives and integrals of non-integer order goes back to Leibniz,
Liouville, Grȕnwald, Letnikov and Riemann. There are many interesting books about fractional calculus and
fractional differential equations [7]. Our main focus is the Caputo definition which turns out to be of great
usefulness in real-life applications.
In the past decade, mathematicians have devoted effort to the study of explicit and numerical solutions to linear
and nonlinear fraction differential equations [3-4]. An extensive amount of research has been done on fractional
calculus, such as [5]. Over the past few years, a number of fractional calculus applications are being used and in the
field of science, engineering and economics [6]. Research on linear and non-linear differential equations and
linearization techniques has gained quite a momentum due the rapidly proliferating use and recent developments
of fractional calculus in these fields [5].

II. Preliminaries
Definition 1
The Riemann–Liouville fractional integral operator of order α ≥ 0 of the function f ∈ Cµ, µ ≥ −1 is defined as

𝑥𝑥
1
𝐽𝐽α 𝑓𝑓(𝑥𝑥) = � (𝑥𝑥 − µ)α−1 𝑓𝑓(µ)𝑑𝑑µ, (1)
Γ(α) 𝑎𝑎

http://dx.doi.org/10.29322/IJSRP.9.03.2019.p8757 www.ijsrp.org
International Journal of Scientific and Research Publications, Volume 9, Issue 3, March 2019 368
ISSN 2250-3153

α > 0, x > 0

𝐽𝐽α 𝑓𝑓(𝑥𝑥) = 𝑓𝑓(𝑥𝑥) (2)

When we formulate the model of real-world problems with fractional calculus, The Riemann–Liouville have
certain disadvantages. Caputo proposed in his work on the theory of viscoelasticity [9] a modified fractional
differential operator 𝐷𝐷∗α .

Definition 2
The fraction derivative of 𝑓𝑓(𝑥𝑥) in Caputo sense is defined as

𝑥𝑥
1
𝐷𝐷𝛼𝛼 𝑓𝑓(𝑥𝑥) = 𝐽𝐽𝑚𝑚−α 𝐷𝐷𝑚𝑚 𝑓𝑓(𝑥𝑥) = � (𝑥𝑥 − η)𝑚𝑚−α−1 𝑓𝑓 (𝑚𝑚) (η) 𝑑𝑑η
Γ(m − α) 0

For (3)
𝑚𝑚
m − 1 < α ≤ m, m ∈ N, x > 0, f ∈ 𝐶𝐶−1
Definition 3
The left sided Riemann–Liouville fractional integral operator of the order 𝜇𝜇3 0 of a function 𝑓𝑓 ∈ 𝐶𝐶𝜇𝜇 ,
𝜇𝜇3 − 1, is defined as

1 1 𝑓𝑓(𝑥𝑥)
𝑗𝑗 𝛼𝛼 𝑓𝑓(𝑥𝑥) = ∫ 𝑑𝑑𝑑𝑑, 𝛼𝛼 > 0, 𝑥𝑥 > 0, (4)
𝛤𝛤(𝑥𝑥) 0 (𝑥𝑥−𝑡𝑡)1−𝛼𝛼

𝑗𝑗 𝛼𝛼 𝑓𝑓(𝑥𝑥) = 𝑓𝑓(𝑥𝑥) (5)

Definition 4
𝑚𝑚
Let ∈ 𝐶𝐶−1 1, 𝑚𝑚 ∈ 𝑁𝑁 ∪ {0}. Then the Caputo fractional derivative of 𝑓𝑓(𝑥𝑥) is defined as
𝑚𝑚,(𝑥𝑥), 𝑚𝑚−1 < 𝛼𝛼 ≤ 𝑚𝑚, 𝑚𝑚𝑚𝑚𝑚𝑚,
𝛼𝛼
𝑗𝑗 𝑚𝑚−𝛼𝛼𝑓𝑓
𝐷𝐷 𝑓𝑓(𝑥𝑥) = � 𝐷𝐷𝑚𝑚 𝑓𝑓(𝑥𝑥) (6)
, 𝛼𝛼 = 𝑚𝑚
𝐷𝐷𝑥𝑥 𝑚𝑚

