You are on page 1of 14

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/325091223

Analytical solution of the nonlinear diffusion equation

Article  in  European Physical Journal Plus · May 2018


DOI: 10.1140/epjp/i2018-12010-6

CITATION READS

1 650

1 author:

Pranay Goswami
Ambedkar University Delhi
57 PUBLICATIONS   234 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Analytical solution of space-time fractional Fokker-Planck equation by homotopy perturbation Sumudu transform method View project

New Trends on Analytic Function Theory View project

All content following this page was uploaded by Pranay Goswami on 11 May 2018.

The user has requested enhancement of the downloaded file.


EPJ Plus
EPJ .org
your physics journal

Eur. Phys. J. Plus (2018) 133: 183 DOI 10.1140/epjp/i2018-12010-6

Analytical solution of the nonlinear diffusion equa-


tion

Ravi Shanker Dubey and Pranay Goswami


Eur. Phys. J. Plus (2018) 133: 183
DOI 10.1140/epjp/i2018-12010-6
THE EUROPEAN
PHYSICAL JOURNAL PLUS
Regular Article

Analytical solution of the nonlinear diffusion equation


Ravi Shanker Dubey1,a and Pranay Goswami2,b
1
Department of Mathematics, AMITY School of Applied Science, AMITY University Rajasthan, Jaipur - 303002, India
2
School of Liberal Studies, School of Liberal Studies, Ambedkar University Delhi - 110006, India

Received: 27 December 2017 / Revised: 6 April 2018


c Società Italiana di Fisica / Springer-Verlag 2018
Published online: 16 May 2018 – 

Abstract. In the present paper, we derive the solution of the nonlinear fractional partial differential equa-
tions using an efficient approach based on the q-homotopy analysis transform method (q-HATM). The
fractional diffusion equations derivatives are considered in Caputo sense. The derived results are graphi-
cally demonstrated as well.

1 Introduction and preliminaries


In applied mathematics, one of the most used concepts is the derivative. The derivative shows the rate of change
in the function. This is also helpful to describe many real phenomena. After discovery of the fractional derivative,
mathematicians solved many complex problems of the real world by the help of fractional calculus. For further detail
about this, we may refer to Caputo [1], Miller and Ross [2], Podlubny [3], Kilbas et al. [4], Hilfer [5], and others [6, 7].
Applications of fractional derivatives can be found in [8–11]. For solving linear and nonlinear differential equations
of integer order, there exist lots of methods, but not may methods exist for fractional differential equations. Few of
the known methods are the Laplace transform method [12], the variational iteration method [13], and the Fourier
transform method [14], etc. Also we know that lots of nonlinear fractional differential equations do not have exact
analytic solutions. To solve such type of problems we use different methods like the homotopy perturbation method [15],
the Adomian decomposition method [16,17], the variation iteration method [18], the homotopy transform method [19]
and so on. The homotopy analysis method has a certain auxiliary parameter h, which provides us an easy way to
control and adjust the convergence of the series solution defined in Lio [20, 21]. Recently, El-Tawil and Huseen [22]
proposed the q-homotopy analysis method (q-HAM). The q-HAM is an efficient approach to solve nonlinear problems
(see [23–25]). This approach consists of auxiliary parameters n and h, when n = 1, q-HAM is reduced into HAM. In this
paper we will consider the fractional model of the nonlinear fractional diffusion equation defined in the following form:
 
α ∂ a ∂u(x, t)
Dt u(x, t) = u , (1)
∂x ∂x
with initial condition
u(x, 0) = ϕ(x), (2)
where Dtα u(x, t)
represents the well-known Caputo fractional derivative of u(x, t). We also discuss and analyze the
different forms of fractional model of the nonlinear diffusion equation. The term ua plays an important role in
the applications of the diffusion process. The fast and slow process of the diffusion is based on a > 0 and a < 0,
respectively. Upon taking a = 0, this equation converts into the fractional partial differential equation. For a detailed
application, see [27–30].
Definition 1. The Riemann-Liouville fractional integral operator of f (t) ∈ Cµ , µ ≥ −1, having order α > 0, is defined
bellow
 t
1
J α [f (t)] = (t − τ )α−1 f (τ )dτ, α > 0, (3)
Γ (α) 0
J 0 [f (t)] = f (t). (4)
For details see [1].
a
e-mail: ravimath13@gmail.com
b
e-mail: pranaygoswami83@gmail.com
Page 2 of 12 Eur. Phys. J. Plus (2018) 133: 183

