Jca 21 08

J ournal of
Classical
A nalysis
Volume 21, Number 2 (2023), 87–152 doi:10.7153/jca-2023-21-08
RECURRENCE RELATIONS FOR THE MOMENTS OF

DISCRETE SEMICLASSICAL FUNCTIONALS OF CLASS s 2
D IEGO D OMINICI
Abstract. We study recurrence relations satisfied by the moments n (z) of discrete linear func-
tionals whose first moment satisfies a holonomic differential equation. We consider all cases
when the order of the ODE is less or equal than 3 .
1. Introduction
Let K be a field (we mostly think of K as R or C) and N0 be the set of nonneg-

ative integers
N0 = N ∪ {0} = {0, 1, 2, . . .} .
Let F = K [[z]] denote the ring of formal power series in the variable z


F=  cn zn : cn ∈ K ,
n=0
and  : F → F be the differential operator defined by [49, 16.8.2]
 = zz , (1)

where z is the derivative operator z = z. We use the notation
(x + c) = (x + c1 ) · · · (x + cm ) , c ∈ Km , m ∈ N,
and for m = 0 we understand that K0 = 0/ and
(x + 0)
/ = x.
We denote by k,n the Kronecker delta

1, k=n
k,n = , k, n ∈ N0 ,
0, k = n
Mathematics subject classification (2020): Primary 33C47; Secondary 33C20, 11B37.

Keywords and phrases: Orthogonal polynomials, moments, recurrences.
c , Zagreb 87
Paper JCA-21-08
88 D. D OMINICI
and say that {n }n0 ⊂ K[x] is a monic basis of K[x] if n is monic and deg (n ) = n
for all n ∈ N0 . The Pochhammer symbol (c)n is defined by [49, 5.2.4]
n−1
(c)0 = 1, (c)n =  (c + j), n ∈ N, c ∈ K,
j=0
and for c ∈ Km we will use the notation [49, 16.1]
(c)n = (c1 )n · · · (cm )n , (0)

/ n = 1, n ∈ N0 .
In this article, we continue the work started in [21], where we studied the moments
n = L p,q [n ] of linear functionals L p,q : K [x] → F (acting on the variable x) defined
by

(a)x zx
L p,q [u] =  u (x) , u ∈ K[x], (2)
x=0 (b + 1)x x!
with a ∈ K p and b ∈ Kq . It follows from (2) that the first moment 0 (z) = L [1] satisfies
a differential equation (in the variable z) with polynomial coefficients  p,q [y] = 0,
where the differential operator  p,q is defined by
 p,q =  ( + b) − z ( + a),
and we assume that x and z are independent variables.

The ODE  p,q [y] = 0 is the (generalized) hypergeometric differential equation
[49, 16.8.3] of order o = max {p, q + 1}, and the first moment 0 (z) can be represented
as
a
0 (z) = p Fq ;z ,
b+1
where the (generalized) hypergeometric function p Fq is defined by [49, 16.2.1],

a (a)x zx
p Fq ;z =  .
b x=0 (b)x x!
Functionals of the form (2) are called discrete semiclassical [24], [43], [47]. If we
define the polynomials
 (x) = x (x + b),  (x) = (x + a), (3)
then the class s of the functional L p,q is given by
s = max {deg ( ) − 1, deg( ) − 1} = o − 1, (4)
and semiclassical functionals of class s = 0 are called classical. In [22], we classified

all discrete semiclassical linear functionals of class s 1 (see also [45], [52], [53]). We
extended our results in [23] to the class s = 2.
In this paper, we will find recurrence relations for the moments of all discrete
semiclassical linear functionals of class s 2.
R ECURRENCE RELATIONS FOR THE MOMENTS 89
2. Previous results
In this section, we give a brief description of the results from our previous work
that will be needed in this paper. For more details and proofs, we refer the reader to
[21].
Since the operator  defined in (1) satisfies
 [zx f ] = zx ( + x) [ f ] , f ∈ F,
it follows using linearity that for all u ∈ K [x] ,
u ( ) [zx f ] = zx u ( + x)[ f ] , f ∈ F. (5)
Let L : K[x] → F be a discrete functional


