Comparing Linear and Nonlinear Differential Equations

of Differential Transformation Method by Other

Numerical Methods
Che Haziqah Che Hussina , Adem Kilicmanb and Arif Mandanganc
a

Preparatory Centre of Science and Technology

Universiti Malaysia Sabah, Jalan UMS, 88400 Kota Kinabalu Sabah, Malaysia.
b

Department of Mathematics and Institute for Mathematical Research

Universiti Putra Malaysia, 43400 UPM, Serdang, Selangor, Malaysia
c

School of Science and Technology

Universiti Malaysia Sabah, Jalan UMS, 88400 Kota Kinabalu Sabah, Malaysia.
a

haziqah@ums.edu.my, bakilicman@putra.upm.edu.my, carifman@ums.edu.my

Abstract. In this study, we solve fifth-order boundary value problems by using the DTM for linear and
nonlinear differential equations and compare the results with other methods such as Adomian Decomposition
Method (ADM), Noor Decomposition Method and Variational Iteration Method. We provide several numerical
examples in order to show the accuracy of the method. Further, we also solve sixth-order nonlinear boundary
value problems and compare the result to ADM. The present study shows that the DTM is able to provide good
results with high accuracy and the method is also easy to apply.
Keywords: Ordinary Differential Equations, Numerical approximation and analysis.
PACS: 02.30.Hq, 02.60.Lj

INTRODUCTION
In 1986, Zhou [1] proposed an approximation concept of differential transformation method (DTM) and
applied to solve electric circuit analysis for linear and nonlinear initial value problems (IVPs). In the literature,
DTM was applied successfully to several different problems in linear and nonlinear boundary value problems
and initial value problems, for example, see [2, 3, 4] and [5].
This transformation technique is based on Taylor series expansion and is a useful tool to obtain analytical
solutions of differential equations. This method constructs the analytical solution in the form of polynomial.
The DTM is an approximation to exact solutions which are differentiable and it has high accuracy. It has minor
error when we compare to exact solutions. Basically, the DTM is different with traditional high order Taylor
series since it needs more computation and it also requires the necessary derivatives [6]. Further, the DTM can
also be applied in the high order differential equations and it is an alternative way to get Taylor series solution
for the given differential equations, for example, see [7, 8, 9, 10, 11].
In addition, Ayaz [12] studied the application of two-dimensional DTM in partial differential equations and
Borhanifar and Abazari [13] also studied partial differential equations for two-dimensional and threedimensional DTM. Ayaz [14] extended the method to solve a system of partial differential equations by DTM.
Wazwaz [15] solved fifth-order boundary value problems by ADM. The author stated that the method provides
solution in the form of a rapidly convergent series. However, Zhang [16] also solved fifth-order boundary value
problems by Variational Iteration Method. The author emphasized that the method presented is highly accurate
solutions. Noor and Mohyuddin [17] solved fifth-order boundary value problems by using Noor decomposition
method. They claimed that method had convergent series solution. Then, Wazwaz [18] applied modified ADM
to solve sixth-order boundary value problems. The author declared that the result demonstrated the reliability
and efficiency of the method.
In this paper, we solved fifth-order and sixth-order boundary value problems by DTM. We also prepared an
analysis on errors for DTM and compare with other numerical methods.

DIFFERENTIAL TRANSFORMATION METHOD

Suppose that, the function
is a continuously differentiable function on the open interval
Then we have the following definitions.
Definition 1
The differential transform of the function

for the -th derivative is defined as follows:

[

where

is the original function and

Definition 2
The inverse differential transform of

is the transformed function.

is defined as follows:

Substitution of equation (1) into equation (2) yields:

Note that, this is the Taylor series of

at
. The following basic operations of differential
transformation can be deduced from equations (1) and (2):
Theorem 1 If
Theorem 2 If
Theorem 3 If

then,
then,

then,

Theorem 4 If

then,

Theorem 5 If

then,

Theorem 6 If

then,

Theorem 7 If

then

Theorem 8 If

then,

Theorem 9 If

then,

Theorem 10 If

where,
.
.
(

then,

Theorem 11 If
then,
(
)
The following theorem was proved in [7].
Theorem 12
The general differential transformation for BVP of linear differential equation,
is given by
[
for integer

)]]

The next two theorems were given in [8,9] respectively.