Definition 2. The definition of fractional derivative of the function f (t) in the Caputo sense is [2]
 t
α m−α n 1
Dt f (t) = J D f (t), = (t − τ )−α+m−1 f (m) (τ )dτ, m − 1 < α ≤ m, (5)
Γ (n − α) 0
where m ∈ N and t > 0.
The property of the Caputo fractional derivative operator and Riemann-Liouville fractional integral operator is
m−1
 f k (0+)tk
Jtα Dtα f (t) = f (t) − . (6)
k!
k=0

Definition 3. The Laplace transform of the Caputo fractional derivative Dtα f (t) is defined below
n−1
L [Dtα f (t)] = sα f¯(s) −

sα−k−1 f k (0+), n − 1 < α ≤ n, (7)
k=0

where L[f (t)] = f¯(s).

2 q-HATM for fractional diffusion equation


On taking the Laplace transform on the fractional diffusion equation (1) and use the initial condition (2), we obtain
  2 
2
ϕ(x) 1 a−1 ∂u a∂ u
L[u(x, t)] = − α L au +u , (8)
s s ∂x ∂x2

we have the nonlinear operator,


 2 
2

ϕ(x) 1 a−1 ∂℘(x, t; q) a ∂ (℘(x, t; q))
N [℘(x, t; q)] = − α L a (℘(x, t; q)) + (℘(x, t; q)) , (9)
s s ∂x ∂x2

where q ∈ [0, 1/n], n ≥ 1 is an embedding parameter and ℘(x, t; q) is a real function of x, t and q.
We are using q-HAM toconstruct the homotopy such as
(1 − nq)L [℘(x, t; q) − u0 (x, t)] = hqN [℘(x, t; q)] , (10)
where u0 (x, t) is an initial guess of u(x, t), L-Laplace transform, h = 0 is an auxiliary parameter, and ℘(x; t; q) is an
unknown function.
It is clear that when q = 0 and q = 1/n, we get from eq. (10)
 
1
℘(x, t; 0) = u0 (x, t), ℘ x, t; = u(x, t). (11)
n
From (11), if we increase q from 0 to 1/n, then our solution ℘(x; t; q) varies from the initial value to the solution, i.e.
u0 (x, t) to u(x, t).
On expanding ℘(x; t; q) by Taylor’s series, we have


℘(x, t; q) = u0 (x, t) + um (x, t)q m , (12)
m=1

where
1 ∂ m ℘(x, t; q) 

um (x, t) = . (13)
m! ∂q m 
q=0

Let u0 (x, t), h, n, be properly selected, the series defined in (12) converges at q = 1/n, hence
∞  m
 1
u(x, t) = u0 (x, t) + um (x, t) , (14)
m=1
n

the result given by eq. (14) must be one of the solutions of eq. (1).
Eur. Phys. J. Plus (2018) 133: 183 Page 3 of 12

Define the vectors ūr = {u0 , u1 , . . . , ur }, differentiating m times eq. (10) with respect to q, and taking q = 0, and
then dividing them by m!, we get
L [um (x, t) − km um−1 (x, t)] = hℜm (um−1 (x, t)), (15)
where
 2 
2
   
ϕ(x) km 1 a−1 ∂(um−1 ) a ∂ (um−1 )
ℜm (um−1 (x, t)) = L [um−1 ] − 1− − α L a (um−1 ) + (um−1 ) (16)
s n s ∂x ∂x2
and
0, m≤1
km = (17)
n, otherwise.
On taking the inverse Laplace transform on eq. (15), we get
um (x, t) = km um−1 (x, t) + hL−1 [ℜm (um−1 )] . (18)
On using the initial guess defined in eq. (2) in recursive relation defined in eq. (18), we get the terms of series solution:

2

u1 (x, t) = −h a(ϕ(x))a−1 (ϕ′ (x)) + (ϕ(x))a ϕ′′ (x) . (19)
Γ (α + 1)
For all m ≥ 2, using eq. (18), we get the following m-th–order approximation:

 
∂ a ∂um−1
um = (n + h)um−1 − h (um−1 ) . (20)
Γ (α + 1) ∂x ∂x
In the end, choosing suitable values of n and h, we get the q-HATM series solution
∞  m
 1
u(x, t) = lim um (x, t) . (21)
N →∞
m=0
n

Due to the factor of (1/n)m , we get the faster convergence of the series.

3 Application of q-HATM
In this section we discuss the following applications of the q-homotopy analysis transform method.

3.1 The fractional heat equation


We finally study the nonlinear diffusion equation
∂ 2 u(x, t)
Dtα u(x, t) = , 0 ≤ x, t ≤ 1; 0 < α ≤ 1, (22)
∂x2
with the initial condition
φ(x) = sin(πx). (23)
Using the q-HATM defined in sect. 3, we get the following successive approximations:
u0 (x, t) = sin(πx), (24)

u1 (x, t) = hπ 2 sin(πx) , (25)
Γ (α + 1)
π 4 t2α
u2 (x, t) = (n + h)u1 + h2 sin(πx), (26)
Γ (2α + 1)
π 4 t2α π 6 t3α
u3 (x, t) = (n + h)u2 + (n + h)h2 sin(πx) + h3 sin(πx),
Γ (2α + 1) Γ (3α + 1)
π 4 t2α
u4 (x, t) = (n + h)u3 + (n + h)2 h2 sin(πx)
Γ (2α + 1)
π 6 t3α π 8 t4α
+ 2h3 (n + h) sin(πx) + h4 sin(πx),
Γ (3α + 1) Γ (4α + 1)
..
.
Page 4 of 12 Eur. Phys. J. Plus (2018) 133: 183

By substituting these values from eqs. (24), (25), (26), . . . , in eq. (21) we get the approximate solution fractional heat
equation.

3.2 The fractional slow diffusion processes

In this section we study the following nonlinear fractional slow diffusion equation
 
α ∂ 2 ∂u(x, t)
Dt u(x, t) = u , (27)
∂x ∂x

with the initial condition


x+b
.ϕ(x) = (28)
2c
Using the q-HATM defined in sect. 3, we obtain the following successive approximations:

x+b
u0 (x, t) = , (29)
2c
α
(x + b)t
u1 (x, t) = −h 3 , (30)
4c Γ (α + 1)
 2
t2α

3h (x + b)
u2 (x, t) = (n + h)u1 + , (31)
8c5 Γ (2α + 1)
(x+b) t2α 9h3 (x+b) t3α 3h3 (x+b)Γ (2α+1) t3α
u3 (x, t) = (n+h)u2 −h2 (n+h) · · + · · + · 2 · ,
8c5 Γ (2α+1) 16 c7 Γ (3α+1) 16 c7 (Γ (α+1)) Γ (3α+1)
3h2 (n + h)2 (x + b) t2α 3h3 (n + h) (x + b)Γ (2α + 1) t3α
u4 (x, t) = (n + h)u3 + · 5
· − · 7 2
·
8 c Γ (2α + 1) 8 c (Γ (α + 1)) Γ (3α + 1)
3 3α 4 4α
3h (n + h) (x + b) t 9h (x + b)Γ (3α + 1) t
− · · + · ·
8 c7 Γ (3α + 1) 16 c9 (Γ (α + 1))(Γ (2α + 1)) Γ (4α + 1)
27h4 (x + b) t4α 9h4 (x + b)Γ (2α + 1) t4α h4 (x + b)Γ (3α + 1) t4α
+ · · + · · + · · ,
32 c9 Γ (4α + 1) 32 c9 (Γ (α + 1))2 Γ (4α + 1) 32 c9 (Γ (α + 1))3 Γ (4α + 1)
..
.

