A sum up method for solving summations of the form

Pn
k=n0 An,k and rising and falling factorial transforms

Parham Zarghami1,∗
arXiv:2312.04144v1 [math.NT] 7 Dec 2023

a University of Tehran, School of Electrical and Computer Engineering, University College

of Engineering, North Kargar street, Tehran, 1439957131, Iran

Abstract
In this paper, we discuss P a method that utilizes the recurrence of An,k to solve
n
summations of the form k=n0 An,k . It is observed that by repeating the pro-
cedure, the upper bound of summation is reduced and tilts toward the lower
bound. This method of summation is mostly suitable for combinatorial se-
quences such as binomial coefficients, Stirling numbers of both kinds, etc. After
the main method is displayed, some examples are illustrated. Some useful identi-
ties about Stirling and r-Stirling numbers are obtained. Finally, two transforms
called rising and falling factorial transforms which turn the basis of power poly-
nomials into factorial basis are derived. These transforms verify and simplify
the results obtained in the examples section. Also, these transforms describe the
relationship between fractional derivatives (or fractional integrals) and falling
factorial (or rising factorial) by its series expansion.
Keywords: Fractional calculus, Summation, Combinatorial sequence,
Factorial, Transformation, Integral representation

1. Introduction

There are several formulas and methods, e.g., telescoping series, that focus on
solving special summations. A summation is a way of accumulating information
in an explicit form. However, some information may be similar and can be
summed up before the summation. The aim of this research is to find a method
of summing up the same terms utilizing recurrences between different terms
of the sequence. This way, the bound of summations dwindles away until it
becomes trivial.
Table 1 shows a number of sequences with their recurrence relations. Table
2 illustrates definition of required functions.

∗ Corresponding author
Email address: [email protected] (Parham Zarghami)
1 Declarations of interest: None

Preprint submitted to Elsevier December 2023

 Table 1: Recurrence relation of sequences
Sequence Recurrent
relation
n n−1
+ n−1



Binomial coefficients (A007318) =
n k k k−1
= (n − 1) n−1 + n−1
  
Stirling numbers of the first kind (A008275) k k k−1
n
= k n−1 + n−1
 
Stirling numbers of the second kind (A008277)
n k k
n−1 k−1
= (n − 1) k r + n−1
 
r-Stirling numbers of the first kind [6] kr k−1 r
n
= k n−1 + n−1
 
r-Stirling numbers of the second kind [6] k r k r k−1 r

Table 2: Definition of required functions

Function Definition
Qn−1 Γ(x+n)
Rising factorial [1, 2] xn = i=0 (x + i) = Γ(x)
Qn−1 Γ(x+1)
Falling factorial [1, 2] (x)n = i=0 (x − i) = Γ(x−n+1)
Pn  n k
Touchard (or Bell) polynomials [7] Tn (x) = k=0 k x
Tn,r (x) = n n+r
xk
P
r-Touchard polynomials [7]
R k=0 k+r r
∞ s−1 −t
Incomplete gamma function [1] Γ(s, x) = x t e dt

The following formulas show some features of incomplete gamma function

[1]:
Z ∞ n
X n! k −x
Γ(n + 1, x) = tn e−t dt = x e (1)
x k!
k=0

Γ(s + 1, x) = sΓ(s + 1, x) + xs e−x (2)

In section 2, we will propose our method and show how some specific sum-
mations tend to assume an explicit form through recurrences. In section 3, some
applications of our proposed method are demonstrated. In section 4, two new
transformation named rising and falling factorial transforms that can change the
basis of polynomial and series from power to factorial ones is proposed. Finally,
suggestions for future works and conclusions are presented.

2. The method

Our proposed method is similar to the Telescoping series. The difference is

in summing up the terms instead of eliminating the similar terms. To illustrate
how our method works, we consider an example recurrence:
m
X
An,k = ai (n, k)An−1,k−i (3)
i=0

Summation on k:
n
X n X
X m n−1
X
An,k = ai (n, k)An−1,k−i = c1,i (n)An−1,i (4)
k=n0 k=n0 i=0 k=n0

2
Then, An,m = 0 and ai (n, m) = 0 for n < m and m < n0 is considered. Based
on above process, c1,i (n) can be defined as follows:
k
X
c1,i (n) = aj (n, i + j) (5)
j=0

