Annals of the University of Craiova, Mathematics and Computer Science Series

Volume 39(2), 2012, Pages 251256

ISSN: 1223-6934

Approximate analytical solutions to the Bagley-Torvik

equation by the Fractional Iteration Method
Toufik Mekkaoui and Zakia Hammouch
Abstract. In this paper we solve the Bagley-Torvik equation, which is an ordinary fractional
differential equation, where the solution procedure is easier, more effective and straightforward.
The validity and the accuracy of this method is shown by the obtained results.
2010 Mathematics Subject Classification. Primary 26A33, 65L05; Secondary 65K10.
Key words and phrases. Bagley-Torvik equation, Fractional Iteration Method (FIM), Series
solution.

1. Introduction
Owing to their frequent appearances in various applications in fluid mechanics,
viscoelasticity, biology, physics and engineering, fractional differential equations have
been the focus of numerous studies. Therefore, considerable attention has been given
to how solving fractional ordinary differential equations, integral equations and fractional partial differential equations. However, there exists no method that yields
exact solutions to differential equations of fractional order, consequently some analytical techniques to handle such equations have been proposed, for example, Adomian
decomposition method [4], [12], the homotopy analysis method [3], the homotopy
perturbation method [1], the Taylor series method [10], variational iteration method
(VIM) [8], [15], etc. Recently, Ghorbani [7] introduced a new alternative technique for
solving nonlinear differential equations of fractional order which requires no Lagrange
multiplier or variational theory. This method can be very effective and reliable for
solving nonlinear fractional differential equations.
The paper is organized as follows: section 2 is devoted to describe some necessary
definitions on fractional calculus which will be used throughout the paper, in section
3 we describe briefly the basic concept of the fractional iteration method (FIM). In
section 4 we apply successfully the FIM to solve the fractional Bagley-Torvik equation.
Some concluding remarks are also given in section 5.
2. Preliminaries definitions
The idea of fractional calculus has been known since the development of the regular calculus, with the first reference probably being associated with Leibniz and
LHospital in 1695 where half-order derivative was mentioned. The commonly used
definitions for the general fractional are Riemann-Liouville and Caputo. In this paper, the derivatives are considered in the Caputo sense, which has the advantage of
defining integer order initial conditions for fractional order differential equations.
Received May 17, 2012.
251

252

T. MEKKAOUI AND Z. HAMMOUCH

Definition 2.1. A real function f (x), x > 0, is said to be in the space C , R if

there exits a real number > such that f (x) = x g(x), where g(x) C[0, ) and
it is said to be in the space Cm if and only if f (m) C for m N.
Definition 2.2. The Riemann-Liouville fractional integral operator of order of a
real function f (x) C , 1, is defined as
Z x
1
J f (x) =
(x t)1 f (t)dt, > 0, x > 0 and J 0 f (x) = f (x). (1)
() 0
The the operators J have the following proprieties, for , 0, , 1 :
J J f (x) = J + f (x),
J J f (x) = J J f (x),
(+1)
J x = (++1)
x+ .
Next we define the Caputo fractional derivatives D of a function f (x) of any real
m
number such that m 1 < m, m N, for x > 0 and f C1
in the terms of
J as
Z x
1
(x t)m1 f (m) (t)dt
(2)
D f (x) = J m Dm f (x) =
(m ) 0
and has the following proprieties for m 1 < m, m N, 1 and f Cm
D J f (x) = f (x),
m1
P (k) + xk
J D = f (x)
f (0 ) k! , for x > 0.
k=0

3. The Variational Iteration Method

In this section we recall briefly the basic concept of VIM [8]. We begin with
considering a differential equation in the general form,
L(y(t)) + N (y(t)) = F (t, y(t)),

(3)

where L is a linear operator, N a nonlinear operator and F is the source term.

According to the variational iteration method [8], one can construct a correction
functional as follow
yn+1 (t) = yn (t) + J (( )[L(yn ( )) + N (
yn ( )) F (, y( ))])

(4)

where is a general Lagrangian multiplier, the subscript n denotes the nth order
approximation, y0 (t) is an initial approximation which can be known according to
the initial conditions or the boundary conditions, and the function yn is a restricted
variation which means yn = 0. It is clear that the successive approximation yn , n
1, can be established by determining a general Lagrangian multiplier , which can
be identified optimally via the variational theory. The successive approximations
yn+1 , n 0 of the solution y(t) will be readily obtained upon using the obtained
Lagrange multiplier and by using any selective function y0 (t). When is known, then
other several approximations yn , n 1, follow immediately. Consequently, the exact
solution can be obtained by using
y(t) = lim yn (t).
n

(5)

APPROXIMATE SOLUTIONS TO THE BAGLEY-TORVIK EQUATION

253

4. The Fractional Iteration Method

Based on [7], we will explain briefly the basic concept of FIM, by considering a
nonlinear fractional differential equation in the following form:
Dt y(t) F (t, y(t)) = 0,

(6)

with the condition

y (k) (0) = ak ,

k = 0, 1, ..., m 1,

(7)

where m 1 m, m N, F is a given nonlinear function of y and y is

the unknown function to be determined. The main task is to fond a solution y to
problem (6), (7) under the form
y(t) = lim yn (t).
n

(8)

Let H(t) 6= 0 denote the so-called auxiliary function. Multiplying (6) by H(t) and
applying the Riemann-Liouville integral operator J on both sides we get
J (H(t)[Dt y(t) F (t, y(t))]) = 0.