By substituting these values from eqs. (29), (30), (31), . . . , in eq. (21) we get the required solution fractional slow
diffusion equation.

3.3 The fractional fast diffusion processes

In this section we study the nonlinear fractional fast diffusion equation


 
α ∂ −2 ∂u(x, t)
Dt u(x, t) = u , (32)
∂x ∂x

with the initial condition


1
ϕ(x) = , (33)
(1 + x2 )
Using the q-HATM defined in sect. 3, we get the following terms:
−1/2
u0 (x, t) = 1 + x2

, (34)

−1/2 −3/2 tα
u1 (x, t) = h 1 + x2 − x2 1 + x2

, (35)
Γ (α + 1)
1 − 2x2 t2α
 
u2 (x, t) = (n + h)u1 − h2 2 5/2
, (36)
(1 + x ) Γ (2α + 1)
Eur. Phys. J. Plus (2018) 133: 183 Page 5 of 12

Fig. 1. The behavior of the q-HATM solution u(x, t) of Case I, with respect to x for different values of h when α = 1, t = 0.5.

(1 − 2x2 ) t2α (34x2 − 4x4 − 7) t3α


   
u3 (x, t) = (n + h)u2 − h2 (n + h) 2 5/2
+ h 3
2 7/2
(1 + x ) Γ (2α + 1) (1 + x ) Γ (3α + 1)
2 3α
   
(3 − 12x ) Γ (2α + 1) t
+ h3 · ,
(1 + x2 )7/2 (Γ (α + 1))2 Γ (3α + 1)
..
.

By substituting these values from eqs. (34), (35), (36), . . . , in eq. (21) we get the approximate solution fractional fast
diffusion equation.

4 Special cases
In this section, we discuss the following special cases of our main result.

Case I
If we take a = 0 and α = 1 in eq. (1), it reduces to the heat equation

∂u(x, t) ∂ 2 u(x, t)
= (37)
∂t ∂x2
and if we put n = 1, h = −1 and φ(x) = sin(πx) in our result (20) we get
2
u(x, t) = sin(πx)e−π t , (38)

which is the same result as obtained by Sulaiman et al. see [29].

Remark 1. If we fix φ(x) = sin(πx) and take various values of n and h, then we get different values of u corresponding
to n and h, which is given in fig. 1.

Case II
If we take a = 2 and α = 1 in eq. (1), it reduces to slow diffusion equation
 
∂u(x, t) ∂ ∂u(x, t)
= u2 , (39)
∂t ∂x ∂x

and, if we put in our result, we obtain


  
x+a 1 3 2 5 3 35 4
u(x, t) = 1 + 2t+ 4t + t + t + ... , (40)
2c 2c 8c 16c6 128c8
Page 6 of 12 Eur. Phys. J. Plus (2018) 133: 183

Fig. 2. The behavior of the q-HATM solution u(x, t) of Case II, with respect to x for different values of h when α = 1, t = 0.5.

and hence consequently the exact solution


(x + a)
u(x, t) = √ , t < c2 , (41)
2 c2 − t
the above infinite series converges to the exact solution of the slow diffusion equation, the same as [26].
Remark 2. If we take various values of n and h, then we get different values of u corresponding to n and h, which is
given in fig. 2.

Case III

If we take a = 2 and α = 1 in eq. (1), it reduces to fast diffusion equation


 
∂u(x, t) ∂ −2 ∂u(x, t)
= u , (42)
∂t ∂x ∂x
and, if we put n = 1, h = −1 and φ(x) = (1 + x2 )−1/2 in our result, we get
x2 t2 (1 − 2x2 )
 
2 −1/2
 1
u(x, t) = 1 + x +t − + + ..., (43)
(1 + x2 )3/2 (1 + x2 )1/2 2! (1 + x2 )5/2
which tends to the exact solution
1
u(x, t) = √ . (44)
x + e2t
2

The above infinite series converges to the exact solution and has the asymptotic behavior, of the Fast Diffusion
equation, same as [26].
Remark 3. If we take various values of n and h, then we get different values of u corresponding to n and h, which is
given in fig. 3.

