International Journal of the Physical Sciences Vol. 6(16), pp.

4102-4109, 18 August, 2011

Available online at http://www.academicjournals.org/IJPS
DOI: 10.5897/IJPS11.146
ISSN 1992 - 1950 2011 Academic Journals


Full Length Research Paper

A two-step Laplace decomposition method for solving
nonlinear partial differential equations

H. Jafari
1,2
*, C. M. Khalique
2
, M. Khan
3
and M. Ghasemi
1

1
Department of Mathematics, Faculty of Mathematical Science, Mazandaran University, P. O. Box 47415-954, Babolsar,
Iran.
2
International Institute for Symmetry Analysis and Mathematical Modelling, Department of Mathematical Sciences,
North-West University, Mafikeng Campus, Mmabatho 2735, South Africa.
3
Department of Sciences and Humanities, National University of Computer and Emerging Sciences Islamabad,
Pakistan.

Accepted 24 June, 2011

The Adomian decomposition method (ADM) is an analytical method to solve linear and nonlinear
equations and gives the solution a series form. Two-step Adomian decomposition method (TSADM) is a
modification on ADM and makes the calculations much simpler. In this paper we combine Laplace
transform and TSADM and present a new approach for solving partial differential equations.

Key words: Two-step Adomian decomposition method, Laplace decomposition method, Adomian
decomposition method.

Mathematics subject classification: 47J30, 49S05.


INTRODUCTION

Most of phenomena in nature are described by nonlinear
differential equations. So scientists in different branches
of science try to solve them. But because of nonlinear
part of these groups of equations, finding an exact
solution is not easy. Different analytical methods have
been applied to find a solution to them. For example,
Adomian (1986, 1988, 1989, 1990, 1991, 1994a, b) has
presented and developed a so-called decomposition
method for solving algebraic, differential, integro-
differential, differential-delay and partial differential
equations. Recently, a modification of ADM was
proposed by Wazwaz (1999). The modified
decomposition method needs only a slight variation from
the standard ADM and has been shown to be
computationally efficient. We consider the following
equation:

). ( ) (
1 1
u N L u R L f u

+ + =


The modified decomposition method was established



*Corresponding author. E-mail: jafari@umz.ac.ir
based on the assumption that the function f can be
divided into two parts and the success of the modified
method depends on the proper choice of the parts
1
f
and
2
f
. The TSADM (Lou, 2005) over comes this difficulty
and explains how we can choose
1
f
and
2
f
properly
without having noise term (Adomian and Race, 1992).
Using Laplace transform in ADM (LDM) proposed by
Khuri (2001, 2004) for the approximate solution of a class
of nonlinear ordinary differential equations, other
scientists have used this method for solving some
important equations (Yusufoglu, 2006; Elgazery, 2008;
Kiymaz 2009; Khan and Gondal, 2010; Hussain and
Khan 2010).
In this work we are interested to use Laplace transform
in TSADM (LTSDM). We illustrate this method with the
help of several examples and compare LDM and LTSDM
with each other.


DESCRIPTION OF LTSDM

Here,the purpose is to discuss the use of Laplace
transform algorithm in TSADM for solving different




equations. We consider general inhomogeneous
nonlinear equation with initial conditions given below:


), , ( t x h Nu Ru Lu = + +
(1)

where L is the highest order derivative which is assumed
to be easily invertible, R is a linear differential operator of
order less than L, Nu represents the non-linear terms and
h(x,t) is the source term. First we explain the main idea of
LDM:
The methodology consists of applying Laplace
transform on both sides of Equation (1)

[ ] [ ] [ ] [ ]. ) , ( ) , ( ) , ( ) , ( t x h t x Nu t x Ru t x Lu = + +
(2)

Using the differential property of Laplace transform and
initial conditions we get

[ ] ) , ( ... ) , ( ) , ( ) , ( 0 0 0
1 2 1
x u x u s x u s t x u s
n n n n



[ ] [ ] [ ]. ) , ( ) , ( ) , ( t x h t x Nu t x Ru = + +
(3)


[ ]
n
n
s
x u
s
x u
s
x u
t x u
) , (
...
) , ( ) , (
) , (
0 0 0
1
2

+ +

+ =



[ ] [ ] [ ]. ) , ( ) , ( ) , ( t x h
s
t x Nu
s
t x Ru
s
n n n
1 1 1
+
(4)