This procedure is repeated as the following:

m
X
c1,k (n)An,k = ai (n, k)An−1,k−i (6)
i=0

n
X n−1
X n−2
X
An,k = c1,k (n)An−1,k = c2,k (n)An−2,k (7)
k=n0 k=n0 k=n0

Again based on above processes, c2,i (n) can be formulated as:

k
X
c2,i (n) = c1,i+j (n)aj (n − 1, i + j) (8)
j=0

After that, the following formulations can be constructed:

n
X n−1
X n−2
X
An,k = c1,k (n)An−1,k = c2,k (n)An−2,k = ... = cn−n0 ,n0 (n)An0 ,n0
k=n0 k=n0 k=n0
(9)
k
X
cs,i (n) = cs−1,i+j (n)aj (n − s + 1, i + j) (10)
j=0

This method may be applied to the super-recurrence of the form:

n n−1 n−1
= f (n, k) + g(n, k) + [n = k = 0] (11)
k k k−1
For instance in [3], the coefficients used are f (n, k) = αn + βk + γ, g(n, k) =
α n + β k + γ . By the definition, 00 = 1 and nk = 0 when n < k. So, the
′ ′ ′

value of nn in the following form can be calculated by the above recurrence:

n n−1
= g(n, n)
n n−1
n (12)
n−2 Y
= g(n, n)g(n − 1, n − 1) = ... = g(i, i)
n−2 i=1
Pn n
(9) and (10) are used for polynomials of the form i=n0 k xk as follows:
n n0
X n k n0 Y
x = xn0 Yn−n0 ,n0 (x) = xn0 Yn−n0 ,n0 (x) g(i, i) (13)
i=n
k n0 i=1
0

3
Where Ym,k (x) has a recurrence as below:
Ym,k (x) = f (n − m + 1, k)Ym−1,k (x) + xg(n − m + 1, k + 1)Ym−1,k+1 (x) (14)
Here n is the upper bound of the summation and Y0,k (x) = 1, Y1,k (x) = f (n, k)+
xg(n, k + 1). (9) and (10) are used for rising factorial polynomials of the form
Pn n k
k=n0 k x as follows:
n n0
X n k n0 Y
x = xn0 yn−n0 ,n0 (x) = xn0 yn−n0 ,n0 (x) g(i, i) (15)
k n0 i=1
k=n0

Where ym,k (x) has a recurrence as below:

ym,k (x) = f (n−m+1, k)ym−1,k (x)+(x+k)g(n−m+1, k +1)ym−1,k+1 (x) (16)
n−1
g(n, n0 ) n0 −1 = 0 is considered at the middle of calculations of both poly-
nomials. In (16), n is the upper bound of the summation and y0,k (x) = 1,
y1,k (x) = f (n, k) + (x + k)g(n, k + 1).

3. Examples
In this section, we will focus on special cases of super-recurrence involved in
polynomials discussed in the previous section. Table 3 illustrates recurrence of
three famous combinatorial sequences:

Table 3: Recurrences of sequences

Sequence Recurrence f (n, k) g(n, k)
n n−1 n−1
Binomial coefficients k
= k
+ k−1
1 1
n n−1 n−1
Stirling numbers of the first kind k
= (n − 1) k
+ k−1
n-1 1
n n−1 n−1
Stirling numbers of the second kind k
=k k
+ k−1
k 1

Recurrence relation of Ym,k (x) and ym,k (x) illustrated as Table 4 and Table
5 respectively.

Table 4: Ym,k (x) of sequences

Sequence Ym,k (x)
Binomial coefficients Ym,k (x) = Ym−1,k (x) + xYm−1,k+1 (x)
Stirling numbers of the first kind Ym,k (x) = (n − m)Ym−1,k (x) + xYm−1,k+1 (x)
Stirling numbers of the second kind Ym,k (x) = kYm−1,k (x) + xYm−1,k+1 (x)

It is observed that in all cases Y0,k (x) = 1 and y0,k (x) = 1 are present.
For solving the above recurrences, we do not need to use complicated methods
while we can use the recurrence unfolding method [3]. It means that we must
replace nth term with (n − 1)th terms involved in the main recurrence. We can
perform iterations in a top-down and bottom-up manner. To avoid repeated
calculations, we have displayed the results only.