(9)

Let h be the so-called auxiliary parameter. Multiplying (9) by h and adding y on

both sides we get
y(t) = y(t) + hJ (H(t)[Dt y(t) F (t, y(t))]).

(10)

Consequently, equation (9) can be solved iteratively as follows

yn+1 (t) = yn (t) + hJ (H(t)[Dt yn (t) F (t, yn (t))]).

(11)

Where the subscript n denotes the iteration order. The convergence of (11) is ensured by the Banachs fixed point theorem, provided that the right hand of (11) is a
contractive mapping [7]. On the other hand the convergence region and the rate of
the convergence of the obtained series solutions can be controlled by the convergence
parameter h as it is described in the framework of the homotopy analysis method
(HAM) [9]. Actually, by plotting the solution (or one of its derivatives) at a particular point with respect to the auxiliary parameter h which is the so-called hcurve,
one can get a proper value of h that ensures the convergence of the obtained solution
series. For more details about the mathematical properties of the hcurve we refer
the readers to a recent paper by Abbasdandy et al. [2].
5. Numerical implementation and discussion
In this section we will solve the Bagley Torvik equation by the so-called fractional
iteration method. The obtained solutions will be compared to the exact ones and to
those obtained via the variational iteration method.
The Bagley-Torvik equation is originally formulated in the studies on behavior of
real material by use of fractional calculus [5]-[13]. It has raised its importance since
than in many engineering and applied sciences applications. In particular, the equation with 21 -order derivative or 32 -order derivative can model the frequency-dependent
damping materials quite satisfactorily. It can also describe motion of real physical
systems, the modeling of the motion of a rigid plate immersed in a Newtonian fluid
and a gas in a fluid, respectively [14]. Approximate solutions have recently been
proposed in the book and papers of Podlubny in which the solution obtained with
approximate methods is compared to the exact solution.

254

T. MEKKAOUI AND Z. HAMMOUCH

Let us consider the following Bagley-Torvik equation with fractional order

3

y 0 (t) + Dt2 y(t) + y(t) = F (t),

(12)

with the initial conditions

y(0) = 1,

y 0 (0) = 1.

(13)

where
t3/2
F (t) = 7 t + 8 + t3 + 1.

(14)

y(t) = t3 + t + 1.

(15)

and the exact solution is

First, we begin by solving the Bagley-Torvik equation by VIM as follows.
The Lagrange multiplier can be identified as = 1, and from (4), we have the
following iteration formula for solving equation (12):


3
yn+1 (t) = yn (t) + J ( )[yn0 (t) + Dt2 yn (t) + yn (t) F (t)] ,
(16)
In view of (13), we choose the initial guess in the form of y0 (t) = 1 + t, we have then
the following recursive relation:
y0 (t) = 1 + t,
16 t5/2
64 t9/2
+ t3 +

5
315

1 6
y2 (t) = 1 + t + t3 3 t2 1/2 t4
t
120
1 6
t3/2
96 t7/2
128 t11/2
512 t15/2

y3 (t) = 1 + t + t3 6 t2 t4
t +8
60
35
1155
675675

.
y1 (t) = 1 + t +

.
.

(17)

and so on. On the other hand, using FIM (11) and starting with the same initial
approximation y0 (t) = 1 + t and the auxiliary function H(t) = 1, we obtain the
following
y0 (t) = 1 + t,
y1 (t) = 1 + t

64 ht9/2
16 ht5/2
ht3

5
315

y2 (t) = 1 + t

32 ht5/2
128 ht9/2
32 h2 t5/2
2 ht3
3 h2 t2

5
315
5

1/2 h2 t4 h2 t3
.

128 h2 t9/2
1 2 6

h t
315
120

.
.

(18)

and so on. In general, by means of the h-curve, it is straightforward to choose a

proper value of h which ensures that the solution series is convergent. This proper
value of h corresponds to the curve segment nearly parallel to the horizontal axis. For
this study we take h = 0.4 (see Figure 1).