The next step is representing solutions as an infinite
series

=
=
0 i
i
t x u t x u ), , ( ) , (
(5)

and the nonlinear operator is decomposed as

=
=
0 i
i
A t x Nu , ) , (
(6)

where
n
A
is Adomian polynomial (Wazwaz, 2002a) of
n
u u u ,... ,
1 0
and can be calculated by the formula

,..., 2 , 1 , 0 , ) (
!
1
0
0
=
(

=
=

i u N
d
d
i
A
i
i
i
i
i
i

(7)

Substitution of Equations (5) and (6) in Equation (4)
yields


n
n
i
i
s
x u
s
x u
s
x u
t x u
) , (
...
) , ( ) , (
) , (
0 0 0
1
2
0

=
+ +

+ =
(

Jafari et al. 4103



[ ] [ ], ) , ( ) , ( t x h
s
A
s
t x Ru
s
n
i
i
n n
1 1 1
0
+
(

=
(8)

On comparing both sides of Equation (8) and by using
standard ADM we have:


[ ] [ ] ) , ( ) , (
) , (
...
) , ( ) , (
) , ( s x k t x h
s s
x u
s
x u
s
x u
t x u
n n
n
= + + +

+ =

1 0 0 0
1
2
0
(9)


[ ] [ ] [ ], ) , ( ) , (
0 0 1
1 1
A
s
t x Ru
s
t x u
n n
=
(10)


[ ] [ ] [ ]. ) , ( ) , (
1 1 2
1 1
A
s
t x Ru
s
t x u
n n
=
(11)

In general, the recursive relation is given by


[ ] [ ] [ ] . , ) , ( ) , ( 0
1 1
1
=
+
i A
s
t x Ru
s
t x u
i
n
i
n
i
(12)

Applying inverse Laplace transform to Equations (9 to
12), our required recursive relation is given by


), , ( ) , ( t x G t x u =
0
(13)


[ ] [ ] [ ] , ) , ( ) , ( 0
1 1
1
1

(

+ =

+
i A
s
t x Ru
s
L t x u
i
n
i
n
i


where G(x,t) represents the term arising from source term
and prescribed initial conditions.
Now we illustrate TSADM. By applying the inverse
operator
1
L
to h(x,t) and using the given conditions we
have:

)), , ( ( t x h L
1
+ =
(14)

where the function indicates the terms arising from
using the given conditions, all are assumed to be
prescribed. For using TSADM we set


, ...
m
+ + + + =
2 1 0
(15)

where
m
, ... , ,
1 0
are the terms arising from applying
inverse operator on the source term h(x,t) and using the
given conditions. We define


, ...
s k k
u
+
+ + =
0
(16)

where k = 0,1,...,m, s = 0,1,...,m-k. Then we verify that
0
u
satisfies the original Equation (1) and given conditions by
substituting. Once the exact solution is obtained we are
4104 Int. J. Phys. Sci.



done. Otherwise, we go to step two. In second step we
set
=
0
u
and continue with the standard ADM:

0
1 1
1
=

+
k A L u R L u
k k k
), ( ) (
(17)

By comparison with ADM and TSADM, it is clear that
TSADM may provide the solution by using one iteration
only and does not have the difficulties arising in the
modified method. Further, the number of terms in

,
namely m, is small in many practical problems. So,
applying TSADM will not be time consuming. Our
purpose in this paper is to combine the LDM and
TSADM. So, we should divide G(x,t) into its components
and check the required conditions for property choice of
) , ( t x u
0
. After applying inverse transform, by TSADM
criterion we can find the exact solution of our equation
after one iteration.


APPLICATIONS

To illustrate LTSDM we now consider some examples.