L [u] =  u (x)  (x) zx , u ∈ K[x], (6)
x=0
where  : N0 → K is a given function. If we set f = 1 in (5), we get
u ( ) [zx ] = u (x) zx , u ∈ K [x]
and therefore
L [u] = u ( ) L [1], u ∈ K [x] . (7)
We conclude from (7) that the moments n (z) of L on any monic basis {n }n0 are
completely determined by the first moment
n = L [n ] = n ( ) L [1] = n ( ) [0 ] , n ∈ N0 . (8)
2.1. Newtonian bases

A convenient choice for {n }n0 is the basis of Newton polynomials [60], [61]
n−1
n (x) =  (x − k ) , n ∈ N, 0 (x) = 1,
k=0
where {k }k0 ⊂ K is a fixed sequence. These polynomials satisfy the 2 -term recur-
rence relation
xn (x) = n+1 (x) + n n (x) . (9)
Setting x =  in (9), multiplying by 0 (z) and using (8), we obtain.
 [n ] = (S + n ) [n ] , (10)
where S denotes the shift operator in n
S [cn ] = cn+1 . (11)

90 D. D OMINICI
R EMARK 1. From (10) and (11) we see that
 [n+1 ] = n+2 + n+1n+1 = S [n+1 + n n ] = S (S + n ) [n ] , (12)
which is different from
(S + n ) S [n ] = (S + n ) [n+1 ] = n+2 + n n+1,
so caution must be exercised when n depends on n .

The Newton polynomials n (x) satisfy the change of bases formula
n
n
xn =  i (x) , (13)
i=0 i

where the coefficients ni are defined by the recurrence

n+1 n n n
= + i , = 1, (14)
i i−1 i n
with boundary conditions
n
= 0, i∈
/ [0, n].
i
Among all Newtonian bases, we will consider the monomial basis ( k = 0) n (x) =
xn and the basis of falling factorial polynomials ( k = k) n (x) = n (x) . The main
reason for choosing the polynomials n (x) is that from their definition
n−1
n (x) =  (x − k) , n ∈ N, (15)
k=0
it follows that
n+1 (x) = (x − n) n (x) = xn (x − 1), n 0, (16)

and therefore n (x) is well suited for dealing with shifts in x . We will call the moments
of L p,q on the monomial basis standard moments
n (z) = L p,q [xn ] , n ∈ N0 ,
and the moments of L p,q on the falling factorial polynomials basis modified moments
n (z) = L p,q [n ] , n ∈ N0 .
Since
zn zn [zx ] = zn n (x) zx−n = n (x) zx ,
we have n (z) = zn zn [0 ] and using the formula [49, 16.3.1]

a (a)n a+n
zn p Fq ;z = p Fq ;z ,
b+1 (b + 1)n b+n+1
we obtain the hypergeometric representation

n (a)n a+n
n (z) = z p Fq ;z , n ∈ N0 .
(b + 1)n b+n+1
Multiplying the ODE  p,q [y] = 0 by  n and using (5), we get
 n+1 ( + b) − z ( + 1)n ( + a) [0 ] = 0. (17)
and using (10) with n = 0 in (17) we have  p,q [0 ] = 0, where the standard moments
recurrence operator  p,q is defined by
 p,q = S n+1 (S + b) − z (S + 1)n (S + a). (18)
The polynomials (x + c) can be written in the monomial basis as

m
(x + c) =  em−k (c) xk , c ∈ Km , (19)
k=0
where the elementary symmetric polynomials en (c) are defined by the generating func-
tion [41]
 m
 en (c)t n =  (1 + tci) , c ∈ Km .
n=0 i=1
Using (19), we can rewrite  p,q in extended form
q n p
n
 p,q =  eq−k (b) S n+k+1 − z   e p− j (a) S k+ j ,
k=0 k=0 k j=0
and the equation  p,q [0 ] = 0 gives a recurrence for the standard moments
q n p
n
 eq−k (b) n+k+1 − z  k  e p− j (a) k+ j = 0
k=0 k=0 j=0
of order n + s + 1, where s is the class of the functional L p,q defined in (4).