5 Conclusion
In this paper we presented an analytical solution of nonlinear diffusion equations. The q-HAM deals with the auxiliary
parameter h and asymptotic parameter n, that provide a simple way to adjust and control the convergence of the
series solutions of linear or nonlinear problems by selecting the suitable values of h and n. Graphical representations
of the solutions of fractional nonlinear diffusion equations are given below (figs. 4 to 8 for Case I; figs. 9 to 13 for Case
II; and figs. 14 to 18 for Case III). Further we have also discussed the effect on the solution of parameter h and n
in our solution (see figs. 1, 2 and 3). Our results show that q-HATM can be applied to many complicated linear and
strongly nonlinear partial differential equations. Thus, we can conclude that q-HAM is a more powerful computational
technique for investigating the fractional problem so it is better than HAM because of the choice of suitable values of
n and h.
Eur. Phys. J. Plus (2018) 133: 183 Page 7 of 12

Fig. 3. The behavior of the q-HATM solution u(x, t) of Case III, with respect to x for different values of h when α = 1, t = 0.5.

Fig. 4. The behavior of the q-HATM solution u(x, t) of Case I, with respect to x and t when α = 1, h = −1, n = 1.

Fig. 5. The behavior of the q-HATM solution u(x, t) of Case I, with respect to t for different values of α when x = 0.3, h = −1,
n = 1.
Page 8 of 12 Eur. Phys. J. Plus (2018) 133: 183

Fig. 6. The behavior of the q-HATM solution u(x, t) of Case I, with respect to h for different values of n when x = 0.3, α = 1,
t = 0.3.

Fig. 7. The behavior of the q-HATM solution u(x, t) of Case I, with respect to h for different values of α when x = 0.3, t = 0.3,
n = 1.

Fig. 8. The behavior of the q-HATM solution u(x, t) of Case I, with respect to n for different values of α when x = 0.3, h = −1,
t = 0.3.
Eur. Phys. J. Plus (2018) 133: 183 Page 9 of 12

Fig. 9. The behavior of the q-HATM solution u(x, t) of Case II, with respect to x and t when α = 1, h = −1, n = 1, b = c = 1.

Fig. 10. The behavior of the q-HATM solution u(x, t) of Case II, with respect to t for different values of α when x = 0.3,
h = −1, n = 1, b = c = 1.

Fig. 11. The behavior of the q-HATM solution u(x, t) of Case II, with respect to h for different values of n when x = 0.3,
α = 1, t = 0.3, b = c = 1.
Page 10 of 12 Eur. Phys. J. Plus (2018) 133: 183

Fig. 12. The behavior of the q-HATM solution u(x, t) of Case II, with respect to h for different values of α when x = 0.3,
t = 0.3, n = 1, b = c = 1.

Fig. 13. The behavior of the q-HATM solution u(x, t) of Case II, with respect to n for different values of α when x = 0.3,
h = −1, t = 0.3, b = c = 1.

Fig. 14. The behavior of the q-HATM solution u(x, t) of Case III, with respect to x and t when α = 1, h = −1, n = 1.
Eur. Phys. J. Plus (2018) 133: 183 Page 11 of 12

Fig. 15. The behavior of the q-HATM solution u(x, t) of Case III, with respect to t for different values of α when x = 0.3,
h = −1, n = 1.

Fig. 16. The behavior of the q-HATM solution u(x, t) of Case III, with respect to h for different values of n when x = 0.3,
t = 0.3, α = 1.

Fig. 17. The behavior of the q-HATM solution u(x, t) of Case III, with respect to h for different values of α when x = 0.3,
t = 0.3, n = 1.
Page 12 of 12 Eur. Phys. J. Plus (2018) 133: 183

Fig. 18. The behavior of the q-HATM solution u(x, t) of Case III, with respect to n for different values of α when x = 0.3,
h = −1, t = 0.3.