Example 1

Consider the nonlinear partial differential equation
(Wazwaz, 1999)


,
2
xt x uu u
x t
+ = +
(18)

with initial conditions


. ) , ( 0 0 = x u


Applying the Laplace transform we have:

[ ] [ ] [ ]
x
uu xt x x u t x u s ) , ( ) , ( + =
2
0
(19)

[ ]. ) , (
x
uu
s s
x
s
x
s x u
1 2
4 2
+ =


Applying inverse Laplace transform we get

[ ] . ) , (
(

+ =

x
uu
s s
xt
xt t x u
1 1
3
1
3
(20)

As we know the solution is in infinite series form:

=
=
0 n
n
t x u t x u ), , ( ) , (
(21)




and nonlinear term is handled with the help of Adomian
polynomials given as follows:

=
=
0 n
n x
u A uu ). (
(22)
By substituting Equations (21) and (22) in (20) we have:


. ) ( ) , ( ) , (

(
(

+ = =
0 0
1
3
1
3
n n
n n
u A
s
xt
xt t x u t x u
(23)
By using LDM we have:


,
3
) , (
3
0
xt
xt t x u + =
(24)

[ ] , ) (
1
) , (
0
1
1
(

=

u A
s
t x u (25)

[ ] . 1 , ) (
1
) , (
1
1

(

=

+
n u A
s
t x u
n n
(26)

By this recursive relation we can find other components
of the solution.


[ ] , ) , (
(

=

x
u u
s
t x u
0 0
1
1
1

15
2
63 3
5 7 3
xt xt xt
=


[ ] , ) ( ) , (
(

=

u A
s
t x u
1
1
2
1

[ ] ,
(

+ =

x x
u u u u
s
1 0 0 1
1
1

(
(

=

xt xt xt xt xt xt
s
4
15
4
189
0 2
63
2
9
2
3
2 1
2 8 6 4
1



2079
2
2835
38
315
22
15
2
11 9 7 5
xt xt xt xt
+ + + =

M

=
= =
0 n
n
xt t x u t x u . ) , ( ) , (
(27)

As we can see for finding next components large amount
of computation should be done. Now using the LTSDM
scheme gives:


,
1 0
+ =
(28)




3
1 0
3
1
xt and xt = =


It is obvious that
1

does not satisfy Equation (18). By

choosing
0 0
= u
and by verifying that
0
u
justifies
Equation (18), the exact solution will be obtained
immediately and we have:

, ) , (
0
xt t x u =
(29)
[ ] , ) ( ) , (
(

=

u A
s
xt
t x u
0
1
3
1
1
3
(30)

1
1
0
1
1

(
(

+
n u A
s
t x u
n
n n
, ) ( ) , (
(31)

So


[ ] , ) , ( 0
1
3
0 0
1
3
1
=
(

=

x
u u
s
xt
t x u


, 0 0 ) , (
1
=
+
n t x u
n


and the solution is


. ) , ( ) , (
0
xt t x u t x u
n
n
= =

=


It is important to note that if we select


,
3
1
) , (
3
0
xt t x u =


we obtain:


[ ]
(

=

x
u u
s
xt t x u
0 0
1
1
1
) , (


(

=
6 1
9
1 1
xt
s
xt


(

=

8
1
9
6
s
x
xt
!

,
63
7
xt
xt =


[ ]
(

=

1
1
2
1
A
s
t x u ) , (


[ ]
(

+ =

x x
u u u u
s
1 0 0 1
1
1


Jafari et al. 4105




(

=
10 4 1
189
2
3
2 1
xt xt
s



(

=

12 6
1
38400 16
s
x
s
x

,
2079
1 2
15
2
1 5
xt xt
+ =


M


As we can see, finding the solution is not possible after
one iteration and we have to calculate the components
like standard ADM.

Example 2

We now consider another nonlinear partial differential
equation (Wazwaz, 2002b)

,
) , (
2
2
u x u u
x
y x u
yy x
+ =

(32)

with initial conditions


, sin ) , ( y y u = 0


. ) , ( 1 0 = y u
x


Applying the Laplace transform we get


[ ] [ ] [ ], ) , ( ) , ( ) , (
yy x x
u u u x y u y su y x u s + + = 0 0
2
(33)

[ ].
sin
) , (
yy x
u u u
s s s
y
s
y s u + + + =
2 4 2
1 1 1


By applying inverse transform we get


[ ] .
!
sin ) , (
(

+ + + =

yy x
u u u
s
x
y x y s u
2
1
3
1
3
(34)

Likewise as in the previous example, we have

(
(

+ + + =
0 0 0
2
1
3
1
3
n n n
u n n n
A y x u
s
x
y x y x u . ) , (
!
sin ) , (
) (
(35)