Theorem 13
Let

is transformable then solution to the high order differential equation

[(

)(
(

is given by
]

[(

)(
(

Theorem 14
The general differential transformation for nonlinear nth-order BVPs,
[

is given by

) ([

)]

More details on the DTM are available in [3,7, 8] and [9].

NUMERICAL EXAMPLES
By applying the theorems in the paper [3,7,8] and [9], we provide several examples for fifth-order and sixthorder boundary value problems.
Example 1.Consider to solve the following linear fifth-order boundary value problems DTM, which is also
solved by ADM [15], Noor decomposition [16] and Variational Iteration Method [17].

subject to the boundary conditions

(5)
Solution:
By Theorem 1 and Theorem 8, the equation (4) is transformed into the equation (6),
[

Then by applying definition 1 at

])]

, we obtain the following transformed boundary conditions :

.

By equation (1),
and
Now by using the transformed equation (6)
together with the transformed boundary conditions, we can easily solve
. The values of and can
be evaluated at
from the following system:
[

][ ]

The solution of this algebraic system gives

. Finally, the following
series solution can be formed by applying the inverse transformation in definition 2 up to

Next we consider, which is also solved by ADM [15], Noor decomposition [16] and Variational Iteration
Method [17].
Example 2 Solve the following nonlinear fifth-order BVP of DTM
subject to the boundary conditions
(8)

By using Theorem 1, Theorem 6 and Theorem 8 the equation (7) is transformed to the following equation (9):

Applying equation (8) to the equation (1) at

found

then the following transformed boundary conditions can be

By using the transformed equation (9) together with the transformed initial conditions, we can easily solve
. The values of and can be evaluated at
from the following system:

The solution of this system gives

. Finally, the following solution
series can be formed by applying the inverse transformation in definition 2 up to

Then we consider
Example 3 Solve the following nonlinear sixth-order BVP of DTM, which is also solved by ADM [18]
subject to the boundary conditions
(11)
By using Theorem 1, Theorem 6 and Theorem 8, the equation (10) is transformed to the following equation
(12):

Applying equation (11) to the equation (1) at

found:

then the following transformed boundary conditions can be

By using the transformed equation (12) together with the transformed initial conditions, we can easily solve
. The values of , and can be evaluated at
from the following system:

The solution of this system gives

. Finally, the
following solution series can be formed by applying the inverse transformation in definition 2 up to

RESULT AND DISCUSSION

There are numerical results for DTM, ADM, NDM and VIM and comparisons of exact solution for Example 1,
2 and 3 which are presented in Table 1, Table 2 and Table 3 respectively.
TABLE (1).Comparison of numerical results for Example 1.

x
0.0
0.1

Exact
solution,y(x)
0
0.09946538264

0.2

0.1954244414

0.3

0.2834703496

0.4

0.3580379275

0.5

0.4121803177

0.6

0.4373085122

0.7

0.4228880689

0.8

0.3560865499

0.9

0.2213642863

1.0

2.50897061e-8

Error of DTM
0

Error of ADM
[15]
0

Error of NDM
[16]

Error of VIM
[17]
0

0
0
Error = analytical solution numerical solution
TABLE (2).Comparison of numerical results for Example 2.

Error of DTM

0.0
0.1

Exact
solution,y(x)
1
1.105170918

0.2

1.221402759

0.3

1.349858807

0.4

1.491824698

0.5

1.648721270

0.6

1.822118801

0.7

2.013752707

0.8

2.225540929

0.9

2.459603112

1.0

2.718281830

0
0

Error of ADM
[15]
0

Error of NDM
[16]
0

Error of VIM
[17]
0
0

From the tables, we can see that there are minor errors for Example 1, 2 and 3. For Example 1, the sum of error
of DTM is
. It has no error at points
and
but the error increases from points
till
The sum of error of ADM is
The errors increase from points
until
. The sum of error of NDM is
The errors increase from points
till
but there are
no error at points
and
. The sum of error of VIM is
. Then the sum of error for
Example 2, the sum of error of DTM is
. It has no error from points
to
but the error
increases for points
and no error at point 0.7. Then the errors increase for points 0.8 till 0.9 and decreases at
point
The sum of error of ADM is
Then the sum of error of NDM is