Similarly, multiplying the ODE  p,q [y] = 0 by n ( − 1) and using (5) and (16),
we get
[( + b) n+1 ( ) − z ( + a) n ( )] [0 ] = 0.
Using (8) and (12), we conclude that  p,q [n ] = 0, where the modified moments recur-
rence operator  p,q is defined by
 p,q = S (S + n + b) − z (S + n + a). (20)

n
When k = k, the coefficients Sn,i = i defined by (14) are called Stirling num-
bers of the second kind [49, 26.8]. Using (13) in (19), we obtain
m k
(x + c) =  em−k (c)  Sk,i i (x) ,
k=0 i=0
92 D. D OMINICI
and from (12) we conclude that

m k
(S + n + c) =  em−k (c + n)  Sk,i S i . (21)
k=0 i=0
Using (21), we can rewrite  p,q in extended form

q k p k
 p,q =  em−k (b + n + 1)  Sk,i S i+1 − z  e p−k (a + n)  Sk,i S i ,
k=0 i=0 k=0 i=0
and the equation  p,q [n ] = 0 gives a recurrence for the modified moments
q k p k
 em−k (b + n + 1)  Sk,i n+i+1 − z  e p−k (a + n)  Sk,i n+i = 0,
k=0 i=0 k=0 i=0
of minimal order s + 1, where s is the class of the functional L p,q defined in (4).
2.2. Transformations
There are 4 canonical transformations of the functional L p,q :
1) The Christoffel transformation at  , which we define by

(a) zx
LCp,q [u] =  (x −  )u (x) (b + 1)
x
,  ∈ K, u ∈ K[x]. (22)
x=0 x x!
Using (9) and (22) we obtain
nC = LCp,q [n ] = L p,q [(x −  )n ] = n+1 + (n −  ) n ,
and in particular
nC = n+1 + (n −  ) n . (23)
Since from (15) it follows that
(− + 1)x
x −  = − , (24)
(− )x
the first moment has the hypergeometric representation

a, − + 1
0C (z;  ) = − p+1 Fq+1 ;z .
b + 1, −
Thus, 0C is a solution of the ODE Cp,q [y] = 0, where the differential operator Cp,q is
defined by
Cp,q =  ( ) ( −  − 1) − z ( ) ( −  + 1), (25)
and the polynomials  ,  where defined in (3).
From (18) and (25) we see that Cp,q 0C = 0, where the recurrence operator Cp,q
is defined by
Cp,q = S n+1 (S + b)(S −  − 1) − z (S + 1)n (S + a)(S −  + 1).
Similarly, (20) and (25) give Cp,q nC = 0, with
Cp,q = S (S + n + b)(S + n −  − 1) − z (S + n + a)(S + n −  + 1).

2) The Geronimus transformation at  , which we define by

u (x) (a) zx
LGp,q [u] =  x −  (b + 1)
x
,  ∈ K \ N0 , u ∈ K[x]. (26)
x=0 x x!
Using (24), we see that the first moment has the hypergeometric representation

G −1 a, −
0 (z;  ) = − p+1 Fq+1 ;z ,
b + 1, − + 1
and therefore is a solution of the ODE Gp,q [y] = 0, where the differential operator Gp,q
is defined by
Gp,q =  ( ) ( −  ) − z ( ) ( −  ) =  p,q ( −  ). (27)
From (9) and (26), we see that
G
n+1 + (n −  ) nG = LGp,q [(x −  )n ] = L p,q [n ] = n ,
and in particular
G
n+1 + (n −  ) nG = n . (28)
From (18) and (27) we see that Gp,q 0G = 0, where the recurrence operator Gp,q
is defined by
Gp,q =  p,q (S −  ) .
Similarly, (20) and (27) give Gp,q nG = 0, with
Gp,q =  p,q (S + n −  ).
3) The Uvarov transformation at  , which we define by
LUp,q [u] = L p,q [u] +  u ( ) z ,  ∈ K, u ∈ K [x] .
The differential operator Up,q is defined by
Up,q = ( −  ) ( −  − 1) p,q, (29)

since
Up,q =  ( ) ( −  )( −  − 1) − z ( ) ( −  + 1)( −  )
= [ ( ) ( −  − 1) − z ( ) ( −  + 1)] ( −  ),
94 D. D OMINICI
and therefore
Up,q [0 ] = 0, Up,q [z ] = 0.
Comparing (29) with (25) and (27), we see that
Up,q = G,C C,G

p,q =  p,q
in the sense of applying a double transformation to the operator  p,q .