4
 Table 5: ym,k (x) of sequences
Sequence ym,k (x)
Binomial coefficients ym,k (x) = ym−1,k (x) + (x + k)ym−1,k+1 (x)
Stirling numbers of the first kind ym,k (x) = (n − m)ym−1,k (x) + (x + k)ym−1,k+1 (x)
Stirling numbers of the second kind ym,k (x) = kym−1,k (x) + (x + k)ym−1,k+1 (x)

3.1. Solving Ym,k (x) = Ym−1,k (x) + xYm−1,k+1 (x)

Ym,k (x) is independent of k, so the recurrence takes the form Ym,k (x) =
(x + 1)Ym−1,k (x) form. Hence:

Yn,k (x) = (x + 1)n (17)

So, for the binomial coefficients involved in following polynomial:

n  
X n i
x = Yn,0 (x) = (x + 1)n (18)
i=0
i

3.2. Solving ym,k (x) = ym−1,k (x) + (x + k)ym−1,k+1 (x)

First, we want to show the results by using top-down unfolding method:
n−1
X 
n−1
yn,k (x) = (x + k + i + 1)(x + k)i (19)
i=0
i
i
We know that ∆x ((x + k)i ) = i(x + k + 1)i−1 = i (x+k)
x+k :

yn,k (x) = (x + k + 1)yn−1,0 (x + k) + (x + k)∆x yn−1,0 (x + k + 1) (20)

Is simplified to:

yn,k (x) = yn−1,0 (x + k) + (x + k)yn−1,0 (x + k + 1) = yn,0 (x + k) (21)

Pn
We used yn,0 (x) = i=0 ni xi to obtain (21). By using x = 1 and (1):

n   n
X n X n!
yn,0 (1) = i! =
i i!
i=0 i=0
Z ∞ (22)
= eΓ(n + 1, 1) = e tn e−t dt
1

Without loss of generality, in (21), we can consider k = 0 and x = m which m

is an integer:

myn,0 (m + 1) = yn+1,0 (m) − yn,0 (m) = ∆n yn,0 (m) (23)

Taking m difference on n (i.e. ∆m−1

n ) from both side of equation results in:
Z ∞
1 m−1 e
yn,0 (m) = ∆ yn,0 (1) = ∆m−1 (tn )e−t dt (24)
(m − 1)! n (m − 1)! 1 n

5
As the difference of exponential function is ∆m−1
n tn = (t − 1)m−1 tn :
Z ∞
e
yn,0 (m) = (t − 1)m−1 tn e−t dt (25)
(m − 1)! 1

m is an integer variable. Thus, a general form like the following is needed:

Z ∞
1
yn,0 (x) = (t + 1)n tx−1 e−t dt (26)
Γ(x) 0

This verifies main recurrence. Then we arrive in an explicit form for yn,k (x) by
(21): Z ∞
1
yn,k (x) = yn,0 (x + k) = (t + 1)n tx+k−1 e−t dt (27)
Γ(x + k) 0
Pn
And integral representation for k=0 nk xk :

n  
X n k
x = yn,0 (m)
k
k=0
Z ∞ (28)
1
= (t + 1)n tx−1 e−t dt
Γ(x) 0

3.3. Solving Ym,k (x) = (n − m)Ym−1,k (x) + xYm−1,k+1 (x)

With unfolding method we can find the fact that Ym,k (x) is independent
from k similar to the recurrence that we discussed in 3.1. So, this recurrence
takes the following form:

Ym,k (x) = (x + n − m)Ym−1,k (x) (29)

We conclude that:
Ym,k (x) = (x + n − m)m (30)
Then, the explicit form of intended polynomial is easily obtained:
n  
X n
xk = Yn,0 (x) = xn (31)
k
k=0

3.4. Solving ym,k (x) = (n − m)ym−1,k (x) + (x + k)ym−1,k+1 (x)

First of all, we should change ym,k (x) to sm,k (x) as below:

sm,k (x) = Γ(x + k)ym,k (x), s0,k (x) = Γ(x + k) (32)

Next, using bottom-up unfolding method we get:

m X j   
X m j
sm,k (x) = (−1)m−j (n − 1)j−i Γ(x + k + i) (33)
j=0 i=0
j i

6
Applying definition of gamma function:
m X j    Z ∞
X m j
sm,k (x) = (−1)m−j (n − 1)j−i tx+k+i−1 e−t dt (34)
j=0 i=0
j i 0

