Professional Documents
Culture Documents
Latifizadeh
January 2014
KeyWords: Mitting-Leffler functions; biological population model; Caputo derivative; Sumudu transform;
Fokker-Plank equation.
well-established mathematical methods, namely, the Definition 2 The left sided RiemannLiouville frac-
homotopy perturbation method and the separation of tional integral of order 0, of a function f
variables method [21, 22, 23]. C , 1 is defined as:
There are numerous integral transforms such as t
1
the Laplace, Sumudu, Fourier, Mellin, and Hankel Jt f (t) = (t )1 f () d, (1)
to solve PDEs. Of these, the Laplace transforma- () 0
tion and Sumudu transformation are the most widely
where > 0, t > 0 and () is the gamma function.
used. The Sumudu transformation method is one of
the most important transform methods introduced in Definition 3 Let f Cn , n N {0}. The left
the early 1990s by Gamage K. Watugala. It is a pow- sided Caputo fractional derivative of f in the Caputo
erful tool for solving many kinds of PDEs in vari- sense is defined by [31] as follows:
ous fields of science and engineering [30, 34]. And
also various methods are combined with the Sumudu D f (t) =
t 1 t n1 (n)
transformation method such as the Homotopy Analy- (n) 0 (t ) f () d ,
sis Sumudu Transform Method (HASTM) [35] which (2)
n 1 < n,
is a combination of the homotopy analysis method t
Dn f (t) = n,
and the Sumudu transformation method. Another ex-
ample is the Sumudu Decomposition Method (SDM) Note that according to [1], Eqs.(1) and (2) be-
[25], which is a combination of the Sumudu transform come:
method and the adomian decomposition method. t
In this paper, a new approach is proposed to use 1
Jt f (x, t) = (t )1 f (x, )d (3)
the homotopy perturbation Sumudu transform method () 0
(HPSTM) to derive the exact solution of various
types of fractional partial differential equations. This for > 0, t > 0, and
method is a combination of the homotopy perturbation Dt f (x, t) =
method and the Sumudu transform method. However, t (4)
(n) 0 (t )
1 n1 f (n) (x, )d
Singh [36] used the homotopy perturbation Sumudu
transform method to obtain the exact solution of non- for n 1 < n.
linear equations which are PDEs of integer order.
The paper is structured in six sections. In section Definition 4 The single parameter and the two pa-
2, we begin with an introduction to some necessary rameters variants of the Mitting-leffler function are
definitions of fractional calculus theory. In section 3 denoted by E (t) and E, (t), respectively, which are
we describe the basic idea of the homotopy perturba- relevant for their connection with fractional calculus,
tion method. In section 4 we describe the homotopy and are defined as:
perturbation Sumudu transform method. In section 5,
we present four examples to show the efficiency of us- tj
E (t) = , > 0, t C, (5)
ing the HPSTM to solve FPDEs and also to compare (j + 1)
j=0
our result with those obtained by other existing meth-
ods. Finally, relevant conclusions are drawn in section
tj
6. E, (t) = , , > 0, t C (6)
(j + )
j=0
Definition 5 Sumudu transform over the following 3 The Basic Idea of the Homotopy
set of functions,
Perturbation Method
{ }
|t|
M, 1 , 2 > 0, |f (t)| < M e j In this section, we will briefly present the algorithm
A = f (t) ,
if t (1)j [0, ) of this method. At first, the following nonlinear dif-
(7) ferential equation is considered:
is defined by
A (u) f (x) = 0, x , (12)
S[f (t)] = G(u) = f (ut)et dt, (8) with the boundary conditions
0
Some special properties of the Sumudu transform are where A, B, f (x) and are a general differential
as follows: function operator, a boundary operator, a known an-
alytical function and the boundary of the domain ,
1. S [1] = 1; respectively.
[ n ] The operator A can be decomposed into a linear
2. S (n+1)
t
= un , n > 0; operator, denoted by L, and a nonlinear operator, de-
[ ] noted by N. Therefore, Eq.(12) can be written as fol-
3. S eat = 1
1au ; lows
L (u) + N (u) f (x) = 0, (14)
4. S [f (x) + g (x)] =S [f (x)] + S [g (x)]
Now we construct a homotopy v(x, p) : [0, 1]
Other properties of the Sumudu transform can be R which satisfies:
found in [3].