From (18) and (29) we see that Up,q 0U = 0, where the recurrence operator
U
 p,q is defined by
Up,q = (S −  ) (S −  − 1) p,q .
Similarly, (20) and (29) give Up,q nU = 0, with
Up,q = (S + n −  )(S + n −  − 1) p,q .
R EMARK 2. Because  −  annihilates the function z , we have Gp,q [z ] = 0

and could have defined

u (x) (a) zx
LGp,q [u] =  x −  (b + 1)
x
x!
+  u ( ) z ,  ∈ K,
x=0 x
as some authors do. Thus, we will not consider the double transformation U,G
p,q , since
we can simplify the composition
U,G G
p,q =  p,q .
If  ( ) = 0, then the differential operator

U( )
 p,q = ( −  − 1) p,q
is called a reduced-Uvarov transformation, since

U( )
 p,q =  ( ) ( −  − 1) − z ( ) ( −  ) = [ ( ) ( −  − 1) − z ( )] ( −  ) ,

U( )
with  ( ) = ( −  )  ( ) , and therefore  p,q z = 0. In this case, we have
U(  )
 p,q = (S −  − 1) p,q ,
and
U( )
 p,q = (S + n −  ) p,q .
The second possibility for a reduced-Uvarov transformation happens when  ( ) =
0. We now have
U( )
 p,q = ( −  )  p,q ,
U( )
 p,q = (S −  )  p,q ,
and
U( )
 p,q = (S + n −  − 1) p,q .
4) The truncation transformation at N , which we define by
N
(a) zx
LTp,q [u] =  u (x) (b + 1)
x
, N ∈ N0 , u ∈ K[x].
x=0 x x!
The first moment admits the hypergeometric representation

T a, −N
0 (z) = p+1 Fq+1 ;z ,
b + 1, −N
and therefore is a solution of the ODE Tp,q [y] = 0, where the differential operator Tp,q
is defined by
Tp,q =  ( ) ( − N − 1) − z ( ) ( − N) = ( − N − 1) p,q . (30)
From (18) and (30) we see that Tp,q 0T = 0, where the recurrence operator Tp,q
is defined by
Tp,q = (S − N − 1) p,q .
Similarly, (20) and (30) give Tp,q nT = 0, with
Tp,q = (S + n − N) p,q .
3. Examples
We now illustrate the application of the formulas that we have derived in the previ-
ous section. We will consider all discrete semiclassical functionals of class s 2, and
also look at the subclasses obtained by applying one or more of the canonical transfor-
mations of the functional L p,q .
3.1. Functionals of class 0 (discrete classical functionals)
The discrete classical orthogonal polynomials (Charlier, Meixner, Krawtchouk)

first appeared in the literature in the years 1905–1934, and were considered at the time
as a generalization of the continuous classical polynomials (Hermite, Laguerre, Jacobi).
The last member of this class (Hahn polynomials) were introduced by Chebyshev
(1875) and Hahn (1949), but we don’t consider them by themselves since they are a
special case (z = 1) of the Generalized Hahn polynomials (see Section 3.2.4).
We will use the notation (p, q; N) to indicate that one of the upper parameters in
the hypergeometric representation of the first moment is a negative integer −N, N ∈ N.
For additional references, see [1], [13], [18], [19], [32], [48], [50], [59].
96 D. D OMINICI
3.1.1. Functional of type (0, 0) (Charlier)

The Charlier polynomials were introduced by Carl Vilhelm Ludwig Charlier (1862–
1934) in his paper [14].
Linear functional

zx
L0,0 [u] =  u (x) .
x=0 x!
First moment differential operator
0,0 =  − z.
Standard moments recurrence operator
0,0 = S n+1 − z (S + 1)n .
Modified moments hypergeometric representation

−
n (z) = zn 0 F0 ; z = zn ez . (31)
−
Modified moments recurrence operator
0,0 = S − z.
R EMARK 3. The Charlier polynomials have the hypergeometric representation

[49, 18.20.8]
(0,0) −n, −x −1
Pn (x; z) = 2 F0 ; −z . (32)
−
3.1.2. Functional of type (1, 0) (Meixner)

The Meixner polynomials were introduced by Josef Meixner (1908–1994) in his
paper [46], although Ladislav Truksa (1891–?) already considered them in his 1931
papers [55], [56], [57] (see [12]).
Linear functional

zx
L1,0 [u] =  u (x) (a)x , z = 1.
x=0 x!
1,0 =  − z ( + a).
1,0 = S n+1 − z (S + 1)n (S + a).
a+n
n (z) = zn (a)n 1 F0 ; z = zn (a)n (1 − z)−a−n . (33)
−
1,0 = S − z (S + n + a).
3.1.2.1. Functional of type (1, 0; N) (Krawtchouk)
The Krawtchouk polynomials were introduced by Mykhailo Pylypovych Kravchuk

