R AlAhmad
R AlAhmad
Rami AlAhmad
Department of Mathematics
Faculty of Science
Yarmouk University
Irbid, Jordan
Abstract
In this paper, we introduce a procedure to approximate the Grünwald-
Leitnikov fractional derivative of order α ∈ (0, 1) for a product of two
functions. This procedure uses the Laplace transform for a product of
functions.
1 Introduction
In fractional calculus, one can take the derivative a non-integer number of
times. An introduction for fractional derivatives was given in [4]. Histori-
cally, the concept of the fractional calculus is more than 300 years old. Oliver
Heaviside introduced the practical use of fractional differential operators in
electrical transmission back in 1890. Recently, scientists and engineers have
come to realize the potential fractional calculus presents in applied science.
Research has been done successfully on applications using models and tools
from fractional calculus , see [7], [8], [9], and [10]. The Grünwald-Letnikov
derivative was introduced by Anton Karl Grünwald in 1867 and by Alek-
sey Vasilievich Letnikov in 1868. The Grünwald-Letnikov derivative [10] is
defined as:
Key words and phrases: Fractional derivatives, Grünwald-Leitnikov
fractional derivatives , Laplace transform.
AMS (MOS) Subject Classifications: 44A10, 26A33.
ISSN 1814-0432, 2021, http://ijmcs.future-in-tech.net
768 R. AlAhmad
We note that the Laplace transform of the product of two functions is not
the product of their Laplace transforms. However, the Laplace transform of a
convolution of two functions equals the product of their Laplace transforms.
For example,
∞ ∞
X tn X tn
ψ1,1 (t) = =
n=1
nΓ(n) n=1 Γ(n + 1)
∞ n
X t
= = et − 1,
n=1
n!
2 Preliminaries
In this section, we will give the fractional derivative of order α for products
of two functions. Also, using these results, Lemma 2.1,Theorem 2.2 and
differential transform allows us to solve fractional differential equations.
Lemma 2.1. [10] For α ∈ (0, 1), the Laplace transform of Grünwald-Letnikov
derivative of (D α f )(t) is sα F (s).
Theorem 2.2. Assume that L(f (t)) = F (s) and L(g(t)) = G(s).
R ∞ R ∞ −(s+ξ)t
If 0 0 e g(ξ)f (t)dtdξ converges absolutely for s > b, then Ls (f (t)G(t))
is given as
Z ∞ Z ∞
Ls f (t)G(t) = Ls f (t)L (g(t)) = g(ξ)F (ξ+s)dξ = g(ξ−s)F (ξ)dξ for s > b.
0 s
(2.3)
R∞R∞
This theorem can be stated as: if 0 0
e−(s+ξ)t (L−1 h)(ξ)f (t)dtdξ con-
verges absolutely for s > b, then
Z ∞ Z ∞
−1
Ls f (t)h(t) = (L h)(ξ)(Lf )(ξ+s)dξ = (L−1 h)(ξ−s)(Lf )(ξ)dξ for s > b.
0 s
(2.4)
770 R. AlAhmad
s
Example 2.3. Using the fact that L (cos(at)) = s2 +a2 , we get
tf (t)
Ls 2 = L (f (t)L(cos(at)))
t + a2
Z ∞
= cos(ξ − s)F (ξ)dξ
s
Z ∞ Z ∞
= cos(s) cos(ξ)F (ξ)dξ − sin(s) sin(ξ)F (ξ)dξ.
s s
3 Main result
Using Theorem 2.2 and Lemma 2.1, we have the following result:
Theorem 3.1. The Grünwald-Letnikov derivative of order α ∈ (0, 1) of f g
is given by:
Z ∞ Z ∞
α −1 α −1 −1 α
(D f g)(t) = L {s (L g)(ξ)(Lf )(ξ+s)dξ} = L {s (L−1 g)(ξ−s)(Lf )(ξ)dξ}.
0 s
Proof. According to Lemma 2.1, the Laplace transform of (D α f g)(t) is sα L f (t)g(t) .
Therefore,
(D α f g)(t) = L−1 {sα L f (t)g(t))}. (3.5)
Now, using Theorem 2.2, we get:
Z ∞ Z ∞
−1
Ls f (t)g(t) = (L g)(ξ)(Lf )(ξ + s)dξ = (L−1 g)(ξ − s)(Lf )(ξ)dξ.
0 s
(3.6)
Combining (3.5) and (3.6), the result follows.
As applications we have the following examples:
Example 3.2. For α ∈ (0, 1), we prove that
D α te−t = t1−α E1,1−α (t) + αE1,2−α (t) .
Now,
Z ∞
α −t −1 α −2
D te =L s u δ(u − s − 1)du
s
−1 α −2
=L s (1 + s) .
Grünwald-Leitnikov fractional derivative for a product of two functions 771
Therefore,
∞
sα X n+1
2
= (−1)n n+2−α
(s + 1) n=0
s
∞
X (n + 1)Γ(n + 2 − α)
= (−1)n .
n=0
Γ(n + 2 − α)sn+2−α
Hence,
sα X ∞ n+1−α
−1 n (n + 1)t
L = (−1)
(s + 1)2 n=0
Γ(n + 2 − α)
∞
X (n + 1)tn+1−α
= (−1)n
n=0
(n + 1 − α)Γ(n + 1 − α)
∞
X (n + 1 − α + α)tn+1−α
= (−1)n
n=0
(n + 1 − α)Γ(n + 1 − α)
∞ ∞
X tn+1−α X tn+1−α
= (−1)n +α (−1)n
n=0
Γ(n + 1 − α) n=0
(n + 1 − α)Γ(n + 1 − α)
∞ ∞
1−α
X (−t)n 1−α
X (−t)n
=t + αt
Γ(n + 1 − α) Γ(n + 2 − α)
n=0 n=0
= t1−α E1,1−α (t) + αE1,2−α (t) .
α 1−α
Therefore, D te −t
=t E1,1−α (t) + αE1,2−α (t) .
Interestingly, limα−→1 D α te−t = −te−t + e−t = D te−t .
772 R. AlAhmad
Hence,
∞ ∞
1 X (−1)k X (−1)k
sα (s + 1) ln(1 + ) − 1 = − k−1−α
− k−α
.
s k=2
ks k=1
ks
∞ ∞
α
e−t + t − 1
−1
X (−1)k X (−1)k
D =L − −
t2 k=2
ksk−1−α k=1 ksk−α
∞ ∞
X (−1)k tk−α−2 X (−1)k tk−α−1
= −
k=2
kΓ(k − 1 − α) k=1 kΓ(k − α)
∞ ∞
X (−t)k X (−t)k
= t−α−2 − t−α−1 .
k=2
kΓ(k − 1 − α) k=1
kΓ(k − α)
References
[1] S. Al-Ahmad, M. Mamat, R. AlAhmad, I. M. Sulaiman, P. L. Ghazali,
M. A. Mohamed, On New Properties of Differential Transform via Dif-
ference Equations, International Journal of Engineering and Technology,
7, 3.28, (2018), 321–324.