transform method in the last section, we consider the By applying the homotopy perturbation method,
following four examples. In example 8, a case of the and substituting Eq.(27) in Eq.(37) we have
one variable time-fractional Fokker-Plank equation is
solved. In example 9, we have generalized biological
population model. In example 10, a two-dimensional pn wn (x, t) = x2 +
fractional wave equation is solved and in example 11, n=0
[ [ (
x)
a systems of linear FPDEs is solved.
1 4 n
pS u S p wn (x, t)
x
x 3
n=0
Example 8 Consider the linear time-fractional ( )
2
2
Fokker-Plank equation with the initial conditions of pn wn (x, t) + 2 pn wn (x, t) .
the form: x
n=0 n=0
( ) (38)
4 x
Dt w(x, t) = w(x, t) w(x, t)
x x 3 Equating the terms with identical powers of p, we get
2
+ 2 w2 (x, t), (33) p0 : w0 (x, t) = x2 , [ [ (
x)
x
1 4
p1 : w1 (x, t) = S u S
w0 (x, t)
where t > 0, x R, 0 < 1, subject to the x]] x 3
2
initial condition w0 (x, t) + 2 w0 2 (x, t)
x
t
w (x, 0) = x2 (34) = x2 ,
( + 1) [ [ (
1 4 x)
p2 : w2 (x, t) = S u S
w2 (x, t)
Taking the Sumudu transform on both sides of x]] x 3
Eq.(33), thus we get 2
w2 (x, t) + 2 w2 2 (x, t)
[ ( ) x
4 x t2
S[Dt w(x, t)] = S w (x, t) w (x, t) = x2
x x ] 3 (2 + 1)
2 2 [ [ (
+ w (x, t) . (35) 1 4 x)
x2 pn : wn (x, t) = S u S
wn (x, t)
x]] x 3
2
Using the property of the Sumudu transform and wn (x, t) + 2 wn 2 (x, t)
x
the initial condition in Eq.(34), we have tn
2
=x (39)
(n + 1)
S [w (x, t)] = x2 +
[ ( ) Thus the solution of Eq.(33) is given by
4 x
u S
w(x, t) w(x, t)
x x 3
]
2 2 w(x, t) = lim pn wn (x, t)
+ w (x, t) , (36) p1
x2 n=0
t t2
= x + x2
2
+ x2 +
( + 1) (2 + 1)
Operating with the Sumudu inverse on both sides
= x2 E (t ). (40)
of Eq.(36) we get
If we put 1 in Eq.(40) or solve Eq. (33) and
w (x, t) = x2 + (34) with = 1, we obtain the exact solution
[ [ ( )
1 4 x
S u S
w(x, t) w(x, t) w (x, t) = x2 et
x x 3
]]
2 2 which is in full agreement with the result in Ref. [21,
+ w (x, t) . (37)
x2 40].