(1892–1942) in his paper [37]. These polynomials are a particular case of the Meixner
polynomials, with −a = N ∈ N.
Linear functional
N
zx
L1,0;N [u] =  u (x) (−N)x x! , z = 1.
x=0
1,0;N =  − z ( − N) .
Standard moments recurrence
1,0;N = S n+1 − z (S + 1)n (S − N).
Modified moments
n (z) = zn (−N)n (1 − z)N−n .
1,0;N = S − z (S + n − N).
3.2. Functionals of class 1

In [22] and [23], we classified the discrete semiclassical functionals of class s = 1.
There are 4 main families and 9 subfamilies, obtained by applying transformations to
the Charlier and Meixner functionals.
For additional references, see [2], [3], [6], [9], [11], [15], [27], [31], [34], [39],
[40], [44], [51].
3.2.1. Functional of type (0, 1) (Generalized Charlier)

Linear functional

1 zx
L0,1 [u] =  u (x) (b + 1)x x!
.
x=0
0,1 =  ( + b) − z.
0,1 = S n+1 (S + b) − z (S + 1)n .

98 D. D OMINICI
zn −
n (z) = 0 F1 ;z
(b + 1)n b+1+n
n−b √
= z 2  (b + 1)Ib+n 2 z ,
where Iv (z) denotes the modified Bessel function of the first kind [49, 10.25.2].
0,1 = S (S + n + b)− z = S 2 + (n + 1 + b)S − z. (34)
R EMARK 4. If we write
n = An pn ,
then the recurrence (34) becomes
(n + b) z
pn+1 + pn − 2 pn−1 = 0.
A A
Choosing
1 z
= −2x, − = 1,
A A2
we get
pn+1 − 2 (n + b)xpn + pn−1 = 0. (35)
The orthogonal polynomials satisfying the 3-term recurrence relation (35) with initial
conditions
p0 = 1, p1 = 2bx
are the modified Lommel polynomials having the hypergeometric representation

− n2 , − n−1
pn (x) = (b)n (2x)n 2 F3 2 ; −x−2 .
b, −n, 1 − b − n
See [17], [28], [38], [42] .

Another possibility is to define
n = (−1)n qn ,
where the monic polynomials qn (b) satisfy the 3 -term recurrence relation
bqn = qn+1 − nqn + zqn−1, q−1 = 0, q0 = 1.
For additional references on the generalized Charlier polynomials, see [16], [35],
[54], [58].
3.2.2. Functional of type (1, 1) (Generalized Meixner)

Linear functional

(a)x zx
L1,1 [u] =  u (x) (b + 1)x x!
.
x=0
1,1 =  ( + b) − z ( + a). (36)
1,1 = S n+1 (S + b) − z (S + 1)n (S + a).
(a)n a+n
n (z) = zn 1 F1 ;z .
(b + 1)n b+1+n
1,1 = S (S + n + b)− z (S + n + a) = S 2 + (n + 1 + b − z)S − z (n + a).
R EMARK 5. If we define
n = (−1)n pn ,
then the monic polynomials pn (b) satisfy the 3 -term recurrence relation
bpn = pn+1 − (n − z) pn + z (n + a − 1) pn−1 .
For additional references on the generalized Meixner polynomials, see [10], [16],
[29].
3.2.2.1. Christoffel Charlier functional

Linear functional

zx
LC0,0 [u] =  (x −  )u (x) x!
.
x=0
C0,0 =  ( −  − 1) − z ( −  + 1),
which is a special case of (36) with
a = − + 1, b = − − 1.
C0,0 = S n+1 (S −  − 1) − z (S + 1)n (S −  + 1).