By using LDM:

,
!
sin ) , (
3
3
0
x
y x y x u + =
(36)

[ ] , 0 ,
1
) , (
2
1
1

(

+ =

+
n A u
s
y x u
n n n
(37)
4106 Int. J. Phys. Sci.



by using the previous recursive relation

[ ]
(

+ =

0 0
2
1
1
1
A u
s
y x u ) , (

[ ]
(

+ =

yy x
u u u
s
0 0 0
2
1
1


(
(

+ =

) ( sin
!
sin
2
1
3
1
2 3
2
1
x
y
x
y x
S

(
(

+ + =

) ( sin
sin

3 5 6 3 4
1
1
3
1 1 1
S
y
S s
y
S
s

4 5 3
72 5
1
3
1
x
y
x x
sin
! !
+ =

M

and

=
+ = =
0 n
n
y x y x u y x u sin ) , ( ) , (
(38)
It is clear that finding the exact solution is time consuming
but by using LTSDM:


,
2 1 0
+ + =
(39)

.
!
sin ,
3
2 1 0
3
1
x and y x = = =


By putting
1 0
,
and
2

we can see none of them

satisfies Equation (29) but
1 0
+
justifies the Equation
(29). So, we select
y x u sin + = + =
1 0 0

and in this
case we have:

, sin ) , (
0
y x y x u + =
(40)

[ ] ,
1

! 3
) , (
0 0 2
1
3
1
(

+ + =

A u
s
x
y x u (41)

[ ] , 1 ,
1
) , (
2
1
1

(

+ =

+
n A u
s
y x u
n n n
(42)

where
) (u A
n
is Adomian polynomials and representing
the nonlinear term of Equation (29).


, sin ) , (
0
y x y x u + =
(43)

[ ] ,
!
) , (
(

+ + =

yy x
u u u
s
x
y x u
0 0 0
2
1
3
1
1
3





[ ] , sin sin
!
(

+ + =

y y x
s
x
2
1
3
1
3

! ! 3 3
3 3
x x
+ =


, 0 =
(44)

0 0
1
=
+
n y x u
n
, ) , (

(45)

and the solution will be

=
+ = =
0 n
n
y x u y x u . sin ) , (
(46)

If we select
x u =
0
, then

[ ]
(

+ + =

0 0
2
1
3
1
1
3
A u
s
x
y y x u
!
sin ) , (

[ ] ,
!
sin
(

+ + =

yy x
u u u
s
x
y
0 0 0
2
1
3
1
3

,
!
sin
(

+ =

4
1
3
1
3 s
x
y

! !
sin
3 3
3 3
x x
y + =



, sin y =
(47)


[ ]
(

+ =

1 1
2
1
2
1
A u
s
y x u ) , (

0 =
(48)


. , ) , ( 1 0
1
=
+
n y x u
n
(49)

By this choice for
0
u
we find the solution after two
iterations. So, LTSDM works better and has less amount
of computation. As we can see
2

does not satisfy

Equation (29). By choosing
! 3
3
2 0
x
u = =
we have:


,
!
) , (
3
3
0
x
y x u =
(50)


[ ]
(

+ + + =

0 0
2
1
1
1
A u
S
y x y x u sin ) , (





[ ] , sin
(

+ + + =

yy x
u u u
s
y x
0 0 0
2
1
1

,
!
sin
(
(

+ + =

3
1
3
2
1
x
s
y x


,
!
sin
5
5
x
y x + =
(51)


[ ]
(

+ =

1 1
2
1
2
1
A u
s
y x u ) , (

[ ] ,
(

+ + =

yy x yy x
u u u u u
s
0 1 1 0 1
2
1
1

, sin
!
sin
(
(

+ + =

y x
x
y x
s
2
5
2
1
2
1
5
1


,
!
sin
!
sin
! 4 7 2 3
4 7 2 3
x
y
x x
y
x
+ + =



M
(52)

As we can see it is necessary to calculate more
components of
) , ( t x u
to get the exact solution.