APPROXIMATE SOLUTIONS TO THE BAGLEY-TORVIK EQUATION

255

Table 1. Comparison of the FIM solutions of (12) with the exact

and the VIM solution
t
yV IM
0.10 1.183140356
0.25 1.43878394
0.50 1.51984451
0.75 0.83083557
1.00 1.11359385

yexact
1.101000
1.265625
1.625000
2.171875
3.000000

yF IM
1.103763584
1.269040456
1.623997167
2.166900262
2.994988879

|yV IM yexact |
0.082140356
0.173158944
0.105155480
1.341039427
4.113593854

|yF IM yexact |
0.00276358
0.00341545
0.00100283
0.00497473
0.00501112

In order to verify numerically whether the proposed methodology leads to high

accuracy, we evaluate the numerical solutions using third-order approximation and
compare it with both of the exact analytical solution and the VIM solution. Table
1 shows the absolute errors between exact solution and the FIM and VIM solutions.
Table 1 and Figures 2(a) and (b) show that the FIM numerical approximate solution
has a high degree of accuracy, compared with the VIM solution. As we know, the
more terms added to the approximate solution, the more accurate it will be.

Figure 1. hcurve of y(t) for the second and third order approximation by FIM.

(a)
Figure 2. Comparison of the exact solution of (12) with: (a) the
FIM solution, (b) both of the FIM and VIM solutions.

(b)

256

T. MEKKAOUI AND Z. HAMMOUCH

6. Conclusion
The fractional iteration method has proven as an efficient technique to solve nonlinear fractional differential equations. Comparison with the variational method has
been shown, the simplicity of the method and the obtained exact results show that it
is a powerful mathematical tool for solving nonlinear fractional differential equations.
The method was used in a direct way without need for the Lagrange multiplier, correction functional, stationary conditions, linearization or discretization. It also provides
more realistic series solutions that converge very rapidly in real physical problems.
References
[1] S. Abbasbandy, Homotopy perturbation method for quadratic Riccati differential equation and
comparison with Adomian decomposition method, Appl. Math. Comput. 172 (2006), 485-490.
[2] S. Abbasbandy, E. Shivanian and K. Vajravelu, Mathematical properties of h-curve in the frame
work of the homotopy analysis method, Comm. in Nonl. Sci. and Num. Sim. 16 (2011), 4268
4275.
[3] F. Abidi and K. Omrani, The homotopy analysis method for solving the Fornberg-Whitham equation and comparison with Adomians decomposition method, Comput. Math. Appl. 59 (2010),
2743-2750.
[4] G. Adomian, Solving Frontier Problems of Physics: The Decomposition Method, Kluwer Academic Publishers, Boston, 1994.
[5] R. L. Bagley and P. J. Torvik, On the appearance of the fractional derivative in the behavior of
real materials, ASME J Appl Mech 5 (1984), no. 2, 294-298.
[6] R. L. Bagley and P. J. Torvik, Fractional calculus a different approach to the analysis of viscoelastically damped structures, AIAA J. 21 (1983), no. 5, 741-748.
[7] A. Ghorbani, Toward a new analytical method for solving nonlinear fractional differential equations, Comput. Methods Appl. Mech. Engrg. 197 (2008), 4173-4179.
[8] J.H. He, Variational iteration method a kind of non-linear analytical technique: some examples,
Internat. J. Non-Linear Mech. 34 (1999), 699-708.
[9] S.J. Liao, On the homotopy analysis method for nonlinear problems, Appl. Math. Comput. 147
(2004), pp.499-513.
[10] Yildiray Keskin, Onur Karaoglu, Sema Servi and Galip Oturanc, The approximate solution of
high-order linear fractional differential equations with variable coefficients in terms of generalized
Taylor polynomials, Mathematical and Computational Applications 16 (2011), no. 3, 617-629.
[11] I. Podlubny, Fractional Differential Equations, Mathematics in Science and Engineering 198,
Academic Press, San Diego, Calif, USA, 1999.
[12] R. Rao, The use of Adomian decomposition method for solving generalized Riccati equations,
Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications
(ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia, 2010.
[13] S. S. Ray and R. K. Bera, Analytical solution of the Bagley-Torvik equation by Adomian
decomposition method, Appl. Math. Comput. 168 (2005), no. 1, 398-410.
[14] M.A.Z. Raja, [Link] and [Link], Solution of fractional order system of Bagley-Torvik
equation using Evolutionary computational intelligence, Mathematical Problems in Engineering
(2011), 1-18.
[15] M. Matinfar, M. Saeidy, M. Mahdavi and M. Rezaei, Variational iteration method for exact solution of gas dynamic equation using Hs polynomials, Bulletin of Math. Analysis and Applications
3 (2011), 50-55.
Moulay Ismail, De
partement de
(Toufik Mekkaoui, Zakia Hammouch) Universite
matiques FST Errachidia BP 509, Boutalamine 52000, Morocco
Mathe
E-mail address: toufik mekkaoui@[Link], [Link]@[Link]