. The sum of error of VIM is
. Finally, for Example 3 we can see that
the sum of error of DTM is very small
. However, the sum of error of ADM is

0.002582383. From the results and graphs, the sum of error of DTM is the smallest error. The DTM
presented the accuracy and efficiency to solve these kinds of problems.
TABLE (3).Comparison of numerical results for Example 3.

x
0.0
0.1

Exact
solution,y(x)
1
1.105170918

0.2

1.221402759

0.000235410

0.3

1.349858807

0.000325718

0.4

1.491824698

0.000385533

0.5

1.648721270

0.000408610

0.6

1.822118802

0.000391896

0.7

2.013752707

0.000336092

0.8

2.225540930

0.000245880

0.9

2.459603114

0.000129888

1.0

2.718281831

FIGURE 1(a). Analysis on error for Example 1.

FIGURE 2(a). Analysis on error for Example 2

Error of DTM
0

Error of ADM
[18]
0
0.000123355

FIGURE 1(b). Analysis on error for Example 1.

FIGURE 2(a). Analysis on error for Example 2

FIGURE 3. Analysis on error for Example 3.

CONCLUSION
The results in this paper support the idea that DTM can be solved easily and it has accuracy in numerical
solution. The errors between the solution of DTM and exact solution are too small which indicates that DTM
has high accuracy for these problems and it also has high efficiency since we can solve the problems easily. The
DTM is the most accurate method, followed by ADM, NDM and VIM. Therefore, we can state that the DTM
method can very successful and is powerful in numerical solution for boundary value problems for these kinds
of problems. The computations in the examples were computed by Maple 13.

ACKNOWLEDGMENTS
The first author would like to thank Universiti Malaysia Sabah for supporting our participation during the 3rd
International Conference on Mathematical Sciences.

REFERENCES
1. J. K. Zhou, Differential Transformation And Its Applications For Electrical Circuits, Huarjung University Press, Wuhan,
China, 1986.
2 I. H. Abdel Halim Hassan and V. S. Erturk, European Journal of Pure and Applied Mathematics 2(3), 426-447 (2009).
3. V. S. Erturk and S. Momani, Applied Mathematics and Computation 188, 1963-1968 (2007).
4. S. Islam, S. Haq, and J. Ali, Commun Nonlinear Sci. Numer. Simulat. 14,1132-1138 (2009).
5. J. M. Jang and C. L. Chen, Applied Mathematics and Computation 88, 137-151 (1997).
6. Z. M. Odibat, C. Bertelleb, M. A. Aziz-Alaouic, and G. H. E. Duchamp, Computer and Mathematics with Applications,
2009, doi:10.1016/j.camwa.2009.11.005.
7. C. H. C. Hussin, A. Kilicman and A. Mandangan, Borneo Science Journal 27, 35-46 (2010).
8. C. H. C. Hussin and A. Kilicman, Mathematical Problems in Engineering, 2011, doi:10.1155/2011/724927.
9. C. H. C. Hussin and A. Kilicman, European Journal of Pure and Applied Mathematics 4(2), 174-185 (2011).
10. C. H. C. Hussin and A. Kilicman, Proceedings of the 20th National Symposium on Mathematical Sciences in Solving
System of Linear Differential Equations by Using Differential Transformation Method, AIP Conference Proceedings
1522, American Institute of Physics, 2013, pp. 245-250; doi: 10.1063/1.4801130.
11. A.Kilicman and O.Altun, Journal of The Franklin Institute 26(5), 917-928 (2012).
12. F. Ayaz, Applied Mathematics and Computation 143(2)(3), 361-374 (2003).
13. A. Borhanifar and R. Abazari, Applied Mathematics and Computation, 2009, doi:10.1007/s12190-009-0338.
14. F. Ayaz, Applied Mathematics and Computation 147, 547-567 (2004).
15. A.M. Wazwaz, Journal of Computational and Applied Mathematics 136, 259-270 (2001).
16. J. Zhang, Computers and Mathematics with Applications 58, 2347-2350 (2009).
17. M. A. Noor, S.T. Mohyud-Din, International Journal of Nonlinear Science 7(2), 143-148 (2009).
18. A.M. Wazwaz, Applied Mathematics and Computation 118, 311-325 (2001).