It is simplified as:
Z m
∞X j   
X m j
sm,k (x) = ( (−1)m−j (n − 1)j−i tx+k+i−1 )e−t dt
0 j=0 i=0
j i
Z ∞X m  
m (35)
= (−1)m−j (t + n − 1)j tx+k−1 e−t dt
0 j=0
j
Z ∞
= (t + n − 1)m tx+k−1 e−t dt
0

And by (32) we get the intended explicit formula:

Z ∞
1
ym,k (x) = (t + n − 1)m tx+k−1 e−t dt (36)
Γ(x + k) 0

Suppose k = 0 and m = n:
Z ∞
1
yn,0 (x) = (t + n − 1)n tx−1 e−t dt
Γ(x) 0
Z ∞ (37)
1
= tn tx−1 e−t dt
Γ(x) 0

Then for the Stirling number of the first kind involved in rising factorial poly-
nomial we have the formula below:
n   Z ∞
X n k 1
x = yn,0 (x) = tn tx−1 e−t dt (38)
k Γ(x) 0
k=0

3.5. Solving Ym,k (x) = kYm−1,k (x) + xYm−1,k+1 (x)

By top-down unfolding method we can derive the following formula:
n  
X n+k
Yn,k (x) = xi = Tn,k (x) (39)
i=0
i+k k

In which Tr,k (x) is r-Touchard polynomials shown in section 1. From [6] we

know that:
  i  
n+k 1X i
= (−1)i−j (k + j)n (40)
i+k k i! j=0 j

7
So:
n  
X n+k
Yn,k (x) = xi
i=0
i + k k
n i  
X 1X i
= (−1)i−j (k + j)n xi
i=0
i! j=0
j
n n (41)
X (k + j)n X xi
= (−1)i−j
j=0
j! i=j
(i − j)!
n n−j
X (k + j)n xj X (−x)i
=
j=0
j! i=0
i!

By (1):
n   Z ∞
e−x X n
Yn,k (x) = (k + j)n xj tn−j e−t dt
n! j=0 j −x
n   (42)
e−x ∞ X n
Z
= (k + j)n xj tn−j e−t dt
n! −x j=0 j
d
Suppose Dx is an operator in Dx = x dx , then, we can obtain Dxn (xi ) = in xi .
So:
n  
(−1)n x−k e−x ∞ X n
Z
Yn,k (x) = (−k − j)n xk+j t−k−j tn+k e−t dt
n! −x j=0 j
n  
(−1)n x−k e−x ∞ X n
Z
x
= Dtn (( )k+j )tn+k e−t dt (43)
n! −x j=0 j t
(−1)n x−k e−x ∞ n x k
Z
x
= Dt (( ) (1 + )n )tn+k e−t dt
n! −x t t
The operator in integral can be vanished with integration by part. Also, this
integral can be solved by Laplace transform or other transforms. For intended
polynomial we have:
n  
(−1)n e−x ∞ n
Z
X n k x
x = Tn (x) = Yn,0 (x) = Dt ((1 + )n )tn e−t dt (44)
k n! −x t
k=0

The same result can be obtained by Dobinski formula [5] as follows:

n
kn k (−1)n ∞ n
Z
X x
x = Dt ((1 + )n )tn e−t dt (45)
k! n! −x t
k=0

Also for Bell numbers [4]:

n
1 X kn (−1)n ∞
1
Z
Bn = = Dtn ((1 + )n )tn e−t dt (46)
e k! e.n! −1 t
k=0

8
Using bottom up unfolding for this recurrence on k and applying Yn,0 (x) =
Tn (x), we can easily observe that:
k  
1 X k
Yn,k (x) = k (−1)k−i Tn+i−1 (x) (47)
x i=0 i

Which is the inverse Stirling transform [3] of following formula:

k  
X k i
Tn+k−1 (x) = x Yn,i (x) (48)
i=0
i

Using the definition of Touchard polynomial and (39):

k X n    n+k k   
X k n+i i+j
X X k n+i
Tn+k−1 (x) = x = ( )xm (49)
i=0 j=0
i i+j i m=0 i=0
i m i

We get:
  X k   
n+k−1 k n+i
= (50)
m i=0
i m i

3.6. Solving ym,k (x) = kym−1,k (x) + (x + k)ym−1,k+1 (x)

First, we consider sn,k (x) = Γ(x + k)yn,k (x), s0,k (x) = Γ(x + k) and use
bottom-up unfolding as in 3.4:
n  
X n+k
sn,k (x) = Γ(x + k + i) (51)
i=0
i+k k