Hence, we have the following properties:

(1) 𝑗𝑗 𝛼𝛼 𝑗𝑗 𝑣𝑣 𝑓𝑓 = 𝑗𝑗 𝛼𝛼+𝑣𝑣 𝑓𝑓, 𝛼𝛼, 𝑣𝑣 > 0, 𝑓𝑓 ∈ 𝐶𝐶𝜇𝜇, 𝜇𝜇 > 0

Γ(γ+1)
𝑗𝑗 𝛼𝛼 𝑥𝑥 𝛾𝛾 = 𝑥𝑥 𝛼𝛼+𝛾𝛾 , 𝛼𝛼 > 0, 𝛾𝛾 > −1, 𝑥𝑥 > 0
Γ(α+γ+1)

𝑥𝑥 𝑘𝑘
𝑗𝑗 𝛼𝛼 𝐷𝐷𝛼𝛼 𝑓𝑓(𝑥𝑥) = 𝑓𝑓(𝑥𝑥) − ∑𝑚𝑚−1 𝑘𝑘 +
𝑘𝑘=0 𝑓𝑓 (0 ) , 𝑥𝑥 > 0, 𝑚𝑚 − 1 < 𝛼𝛼 ≤ 𝑚𝑚
𝑘𝑘

𝐷𝐷𝑗𝑗 𝛼𝛼 𝑓𝑓(𝑥𝑥) = 𝑓𝑓(𝑥𝑥), 𝑥𝑥 > 0, 𝑚𝑚 − 1 < 𝛼𝛼 ≤ 𝑚𝑚

𝐷𝐷𝐷𝐷 = 0, 𝐶𝐶 𝑖𝑖𝑖𝑖 𝑡𝑡ℎ𝑒𝑒 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐

http://dx.doi.org/10.29322/IJSRP.9.03.2019.p8757 www.ijsrp.org
International Journal of Scientific and Research Publications, Volume 9, Issue 3, March 2019 369
ISSN 2250-3153

0, 𝛽𝛽 ∈ 𝑁𝑁0 , 𝛽𝛽 < [𝛼𝛼 ],

� Ã(𝛽𝛽+1)
𝐷𝐷𝛼𝛼 𝑥𝑥 𝛽𝛽 = , 𝛽𝛽 ∈ 𝑁𝑁0 , 𝛽𝛽 ≥ [𝛼𝛼 ]
Ã(𝛽𝛽−𝛼𝛼+1)𝑋𝑋 𝛽𝛽−𝛼𝛼

Where [α] denoted the smallest integer greater than or equal to α and 𝑁𝑁0 = {0,1,2, … }

Definition 5
An integral equation is an equation in which the unknown function 𝑦𝑦(𝑥𝑥) appears under an integral-signs (4) and
(5).

A standard integral equation 𝑦𝑦(𝑥𝑥) is of the form:

ℎ(𝑥𝑥)
𝑦𝑦(𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫𝑔𝑔(𝑥𝑥) 𝐾𝐾(𝑥𝑥, 𝑡𝑡)𝑦𝑦(𝑡𝑡)𝑑𝑑𝑑𝑑 (7)

Where 𝑔𝑔(𝑥𝑥) and 𝑦𝑦(𝑥𝑥) are the limits of integration, λ is a constant parameter, and 𝐾𝐾(𝑥𝑥, 𝑡𝑡) is a function of two
variable 𝑥𝑥 and 𝑡𝑡 called the kernel or the nucleolus of the integral equation. The function 𝑦𝑦(𝑥𝑥) that will be
determined appears under the integral sign and also appears inside and outside the integral sign as well. It is to be
noted that the limits of integration 𝑔𝑔(𝑥𝑥) and ℎ(𝑥𝑥) may be both variables, constants or mixed [8].