Example 3

Consider the system of inhomogeneous partial differential
equations (Wazwaz, 2003)

=
+ + =
), ( 1
), ( 1
t x Exp uv v
t x Exp uv u
x x
x
(53)

with initial conditions

), ( ) , ( t Exp t u = 0

). ( ) , ( t Exp t v = 0


Applying the Laplace transform on both sides of Equation
(53) we get

[ ] [ ] ) )( ( ) . ( ) , (
1
1 1
0

+ =
s
t Exp
s
uv t u t x u s
(54)

[ ], ) (
) (
) (
) , ( uv
s
t Exp
s s s
t Exp
s
t s u
1 1
1
1
2
+ +

+ =


and by applying inverse transform
Jafari et al. 4107



[ ] . ) ( ) , (
(

+ + + =

uv
s
t x Exp x t x u
1
1
(55)

For second equation

[ ] [ ] ) )( ( ) , ( ) , (
1
1 1
0
+
=
S
t Exp
s
uv t v t x v s
x
(56)


[ ]. ) (
) (
) (
) , (
x
uv
s
t Exp
s s s
t Exp
s
t s v
1 1
1
1
2
+ +
+

=


By applying inverse transform we get

[ ] . ) ( ) , (
(

+ + =

x
uv
s
t x Exp x t s v
1
1
(57)

By using LDM:


) ( ) , (
0
t x Exp x t x u + + =
(58)

) ( ) , ( t x Exp x t x v + =
0
(59)

, , ) ( ) , ( 0
1
0
1
1

(
(

+
n u A
s
t x u
n
n n
(60)

, , ) ( ) , ( 0
1
0
1
1

(
(

+
n u A
s
t x v
n
n n
(61)

by using the previous recursive relation

[ ]
(

=

0
1
1
1
A
s
t x u ) , (

[ ]
(

=

0 0
1
1
v u
s


[ ]
(

+ + + =

)) ( ))( ( ( t x Exp x t x Exp x
s
1
1

(

+ =

2 2 4 2
1
1 1
2 1
) (
) (
) (
) (

s s
t Exp
s s
t Exp
s s

), )( ( ) ( ) )( ( ) ( 1 1
3
1
3
+ + + + + = x t x Exp t Exp x t x Exp t Exp x x

M

[ ]
(

=

0
1
1
1
A
s
t x v ) , (


[ ]
(

=

x
v u
s
0 0
1
1


4108 Int. J. Phys. Sci.



[ ]
(

+ + =

)) ( ))( ( ( t x Exp t x Exp x
s
1
1
1

(

+ =

2 2
1
1 1
1 1
) (
) (
) (
) (

s s
t Exp
s
t Exp
s s


), ( ) ( x t xExp t x Exp x + + + = 1


M

and

=
+ = =
0 n
n
t x Exp t x u t x u ), ( ) , ( ) , (
(62)

=
= =
0 n
n
t x Exp t x u t x v ). ( ) , ( ) , (
(63)

For finding the solution we should calculate other
components of
) , ( ), , ( t x v t x u
and obviously it is time
consuming. Now we can see by using LTSDM, getting
the solution is very simple. For first equation we have

,
1 0
+ =
(64)


), ( t x Exp and x + = =
1



and for second equation


,
1 0
+ =
(65)


). ( t x Exp and x = =
1 0



By putting these functions in Equation (53), we can see
1

and

1

satisfy Equation (53).


), ( ) , ( t x Exp t x u + =
0
(66)

), ( ) , ( t x Exp t x v =
0
(67)


[ ]
(

+ =

0
1
1
1
A
s
x t x u ) , (

[ ]
(

+ =

0 0
1
1
v u
s
x


[ ]
(

+ + =

)) ( ))( ( ( t x Exp t x Exp
s
x
1
1

(

+ =

2
1
1
s
x


0 = = x x
(68)




, , ) ( ) , ( 0 0
1
0
1
1
=
(
(

+
n u A
s
t x u
n
n n
(69)


[ ]
(

=

0
1
1
1
A
s
x t x v ) , (


[ ]
(

=

x
v u
s
x
0 0
1
1



[ ]
(

+ =

)) ( ))( ( ( t x Exp t x Exp
s
x
1
1


(

=

2
1
1
s
x



0 = = x x
(70)


, . ) ( ) , ( 0 0
1
0
1
1
=
(
(

+
n u A
s
t x u
n
n n
(71)

So

=
+ = =
0 n
n
t x Exp t x u t x u ), ( ) , ( ) , (
(72)

=
= =
0
), ( ) , ( ) , (
n
n
t x Exp t x v t x v
(73)

Same as previous examples, by choosing
x u = =
0 0


and
x v = =
0 0

that do not justify Equation (53) we will
have the same size of computation with standard ADM.