Now by the definition of gamma function we get:

n   Z ∞
X n+k
sn,k (x) = tx+k+i−1 e−t dt
i=0
i + k k 0
Z ∞X n  
n+k (52)
= tx+k+i−1 e−t dt
0 i=0
i + k k
Z ∞
= Tn,k (t)tx+k−1 e−t dt
0

Then, by sn,k (x) = Γ(x + k)yn,k (x) we get:

Z ∞
1
yn,k (x) = Tn,k (t)tx+k−1 e−t dt (53)
Γ(x + k) 0
Finally, we have the explicit formula of the following rising factorial polynomial:
n  
X n k
x = yn,0 (x)
k
k=0
Z ∞ (54)
1
= Tn (t)tx−1 e−t dt
Γ(x) 0

9
4. Rising and falling factorial transforms

In this section, we discuss the transformation of power polynomials into ris-

ing and falling factorial polynomials. These transforms justify the results that
we obtained in the previous section. First, we consider a rising factorial polyno-
mial and a power polynomial with the same coefficient as follows respectively:
n
X
Rn (x) = an,k xk (55)
k=0

n
X
Pn (x) = an,k xk (56)
k=0

We can apply the definition of rising factorial as below:

n
X Γ(x + k)
Rn (x) = an,k (57)
Γ(x)
k=0

As well as the definition of gamma function:

RF T (Pn (x)) = Rn (x)
n Z ∞
1 X
= an,k tx+k−1 e−t dt
Γ(x) 0
k=0

1
Z ∞ X n (58)
= ( an,k tk )tx−1 e−t dt
Γ(x) 0
k=0
Z ∞
1
= Pn (t)tx−1 e−t dt
Γ(x) 0
Here we encounter the fact that operation can be done on two distinct types
of polynomials (i.e. power and factorial polynomials). This is useful to derive
formulas
n for RF T (RF T (P
Pnn(x)))
nand RF T −1(RF T −1 (Pn (x))) utilizing xn =
P n n k
k=0 k (x)k and x = k=0 k x on (55) and (56), respectively:

n n Xk  
X X k
Pn (−x) = an,k (−1)k xk = an,k (−1)k (x)i
i=0
i
k=0 k=0
n n  
(59)
X
i
X k
= (−x) an,k (−1)k−i
i=0
i
k=i

RF T −1
(RF T −1
(Rn (x))) = RF T −1 (Pn (x))
n n  
X X k
= xi an,k (−1)k−i
i=0
i (60)
k=i
n
X
= an,k (−1)k Tk (−x)
k=0

10
 n n X k   n n  
X X k X X k
Rn (x) = an,k xk = an,k xi = xi an,k (61)
i i
k=0 k=0 i=0 i=0 k=i
n n  
X
i
X k
RF T (RF T (Pn(x))) = RF T (Rn (x)) = x an,k
i=0
i
k=i
n k   Z ∞ (62)
X X k i 1
= an,k ( x)= Rn (t)tx−1 e−t dt
i=0
i Γ(x) 0
k=0

Consider the following polynomial to which this form ∆k f (x) = (E − I)k f (x) =
Pk k

k−i
i=0 i (−1) f (x + i) and Cauchy multiplication are applied:
∞ ∞ k  
X ∆k Pn (−x) k X y k X k
y = (−1)k−i Pn (−x − i)
k! k! i=0 i
k=0 k=0
∞ k
X X (−1)k−i Pn (−x − i)
= yk (63)
i=0
(k − i)! i!
k=0
∞
X Pn (−x − k)
=e −y
yk
k!
k=0
k
This is generalization of Dobinski formula [5]. Also, we can use ∆ f (x)
k! =
P∞  i f (i) (x)
i=k k i! from [8] for above polynomial:
∞ ∞ ∞   (i)
X ∆k Pn (−x) k X k X i Pn (−x)
y = y
k! k i!
k=0 k=0 i=k
i
(64)
∞ (i)   ∞ (i)
X Pn (−x) X i k X Pn (−x)
= y = Ti (y)
i=0
i! k i=0
i!
k=0