The authors express their sincere gratitude to the anonymous referee for his/her sincere comments to improve this paper.

References
1. M. Caputo, Elasticità e Dissipazione (Zanichelli, Bologna, 1969).
2. K.S. Miller, B. Ross, An Introduction to the Fractional Calculus and Fractional Differential Equations (John Wiley and
Sons, New York, Chichester, Brisbane, Toronto and Singapore, 1993).
3. I. Podlubny, Fractional Differential Equations (Academic Press, New York, 1999).
4. A.A. Kilbas, H.M. Srivastava, J.J. Trujillo, Theory and Applications of Fractional Differential Equations, in North-Holland
Mathematical Studies, Vol. 204 (Elsevier North-Holland Science Publishers, Amsterdam, London and New York, 2006).
5. R. Hilfer (Editor), Applications of Fractional Calculus in Physics (World Scientific Publishing Company, Singapore, New
Jersey, London and Hong Kong, 2000) pp. 87–130.
6. V.B.L. Chaurasia, R.S. Dubey, F.B.M. Belgacem, Math. Eng. Sci. Aerospace 3, 179 (2012).
7. V.B.L. Chaurasia, R.S. Dubey, ISRN Appl. Math. 2011, 682381 (2011).
8. M. Majumder, N. Chopra, R. Andrews, Nature 438, 44 (2005).
9. B.S. Alkahtani, O.J. Alkahtani, R.S. Dubey, P. Goswami, J. Nonlinear Sci. Appl. 10, 299 (2017).
10. Y. Xuan, W. Ro, Int. J. Heat Mass Transfer 43, 3701 (2008).
11. B.S. Alkahtani, O.J. Algahtani, R.S. Dubey, P. Goswami, Entropy 19, 114 (2017).
12. X.K. Li, J.G. Peng, Appl. Math. Lett. 24, 2019 (2011).
13. X. Yang, D. Baleanu, Therm. Sci. 17, 625 (2013).
14. L. Gaul, P. Klein, S. Kenpe, Mech. Res. Commun. 16, 586 (1989).
15. H.-J. He, Comput. Math. Appl. Mech. Eng. 178, 257 (1999).
16. G. Adomian, J. Math. Anal. Appl. 135, 501 (1988).
17. A.M. Wazwaz, S.M. El-Sayed, Appl. Math. Comput. 122, 393 (2001).
18. H. Jafari, H. Tajadodi, Int. J. Differ. Equ. 2010, 764738 (2010).
19. R.S. Dubey, B.S.T. Alkahtani, A. Atangana, Math. Prob. Eng. 2015, 780929 (2015).
20. S.J. Liao, Commun. Nonlinear Sci. Numer. Simul. 14, 83 (2009).
21. S.J. Liao, Appl. Math. Comput. 169, 86 (2005).
22. M.A. El-Tawil, S.N. Huseen, Int. J. Appl. Math. Mech. 8, 51 (2012).
23. M.A. El-Tawil, S.N. Huseen, Int. J. Contemp. Math. Sci. 8, 481 (2013).
24. S.N. Huseen, S.R. Grace, J. Appl. Math. 2013, 569674 (2013).
25. S.N. Huseen, S.R. Grace, M.A. El-Tawil, Int. J. Comput. Technol. 11, 2859 (2013).
26. A.M. Wazwaz, Comput. Math. Appl. 54, 933 (2007).
27. G. Gurarslan, M. Sari, Int. J. Numer. Math. Biomed. Eng. 27, 69 (2009).
28. E.A. Saired, Appl. Math. Comput. 98, 103 (1999).
29. J. Sulaiman, M. Othman, M.K. Hasan, Int. J. Comput. Math. 81, 1559 (2004).
30. V.B.L. Chaurasia, R.S. Dubey, Fract. Differ. Calculus 3, 21 (2013).

View publication stats

You might also like