Operating with the Sumudu inverse on both sides of where 0 < , < 1, subject to the initial condition
Eq.(50) we get
w (x, 0) = sinh (x) , v (x, 0) = cosh (x) , (56)
w (x, y, t) = sin (x) sin (y) +
[ [ ( ) ]] Taking the Sumudu transform on both sides of
S1 u S 2 Dx2 + Dy2 w . (51) Eq.(55), thus we get
{
By applying the homotopy perturbation method, and S [Dt w] = S [vx v w] ,
substituting Eq.(27) in Eq.(51) we have (57)
S [Dt v] = S [wx v w] ,
Using the property of the Sumudu transform and the
pn wn (x, y, t) = sin (x) sin (y) + initial condition in Eq.(56), we have
n=0
[ [ ( )]] {
( ) u S [w (x, t)] u w (x, 0) = S [vx v w] ,
pS1 u S 2 Dx2 + Dy 2
pn wn (x, y, t) u S [v (x, t)] u v (x, 0) = S [wx v w] ,
n=0
(52) and
{
Equating the terms with identical powers of p, we get u S [w (x, t)] = u sinh (x) + S [vx v w] ,
u S [v (x, t)] = u cosh (x) + S [wx v w] ,
p0 : w0 (x, y, t) = sin(x) sin(y), (58)
1 4t Operating with the Sumudu inverse on both sides of
p : w1 (x, y, t) = sin(x) sin(y),
( + 1) Eq.(58) we get
42 t2 {
p2 : w2 (x, y, t) = sin(xt) sin(y), w (x, t) = sinh (x) + S1 [u S [vx v w]] ,
(2 + 1)
v (x, t) = cosh (x) + S1 [u S [wx v w]] ,
(1)n 4n tn (59)
pn : wn (x, y, t) = sin(x) sin(y).
(n + 1) By applying the homotopy perturbation method, and
(53) substituting Eq.(27) in Eq.(59) we have
Thus the solution of Eq.(48) is given by
pn wn (x, t) = sinh (x) +
w(x, y, t) = lim pn wn (x, y, t) n=0
[ [ ( )
p1
n=0
= sin(x) sin(y) pS1 u S q n vn (x, t)
( ) x
4t 42 t2 n=0
]]
1 + +
( + 1) (2 + 1)
q n vn (x, t) pn wn (x, t) ,
(1)n 4n tn n=0 n=0
= sin (x) sin (y)
(n + 1)
n=0 and
= sin (x) sin (y) E (4t ) (54)
If we put 2 in Eq.(44) or solve Eq. (38) and (39) q n vn (x, t) = cosh (x) +
with = 2 , we obtain the exact solution n=0
[ [ (
)
n
(1) (2t) n
w (x, y, t) = sin (x) sin (y) pS1 u S pn wn (x, t)
(2n + 1) x
n=0
n=0
]]
= sin (x) sin (y) cos(2t) q vn (x, t)
n n
p wn (x, t) . (60)
which is in full agreement with the result in Ref. [16, n=0 n=0
38]. Equating the terms with identical powers of p, we get
{
Example 11 Consider the following system of linear w0 (x, t) = sinh(x)
FPDEs p0 :
v (x, t) = cosh(x)
{ { 0 t
Dt w vx + v + w = 0, 1
w1 (x, t) = cosh(x) (+1)
(55) p : t
Dt v wx + v + w = 0, v1 (x, t) = sinh(x) (+1)
t+ cosh(x) t+ sinh(x)
w2 (x, t) = (++1) + (++1)
6 Conclusion
2 cosh(x)
+ t (2+1) Recently, a new transform technique known as
p2 : t+ sinh(x)
v2 (x, t) = t+ cosh(x) Sumudu transform has been introduced by Watugala
(++1) + (++1) [37]. This transform technique has a great potential
2 sinh(x)
+ t (2+1) of applicability to solve FPDEs. In addition, the Chi-
(61) nese mathematician Ji-Huan He [12, 14] introduced
the homotopy perturbation method. This method was
Thus the solution of Eq.(55) is given by a powerful tool to solve various kinds of nonlinear
problems.
( ) In this paper, we have introduced a combination
t+
w(x, t) = sinh(x) 1 + + of the homotopy perturbation method and the Sumudu
( ( + + 1) transform method for time fractional problems. This
t t+ combination builds a strong method called the HPST
cosh(x) +
( + 1)) ( + + 1) method. This method has been successfully applied
t2 to one- and two-dimensional fractional equations and
+
(2 (
+ 1) also for systems of more than two linear and nonlin-
)
t+ ear partial differential equations. The HPSTM is an
v(x, t) = cosh(x) 1 + + analytical method and runs by using the initial con-
( ( + + 1)
ditions only. Thus, it can be used to solve equations
t t+
sinh(x) + with fractional and integer order with respect to time.