Definition 6

An integro-differential equation is an equation in which the unknown function 𝑦𝑦(𝑥𝑥) appears under an integral sign
and contain ordinary derivatives 𝑦𝑦 (𝑛𝑛) (𝑥𝑥) as well. A standard integro-differential equation is of the form:

ℎ(𝑥𝑥)
𝑦𝑦 (𝑛𝑛) (𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫𝑔𝑔(𝑥𝑥) 𝐾𝐾(𝑥𝑥, 𝑡𝑡)𝑦𝑦(𝑡𝑡)𝑑𝑑𝑑𝑑 (8)

Integral equations and integro-differential equations are classified into distinct types according to limits of
integration and the kernel 𝐾𝐾(𝑥𝑥; 𝑡𝑡) are as prescribed before [7]

1. If the limits of the integration are fixed, then the integral equation is called a Fredholm integral
equation and is of the form:
𝑏𝑏
𝑦𝑦(𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫𝑎𝑎 𝐾𝐾(𝑥𝑥, 𝑡𝑡)𝑦𝑦(𝑡𝑡)𝑑𝑑𝑑𝑑 (9)

2. If at least one limits is a variable, then the equation is called a Volterra integral equation and is given
as:

𝑏𝑏
𝑦𝑦(𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫𝑎𝑎 𝐾𝐾(𝑥𝑥, 𝑡𝑡)𝑦𝑦(𝑡𝑡)𝑑𝑑𝑑𝑑 (10)

III. Fundamentals of The Picard Iteration Method

The
successive approximations method, also called the Picard iteration method provides a scheme that can be used
for solving initial value problems or integral equations. This method solves any problem by finding successive
approximations to the solution by starting with an initial guess, called the zeroth approximation. As will be seen,
the zeroth approximation is any selective real-valued function that will be used in a recurrence relation to
determine the other approximations [8].
Given the linear Volterra integral equation of the second kind
http://dx.doi.org/10.29322/IJSRP.9.03.2019.p8757 www.ijsrp.org
International Journal of Scientific and Research Publications, Volume 9, Issue 3, March 2019 370
ISSN 2250-3153
𝑥𝑥
𝑢𝑢(𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫0 𝐾𝐾(𝑥𝑥, 𝑡𝑡)𝑢𝑢(𝑡𝑡)𝑑𝑑𝑑𝑑, (11)
Where u(x) is the unknown function to be determined, 𝐾𝐾(𝑥𝑥, 𝑡𝑡) is the kernel, and λ is a parameter. The
successive approximations method introduces the recurrence relation
𝑥𝑥
𝑢𝑢𝑛𝑛 (𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫0 𝐾𝐾(𝑥𝑥, 𝑡𝑡) 𝑢𝑢𝑛𝑛−1 (𝑡𝑡)𝑑𝑑𝑑𝑑, 𝑛𝑛 ≥ 1, (12)
Where the zeroth approximation 𝑢𝑢0 (𝑥𝑥) can be any selective real valued function. We always start with an initial
guess for 𝑢𝑢0 (𝑥𝑥), mostly we select 0,1, 𝑥𝑥 for 𝑢𝑢0 (𝑥𝑥).
In this paper, we will consider
𝑢𝑢0 (𝑥𝑥) = 0 (13)
and by using (12), several successive approximations 𝑢𝑢𝑘𝑘 , 𝑘𝑘 ≥ 1 will be determined as
𝑥𝑥
𝑢𝑢1 (𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫0 𝐾𝐾(𝑥𝑥, 𝑡𝑡)𝑢𝑢0 (𝑡𝑡)𝑑𝑑𝑑𝑑,

𝑥𝑥
𝑢𝑢2 (𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫0 𝐾𝐾(𝑥𝑥, 𝑡𝑡)𝑢𝑢1 (𝑡𝑡)𝑑𝑑𝑑𝑑,

𝑥𝑥
𝑢𝑢3 (𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫0 𝐾𝐾(𝑥𝑥, 𝑡𝑡)𝑢𝑢2 (𝑡𝑡)𝑑𝑑𝑑𝑑,
.
. (14)
.
𝑥𝑥
𝑢𝑢𝑛𝑛 (𝑥𝑥) = 𝑓𝑓(𝑥𝑥) + 𝜆𝜆 ∫0 𝐾𝐾(𝑥𝑥, 𝑡𝑡)𝑢𝑢𝑛𝑛−1 (𝑡𝑡)𝑑𝑑𝑑𝑑,