100 D. D OMINICI

n−+1
nC (z) = (n −  )zn 1 F1 ;z .
n−
Using Kummer’s transformation [49, 13.2.39]

a z b−a
1 F1 ; z = e 1 F1 ; −z ,
b b
we get

−1
nC (z;  ) n z
= (n −  )z e 1 F1 ; −z
n−

z
= (n −  )zn ez 1 + ,
n−
in agreement with (23), since
nC = n+1 + (n −  ) n = (z −  + n)zn ez . (37)

C0,0 = S (S + n −  − 1) − z (S + n −  + 1)
(38)
= S 2 + (n −  − z) S − z (n −  + 1).
R EMARK 6. Using (37), we see that the modified moments satisfy the first order
C [ ] = 0, with
recurrence 0,0
C
0,0 = (n −  + z) S − z (n −  + 1 + z).
This agrees with (38), since

C
(S + n + 1 −  ) 0,0 = (z + n + 1 −  )C0,0 .
3.2.2.2. Geronimus Charlier functional

Linear functional

u (x) zx
LG
0,0 [u] =  , ∈
/ N0 .
x=0 x −  x!
G G
0,0 = 0,0 ( −  ) =  ( −  ) − z ( −  ),
a = − , b = − .
G
0,0 = 0,0 (S −  ) = S
n+1
− z (S + 1)n (S −  ).

zn n−
nG (z;  ) = F
1 1 ; z .
n− n− +1
Using the identity [49, 13.6.5]

a
1 F1 ; −z = az−a  (a, z) ,
a+1
where  (a, z) is the incomplete gamma function defined by [49, 8.2.1]

z
 (a, z) = t a−1 e−t dt, Re (a) > 0,
0
we obtain
nG (z;  ) = (−1)n (−z)  (n −  , −z).
Since the function  (a, z) satisfies the recurrence [49, 8.8.1]
 (a + 1, z) = a (a, z) − za e−z ,
we see that
G
n+1 + (n −  ) nG = zn ez = n ,
in agreement with (28).
G 2
0,0 = 0,0 (S + n −  ) = S + (n + 1 −  − z) S − z (n −  ).
3.2.2.3. Reduced-Uvarov Charlier functional

Since for the Charlier functional
 ( ) =  ,
we will have a reduced Uvarov transformation U ( ) for it if  = 0.

Linear functional

zx
 u (x)
U(0)
L0,0 [u] = +  u (0) .
x=0 x!

U(0)
0,0 = ( − 1)0,0 =  ( − 1) − z ,
102 D. D OMINICI
a = 0, b = −1.

U(0)
0,0 = (S − 1)0,0 = S n+1 (S − 1) − z (S + 1)n S .

U(0)
0,0 = (S + n)0,0 = S 2 + (n − z)S − nz.
For additional references, see [7], [26].
3.2.2.4. Truncated Charlier functional

Linear functional
N
zx
LT0,0 [u] =  u (x) x!
, N ∈ N0 .
x=0
First moment
N
zx
0T (z) =  x!
= N (z) ,
x=0
where N (z) denotes the truncated exponential series [49, 8.4.11].
T0,0 = ( − N − 1)0,0 =  ( − N − 1) − z ( − N) ,
a = −N, b = −N − 1.
T0,0 = (S − N − 1)0,0 = S n+1 (S − N − 1) − z (S + 1)n (S − N) .

n−N zn
nT (z) = zn 1 F1 ;z = U (n − N, n − N, z),
n−N (N − n)!
where U (a, b, z) is Tricomi’s function [49, 13.2.6]. Using the identity [49, 13.6.6]
U (a, a, z) = ez  (1 − a, z),
where  (a, z) is the incomplete gamma function defined by [49, 8.6.5]


 (a, z) = t a−1 et dt,
z
we conclude that
zn ez
nT (z) =  (N − n + 1, z).
(N − n)!
Comparing with (31), we see that
 (N − n + 1, z)
nT (z) = n (z) , 0 n N.
(N − n)!
T0,0 = (S + n − N)0,0 = S 2 + (n − N − z) S − (n − N)z.
For additional references, see [33].
3.2.3. Functional of type (2, 0; N) (Generalized Krawtchouk)

Linear functional
N
zx
L2,0;N [u] =  u (x) (−N)x (a)x x! , N ∈ N0 .
x=0
2,0;N =  − z ( − N) ( + a).
2,0;N = S n+1 − z (S + 1)n (S − N) (S + a).
n − N, a + n
n (z) = zn (−N)n (a)n 2 F0 ;z .
−
(0,0)
Using the hypergeometric representation (32) of the Charlier polynomials Pn (x; z) ,
we can write
(0,0)
n (z) = zn (−N)n (a)n PN−n −a − n; −z−1 .
2,0;N = S − z (S + n − N)(S + n + a)
= −zS 2 + [1 − z (2n + 1 − N + a)] S + z (N − n)(n + a).
R EMARK 7. If we set z−1 = x, we see that the modified moments are a family of
monic orthogonal polynomials pn (x) , satisfying the 3-term recurrence relation
xpn = pn+1 + (2n − 1 − N + a) pn + (n + a − 1)(n − N − 1) pn−1 .