( + 1)) ( + + 1) An important advantage of the new approach is its low
t2 computational load.
+ (62)
(2 + 1)
Acknowledgements: The authors would like to
Setting = in Eqs. (62), we reproduce the solution thank Prof. Bjorn-Erik Erlandsson, Maryam Shabany,
of [18] as follows: Mohammad Karbalaie and Mehran Shafaiee for the
friendly scientific atmosphere which resulted in inter-
( ) acting outcome.
t2
w(x, t) = sinh(x) 1 + +
( (2 + 1) )
t t3 References:
cosh(x) + + ...
( (+1) 2 (3+1)) [1] A. A. M. Arafa, S. Z. Rida, H. Mohamed, Ho-
t motopy Analysis Method for Solving Biological
v(x, t) = cosh(x) 1 + + ...
( (2 + 1) ) Population Model, Communications in Theoret-
t t3 ical Physics, Vol. 56, No. 5, 2011.
sinh(x) + + ...
( +1) (3+1) [2] G. Adomian, Solving frontier problems of
(63) physics: The decomposition method, Kluwer
Academic Publishers, Boston and London,
If we put 1 in Eq.(63) or solve Eq. (55) and 1994.
(56) with = = 1 , we obtain the exact solution [3] F. B. M. Belgacem and A. A. Karaballi,
( ) Sumudu transform fundamental properties in-
t2 t4 vestigations and applications, International J.
w(x, t) = sinh(x) 1 + + + Appl. Math.Stoch. Anal, Vol. 2006, DOI
( 2! 3 4! 5 ) 10.1155/JAMSA/2006/91083, 2005, pp 123.
t t
cosh(x) t + + + [4] V. Daftardar-Gejji, S. Bhalekar, Solving frac-
3! 5!
= sinh(x (t) )
tional boundary value problems with Dirich-
t2 t4 let boundary conditions using a new iterative
v(x, t) = cosh(x) 1 + + + method, Computers & Mathematics with
( 2! 3 4! 5 ) Applications, Vol. 59, No. 5, 2010, pp. 1801
t t
sinh(x) t + + + 1809.
3! 5!
[5] S. Bhalekar, V. Daftardar-Gejji, New iterative
= cosh(x t) (64)
method: Application to partial differential equa-
tions, Applied Mathematics and Computer, 203
which is in full agreement with the result in Ref. [19]. (2), 2008, pp 778783.
[6] V. Daftardar-Gejji, S. Bhalekar, Solving frac- [18] H. Jafari, S. Seifi, Solving system of nonlinear
tional boundary value problems with Dirich- fractional partial differention equations by ho-
let boundary conditions using a new iterative motopy analysis method, Commun. Nonlinear
method,Computers and Mathematics with Appli- Sci. Numer.Simul. Vol. 14, 2009, pp. 1962-1969.
cations, Vol. 59, No 5, 2010, pp. 18011809. [19] H. Jafari, M. Nazari, D. Baleanu, and C. M.
[7] J. F. Cheng, Y. M. Chu, Solution to the Khalique. A new approach for solving a sys-
linear fractional differential equation using tem of fractional partial differential equations.
Adomian decomposition method, Mathemati- Computers and Mathematics with Applications,
cal Problems in Engineering, Vol. 2011, 2011, 2012, doi.org/10.1016/j.camwa.2012.11.014.
doi:10.1155/2011/587068. [20] E. Hesameddini, H. Latifzadeh, An optimal
[8] H. Eltayeb, A. Kilicman. Application of choice of initial solutions in the homotopy per-
Sumudu Decomposition Method to Solve Non- turbation method. International Journal of Non-
linear System of Partial Differential Equa- linear Sciences and Numerical Simulation, Vol.
tions. Abstract and Applied Analysis. Hin- 10(11-12), 2011, pp 13891398.
dawi Publishing Corporation, Vol 2012, 2012, [21] A. Karbalaie, M. M. Montazeri, H. H.