Pn(i) (−x)
Consequently, by i! = an,k (−1)k we have:
x=0
n ∞ ∞
X X ∆k Pn (−x) X Pn (−k) k
an,k (−1)k Ti (y) = y k = e−y y (65)
i=0
k! k!
k=0 x=0 k=0

We can derive inverse rising factorial transform:

n
X
RF T −1 (Pn (x)) = an,k (−1)k Ti (−x)
i=0
∞
X ∆k Pn (−z)
= (−x)k (66)
k!
k=0 z=0
∞
X (−1)k Pn (−k) k
= ex x
k!
k=0

11
A power and a falling factorial polynomial can be represented as follows:
n
X
Pn (x) = an,k xk (67)
k=0
n
X
Fn (x) = an,k (x)k (68)
k=0
FFT
Pn (x) −−−→ Fn (x) (69)
By [9]:
1 1 et
Z
= dt (70)
Γ(z) 2πi C tz
Which C is any deformed anticlockwise Bromwich contour [9]. Using the defi-
nition of falling factorial:
Γ(x + 1) Γ(x + 1) et
Z
(x)n = = x−n+1
dt (71)
Γ(x − n + 1) 2πi C t

So, we get:
Γ(x + 1) Pn (t)et
Z
Fn (x) = dt (72)
2πi C tx+1
By [10]:

Fn (x) = F F T (Pn (x)) = ∂tx (Pn (t)et ) (73)

t=0
This formula demonstrate that falling factorial transform is equivalent to xth
derivative of Pn (t)et .
(58) is recall us fractional integral in (0, ∞). So we can rewrite (58) for RF T
as follows:
RF T (Pn (x)) = 0 D∞ −x
(Pn (t)et ) (74)
Also, There is another way to calculate RF T owing to the definition of Mellin
transform [10]:
1
RF T (Pn (x)) = M(Pn (t)e−t ) (75)
Γ(x)

5. Future works
The following suggestions are made for further studies:
1. Generalizing of this method: We proposed a method that was used to
solve finite summation with a restrictive parameter. Hence, infinite sum-
mations with more variables can be considered. Also, this method can be
generalized to multidimensional summation;
2. Solving more examples: More examples can be solved by this method,
especially polynomials (e.g., power or factorial polynomials);
3. Extending rising and falling factorial transforms.

12
6. Conclusion

We proposed a method that solves the summations that are finite and have
a restrictive parameter. We illustrated some examples from which some use-
ful identities emerge. We also investigated two new transforms that unify the
results obtained in the example section. In addition, they are suitable for the
transformation of the power series into the Newton series and vice versa. By
the inversion formula of both transforms, almost all series can be solved. We
can conclude from the results of section 4 that falling and rising factorial trans-
forms can be represented as fractional derivatives and integrals of the intended
function, respectively.

References

[1] M. Abramowitz, I. A. Stegun, Handbook of Mathematical Functions with

Formulas, Graphs, and Mathematical Tables, Ninth Printing. New York:
Dover, 1972.
[2] L. Comtet, Advanced Combinatorics: The Art of Finite and Infinite Ex-
pansions, rev. enl. ed. Dordrecht, Netherlands: Reidel, 1974.
[3] R. L. Graham, D. E. Knuth, and O. Patashnik, Concrete Mathematics,
Second Ed., Addison-Wesley, 1994.
[4] E. T. Bell, The Iterated exponential integers, Annals of Mathematics, Vol.
39, No. 3, July 1938.
[5] G. C. Rota, The number of partitions of a set, The American Mathematical
Monthly, Vol. 71, No. 5, May 1964.
[6] A. Z. Broder, The r-Stirling numbers, Discrete Mathematics, Vol. 49, May
1984.
[7] M. Mihoubi, H. Belbachir, Linear recurrences for r-Bell polynomials, Jour-
nal of Integer Sequences, Vol. 17, 2014.
[8] F. Olver, D. Lozier, R. Boisvert, C. Clark, NIST Handbook of Mathemat-
ical Functions, Cambridge University Press, New York, 2010.
[9] T. Schmelzer, L. N. Trefethen, Computing the gamma function using con-
tour integrals and Rational Approximation, SIAM Journal on Numerical
Analysis, Vol. 45, No. 2, pp. 558-571, January 2007.
[10] L. Debnath, D. Bhatta, Integral Transforms and Their Applications, Third
Ed., CRC Press, 2015.

13