104 D. D OMINICI
3.2.4. Functional of type (2, 1) (Generalized Hahn of type I)

Linear functional

(a1 )x (a2 )x zx
L2,1 [u] =  u (x) (b + 1)x x!
.
x=0

2,1 =  ( + b) − z ( + a1 ) ( + a2) . (39)
2,1 = S n+1 (S + b) − z (S + 1)n (S + a1 ) (S + a2) .
(a1 )n (a2 )n a1 + n, a2 + n
n (z) = zn 2 F1 ;z . (40)
(b + 1)n b+1+n
2,1 = S (S + n + b)− z (S + n + a1) (S + n + a2) (41)
2
= (1 − z)S + [b + n + 1 − z (2n + 1 + a1 + a2)] S − z (n + a1) (n + a2) .
R EMARK 8. If we set b = −x, we see that the modified moments are a family of
orthogonal polynomials pn (x) , satisfying the 3-term recurrence relation
xpn = (1 − z) pn+1 + [n − z (2n − 1 + a1 + a2 )] pn
+ z (n − 1 + a1) (n − 1 + a2) pn−1 .
3.2.4.1. Hahn functional

When z = 1, the generalized Hahn functional of type I becomes the Hahn func-
tional [36]. Note that in this case (40) can be reduced using the identity [49, 15.4.20]
a1 , a2 (b − a2)−a1
2 F1 ;1 = , Re (b − a2) > Re (a1 ) .
b (b)−a1
Choosing a1 = −N, N ∈ N, a2 = a, we get
(−N)n (a)n (b + 1 − a)N−n
n (1) = ,
(b + 1)n (b + 1 + n)N−n
or
(b + 1 − a)N (−N)n (a)n
n (1) = (−1)n ,
(b + 1)N (a − b − N)n
which is a solution of (41) when z = 1, a1 = −N, and a2 = a
[(b − a + N − n)S − (n − N)(n + a)] [n (1)] = 0.
For additional references, see [20], [30].
3.2.4.2. Christoffel Meixner functional

Linear functional

zx
LC1,0 [u] =  (x −  )u (x) (a)x x! .
x=0

C1,0 =  ( −  − 1) − z ( + a)( −  + 1),
a1 = a, a2 = − + 1, b = − − 1.
C1,0 = S n+1 (S −  − 1) − z (S + 1)n (S + a)(S −  + 1).

n + a, n −  + 1
nC (z) = (n −  )(a)n zn 2 F1 ;z .
n−
Using the identity

a, b c−a−b c − a, c − b
2 F1 ; z = (1 − z) 2 F1 ;z ,
c c
we get

−a−n−1 − − a, −1
nC (z;  ) n
= (n −  )(a)n z (1 − z) 2 F1 ;z
n−
(a +  )z
= (n −  )(a)n zn (1 − z)−a−n−1 1 + ,
n−
in agreement with (23), since
nC = n+1 + (n −  ) n = (z + az + n −  )(1 − z)−a−n−1 (a)n zn . (42)
C1,0 = S (S + n −  − 1) − z (S + n + a)(S + n −  + 1)
(43)
= (1 − z)S 2 + [n −  − z (2n + 2 + a −  )] S − z (n + 1 −  )(n + a).
R EMARK 9. From (42), we see that the modified moments satisfy the first order
C [ ] = 0, where
recurrence 1,0
C
1,0 = (1 − z)(n −  + z + az)S − z (n + a)(n + 1 −  + z + az).
This agrees with the second order recurrence (43), since
C
(S + n + 1 −  ) 1,0 = (n + 1 −  + z + az)C1,0 .
106 D. D OMINICI
3.2.4.3. Geronimus Meixner functional