doi:10.1155/2012/412948. Muhammed, New Approach to Find the Exact
Solution of Fractional Partial Differential Equa-
[9] K. A. Gepreel, The homotopy perturbation
tion. WSEAS Transactions on Mathematics, Vol.
method applied to the nonlinear fractional Kol-
11(10), 2012, pp. 908917.
mogorov Petrovskii Piskunov equations, Ap-
plied Mathematics Letters, Vol. 24, No. 8, 2011, [22] A. Karbalaie, H. H. Muhammed, B. E. Erlands-
pp. 14281434 son, Using Homo-Separation of Variables for
Solving Systems of Nonlinear Fractional Partial
[10] A. Ghorbani, Beyond Adomian polynomials: He Differential Equations, International Journal of
polynomials, Chaos, Solitons and Fractals, Vol. Mathematics and Mathematical Sciences, Vol.
39, No. 3, 2009, pp. 1486-1492. 2013, 2013, doi.org/10.1155/2013/421378.
[11] V. G. Gupta and B. Sharma, Application of [23] A. Karbalaie, H. H. Muhammed, M. Shabani, M.
Sumudu Transform in Reaction-Diffusion Sys- M. Montazeri, Exact Solution of Partial Differ-
tems and Nonlinear Waves, Applied Mathemat- ential Equation Using Homo-Separation of Vari-
ical Sciences, Vol. 4, 2010, pp 435-446. ables, International Journal of Nonlinear Sci-
[12] J. H. He, Approximate analytical solution for ence, 2013,
seepage flow with fractional derivatives in [24] A. A. Kilbas, M. Saigo, R. K. Saxena, Gener-
porous media, Computer Methods in Applied alized Mittag-Leffler function and generalized
Mechanics and Engineering, vol. 167 (1-2), fractional calculus operators, Integral Trans-
1998, pp. 57- 68. forms and Special Functions, Vol. 15, 2004, pp.
[13] J. H. He, Limit cycle and bifurcation of nonlin- 31-49.
ear problems, Chaos, Solitons & Fractals, Vol. [25] D. Kumar, J. Singh, S. Rathore, Sumudu De-
24, 2005, pp. 827833 composition Method for Nonlinear Equations, In
International Mathematical Forum, Vol. 7, No.
[14] J. H. He, A new approach to nonlinear partial 11, 2012, pp. 515521.
differential equations. Commun Nonlinear Sci
[26] S. Kumar, A. Yildirim, L. Wei, A fractional
Numer Simul, Vol. 2(4), 1997, pp. 230-235.
model of the diffusion equation and its analyt-
[15] R. Hilfer (Ed.), Applications of Fractional Cal- ical solution using Laplace transform. Scientia
culus in Physics, World Scientific, Singapore, Iranica, Vol. 19(4), 2012, pp. 11171123.
2000. [27] F. Mainardi, Fractional diffusive waves in vis-
[16] H. Jafari, S. Momani, Solving fractional dif- coelastic solids. in: J.L. Wegner, F.R. Norwood.
fusion and wave equations by modified homo- (Eds.) IUTAM Symposium-Nonlinear Waves in
topy perturbation method. Physics Letters A, Solids, Fairfield, 1995, pp. 9397.
Vol. 370(5), 2007, pp. 388396 [28] R. L. Magin and M. Ovadia, Modeling the car-
[17] H. Jafari, S. Seifi, A. Alipoor, M. Zabihi, An diac tissue electrode in-terface using fractional
iterative method for solving linear and nonlin- calculus, J. Vibration and Control , Vol. 14,
ear fractional diffusion-wave equation, J. Non- 2008, pp. 14311442.
linear Fract.International e-Journal of Numeri- [29] K. S. Miller, B. Ross, An Introduction to the
cal Analysis and Related Topics, Vol. 3, 2009, fractional Calculus and Fractional Differential
pp. 2032. Equations. John Wiley and Sons, 1993.