Conclusions

In this work, we have carefully combined two-step
Adomian decomposition method with Laplace transform.
In the three illustrated examples we showed that LTSDM
consists of three steps: the first step was applying
Laplace transform on our equation and then inverse
transform, the second step was verifying that the zeroth
component of the series solution included the exact
solution. If yes, we are done. Otherwise, we should go to
the third step and continue with the standard ADM. The
obtained results in examples indicated that LTSDM was
feasible, effective and we do not have the noise term.
The LTSDM overcomes the difficulties arising in the
modified decomposition method established in (Wazwaz,
1999).The power of LTSDM depends on the proper
choice of
0
u
and
1
u
and the occurrence of the exact
solution in the zeroth term. If the exact solution exists in
the zeroth component LTSDM requires much less




calculation in comparison with LDM. If not, LTSDM
requires only a little effort than LDM.


ACKNOWLEDGMENT

The First author (H. Jafari) gratefully acknowledge the
support from the research project of the National Elite
Foundation of Iran.


REFERENCES

Adomian G (1986). Nonlinear Stochastic Operator Equations: Acad.
press, Can Diego, CA.
Adomian G (1988). A review of the decomposition method in applied
mathematics: J. Math. Anal, Appl., 135 : 501-544
Adomian G (1989). Nonlinear Stochastic Systems Theory and
Applications to Physics: Kluwer Academic.
Adomian G (1990). A review of the decomposition method and some
recent results for nonlinear equation: Math. Comp. Model. 13(7): 17-
43.
Adomian G, Rach R (1990). Equality of partial solutions in the
decomposition method for linear or nonlinear partial differential
equations: Comp. Math. Appl., 19(2):9-12.
Adomian G (1991). A review of the decomposition method and some
recent results for nonlinear equations: Comp. Math. Appl., 21(5): 101-
127.
Adomian G, Race R (1992). Noise term in decomposition solution
series: Comput. Math. Appl., 24: 61-64.
Adomian G (1994a). Solving Frontier Problems of Physics: The
Decomposition Method: Kluwer Academic publisher, Boston.
Adomian G (1994b). Solution of physical problems by decomposition:
Comp. Math. Appl., 27(9/10): 145-154.
Elgazery NS (2008). Numerical solution for the Falkner-Skan equation:
Chaos Soliton and Fractals, 35: 738-746.
Hussain M, Khan M (2010). Modied Laplace decomposition method:
Appl. Math. Sci., 36(3): 1769-1783.































Jafari et al. 4109



Khuri SA (2001). A Laplace decomposition algorithm applied to class of
non-linear differential equation: J. Appl. Math., 4: 141-145.
Khuri SA (2004). A new approach to Bratus problem: Appl. Math.
Comput., 147: 131-136.
Kiymaz O (2009). An algorithm for solving Initial Value Problems using
Laplace decomposition method: Appl. Math. Sci., 3: 1453-1459.
Khan M, Gondal MA (2010). A new analytical solution of foam drainage
equation by Laplace decomposition method: J. Adv. Res. Differential
Equation., 2(3): 53-64.
Lou XG (2005). A two-step Adomian decomposition method: Appl.
Math. Comput., 170: 570-583.
Wazwaz AM (1999). A A reliable modification of Adomian
decomposition method: Appl. Math. Comput., 102:77-86.
Wazwaz AM (2002a). A new technique for calculating Adomian
polynomials for nonlinear polynomials: Appl. Math. Comput., 111: 33-
51.
Wazwaz AM (2002b). Partial differential equations methods and applied
Netherland Balkema publisher.
Wazwaz AM (2003). The existence of noise terms for system of
inhomogeneous dierential and integral equations: Appl. Math.
Comput., 146: 81-92.
Yusufoglu E (2006). Numerical solution of Dung equation by the
Laplace decomp. algorithm: Appl. Math. Comput., 177: 572-580.