IV. Numerical Examples

Program 1:
Consider the following fractional integro-differential equation:
3
6𝑡𝑡 9/4 −𝑡𝑡 2 exp(𝑡𝑡) 𝑡𝑡
𝐷𝐷4 (𝑦𝑦(𝑡𝑡)) = 13 + � � 𝑦𝑦(𝑡𝑡) + ∫0 𝑒𝑒𝑒𝑒𝑒𝑒(𝑡𝑡)𝑥𝑥𝑥𝑥(𝑥𝑥)𝑑𝑑𝑑𝑑, 0 ≤ 𝑥𝑥 ≤ 1, (15)
Γ� � 5
14

Subject to 𝑦𝑦(0) = 0 with exact solution 𝑦𝑦(𝑡𝑡) = 𝑡𝑡 3 (16)

Applying 𝐽𝐽3/4 to both sides of (15) and applying (6) yields

6 1 3 𝑡𝑡 2 exp(𝑡𝑡) 𝑡𝑡
𝑦𝑦(𝑡𝑡) = 13 𝐽𝐽3/4 �𝑡𝑡 9/4 � − 𝐽𝐽4 � � 𝑦𝑦(𝑡𝑡) + 𝐽𝐽3/4 �exp(𝑡𝑡) ∫0 𝑥𝑥𝑥𝑥(𝑥𝑥)𝑑𝑑𝑑𝑑 � (17)
Γ� � 5 5
14

Applying:
𝑦𝑦0 (𝑡𝑡) = 0 and (12), we have the following iterations:
http://dx.doi.org/10.29322/IJSRP.9.03.2019.p8757 www.ijsrp.org
International Journal of Scientific and Research Publications, Volume 9, Issue 3, March 2019 371
ISSN 2250-3153

𝑦𝑦1 (𝑡𝑡) = 𝑦𝑦2 (𝑡𝑡) = 𝑦𝑦3 (𝑡𝑡) = 𝑦𝑦4 (𝑡𝑡) = 𝑦𝑦5 (𝑡𝑡) = 𝑦𝑦6 (𝑡𝑡) = 𝑡𝑡 3

Which coincides with the exact solution.

Problem 2:

Consider the following fractional integro-differential equation:

3
8
1 � �𝑥𝑥 2 −2𝑥𝑥 1/2 𝑥𝑥 1
𝐷𝐷 (𝑢𝑢(𝑥𝑥)) =
2
4
+ + ∫0 𝑥𝑥𝑥𝑥𝑥𝑥(𝑡𝑡)𝑑𝑑𝑑𝑑, 0 ≤ 𝑥𝑥 ≤ 1, (18)
√𝜋𝜋 12
Subject to 𝑢𝑢(0) = 0 with exact solution 𝐵𝐵(𝑥𝑥) = 𝑥𝑥 2 − 𝑥𝑥 (19)

Applying 𝐽𝐽1/2 (Half fractional integral) to both sides of (18) and applying (6) yields

1 1
𝑢𝑢(𝑥𝑥) = 𝑥𝑥 2 − 𝑥𝑥 + 𝐽𝐽0.5 [𝑥𝑥 ] + 𝐽𝐽0.5 �∫0 𝑥𝑥𝑥𝑥𝑥𝑥(𝑡𝑡)𝑑𝑑𝑑𝑑� (20)
2

Choosing 𝑢𝑢0 (𝑥𝑥) = 0 and applying (12), gives:

𝑥𝑥 3/2
𝑢𝑢1 (𝑥𝑥) = 𝑥𝑥 2 − 𝑥𝑥 +
9√𝜋𝜋
2 8𝑥𝑥 3/2
𝑢𝑢2 (𝑥𝑥) = 𝑥𝑥 − 𝑥𝑥 +
189𝜋𝜋
32𝑥𝑥 3/2
𝑢𝑢3 (𝑥𝑥) = 𝑥𝑥 2 +
567𝜋𝜋√𝜋𝜋
256𝑥𝑥 3/2
𝑢𝑢4 (𝑥𝑥) = 𝑥𝑥 2 − 𝑥𝑥 +
11907𝜋𝜋2
2048𝑥𝑥 3/2
𝑢𝑢5 (𝑥𝑥) = 𝑥𝑥 2 − 𝑥𝑥 +
250047𝜋𝜋5/2
3
16384𝑥𝑥 2
𝑢𝑢6 (𝑥𝑥) = 𝑥𝑥 2 − 𝑥𝑥 +
5250987𝜋𝜋3

Therefore, the series solution is

3
2
16384𝑥𝑥 2
𝑢𝑢(𝑥𝑥) = 𝑥𝑥 − 𝑥𝑥 +
5250987𝜋𝜋 3

Table of Numerical results for Problem 2

X Picard(n=6) Exact Solution Abs. Error
0 0 0 0
0.5 -0.249964436 -0.25 3.55644E-05
1 0.000100591 0 0.000100591
1.5 0.750184798 0.75 0.000184798

http://dx.doi.org/10.29322/IJSRP.9.03.2019.p8757 www.ijsrp.org
International Journal of Scientific and Research Publications, Volume 9, Issue 3, March 2019 372
ISSN 2250-3153
2 2.000284515 2 0.000284515
2.5 3.750397622 3.75 0.000397622
3 6.000522688 6 0.000522688
3.5 8.750658662 8.75 0.000658662
4 12.00080473 12 0.00080473
4.5 15.75096024 15.75 0.000960239
5 20.00112465 20 0.001124645

Graphical Representation of Problem 2:

Picard Iteration (n=6)

Picard Iteration(n=6) Exact Solution

25

20

15

10

0
1 2 3 4 5 6 7 8 9 10 11
-5

V. Conclusion
In this paper, Successive approximation method also known as Picard Iteration method (PIM) was
applied to a transformed fractional integro-differential equations, the results when compared with other
methods yields a better results.

Reference

1. Nazari D, Shahmorad S (2010) Application of the fractional differential transform method to

fractional-order integro-differential equations with nonlocal boundary conditions. J Comput
Appl Math 3: 883-891.
2. Caputo M (1967) Linear models of disssipation whose Q is almost frequency Independent. Part
II, J Roy Austral Soc 529-539.
3. K.B. Oldham, J. Spanier, The Fractional Calculus, Academic Press, New York, 1974.
4. K.S. Miller, B. Ross, An Introduction to The Fractional Calculus and Fractional Differential
Equations, Wiley, New York, 1993.

http://dx.doi.org/10.29322/IJSRP.9.03.2019.p8757 www.ijsrp.org
International Journal of Scientific and Research Publications, Volume 9, Issue 3, March 2019 373
ISSN 2250-3153
5. S.S. Ray, K.S. Chaudhuri, R.K. Bera, Analytical approximate solution of nonlinear dynamic system
containing fractional derivative by modified decomposition method, Appl. Math. Comput. 182
(2006) 544_552.
6. S. Das, Functional Fractional Calculus for System Identification and Controls, Springer, New
York, 2008.
7. Oyedepo T, Taiwo OA, Abubakar JU, Ogunwobi ZO (2016) “Numerical Studies for Solving
Fractional Integro-Differential Equations by using Least Squares Method and Bernstein
Polynomials”. Fluid Mech Open Acc 3: 142. doi: 10.4172/2476-2296.1000142.
8. Wazwaz, A.M. (2011). “Linear And Nonlinear Integral Equations”: Methods and Applications.
Springer Saint Xavier University Chicago, USA.
9. M. Caputo, Linear models of dissipation whose Q is almost frequency independent, Part II, J. Roy.
Astr. Soc. 13 (1967) 529.

http://dx.doi.org/10.29322/IJSRP.9.03.2019.p8757 www.ijsrp.org