Linear functional

u (x) zx
LG
1,0 [u] =  x− (a)x
x!
, ∈
/ N0 .
x=0
G
1,0 = 1,0 ( −  ) =  ( −  ) − z ( + a)( −  ),
a1 = a, a2 = − , b = − .
G
1,0 = 1,0 (S −  ) = S
n+1
− z (S + 1)n (S + a) (S −  ).
zn a + n, n − 
nG (z;  ) = (a)n 2 F1 ;z .
n− n−+1
Using the identity [49, 8.17.7]
a, 1 − b
2 F1 ; z = az−a Bz (a, b),
a+1
we conclude that
nG (z;  ) = (a)n z Bz (n −  , 1 − a − n),
where Bz (a, b) is the incomplete beta function defined by [49, 8.17.1]
1
Bz (a, b) = z a
t a−1 (1 − zt)b−1 dt. (44)
0
Since the function Bz (a, b) satisfies the recurrence [49, 8.17(iv)]
aBz (a, b + 1) − bBz (a + 1, b) = za (1 − z)b ,
we see that
G
n+1 + (n −  ) nG = zn (a)n (1 − z)−a−n = n ,
in agreement with (28).
G
1,0 = 1,0 (S + n −  )
= (1 − z)S 2 + [n −  + 1 − z (1 + a −  + 2n)] S − z (n −  )(n + a).
3.2.4.4. Reduced-Uvarov Meixner functional

Since for the Meixner functional we have
 ( ) =  ,  ( ) =  + a,
we will have reduced cases for its Uvarov transformation U ( ) if
 = 0, −a.
i)  = 0 Linear functional

zx
 u (x) (a)x
U(0)
L1,0 [u] = +  u (0) .
x=0 x!
U(0)
1,0 = ( − 1)1,0 =  ( − 1) − z ( + a)  ,
a1 = a, a2 = 0, b = −1.
U(0)
1,0 = (S − 1)1,0 = S n+1 (S − 1) − z (S + 1)n (S + a)S .
U(0)
1,0 = (S + n)1,0 = (1 − z)S 2 + [n − z (2n + 1 + a)] S − zn (n + a).
For additional references, see [4], [8], [25].
ii)  = −a Linear functional


zx
 u (x) (a)x
U(−a)
L1,0 [u] = +  u (−a) z−a .
x=0 x!
U(−a)
1,0 = ( + a)1,0 =  ( + a) − z ( + a)( + a + 1),
a1 = a, a2 = a + 1, b = a.
U(−a)
1,0 = (S + a)1,0 = S n+1 − z (S + 1)n (S + a + 1) (S + a).
Modified moments’ recurrence operator
U(−a)
1,0 = (S + n + a + 1)1,0
= (1 − z)S 2 − (n + a + 1)[(2z − 1)S + z (n + a)] .
108 D. D OMINICI
3.2.4.5. Truncated Meixner functional

Linear functional
N
zx
LT1,0 [u] =  u (x) (a)x x!
, N ∈ N0 .
x=0

T1,0 = ( − N − 1)1,0 =  ( − N − 1) − z ( + a)( − N) ,

a1 = a, a2 = −N, b = −N − 1.

T1,0 = (S − N − 1)1,0 = S n+1 (S − N − 1) − z (S + 1)n (S + a)(S − N) .

n − N, n + a
nT (z) = (a)n zn 2 F1 ;z .
n−N
Using the transformation [49, 15.8.7]

−N, b (c − b)N −N, b
F
2 1 ; z = F
2 1 ; 1 − z ,
c (c)N b−c−N+1
we obtain

n (−N − a)N−n n + a, n − N
nT (z) = (a)n z 2 F1 ;1 − z .
(n − N)N−n n+a+1
Since the incomplete beta function (44) has the hypergeometric representation [49,
8.17.7]
za 1 − b, a
Bz (a, b) = 2 F1 ;z ,
a a+1
we conclude that
(a)N+1 n
nT (z) = z (1 − z)−a−n B1−z (a + n, N − n + 1),
(N − n)!
and comparing with (33), we see that
(a + n)N−n+1
nT (z) = B1−z (a + n, N − n + 1) n (z) .
(N − n)!
T1,0 = (S + n − N)1,0
= (1 − z)S 2 + [n − N − z (2n + 1 − N + a)] S − z (n − N)(n + a).
3.3. Functionals of class 2

In [23], we classified the discrete semiclassical functionals of class s = 2. There
are 6 main families and 58 subfamilies, obtained by applying transformations to the
functionals of class s = 1 , or double transformations to the functionals of class s = 0.
3.3.1. Functional of type (0,2)

Linear functional

1 zx
L0,2 [u] =  u (x) (b1 + 1)x (b2 + 1)x x!
.
x=0

0,2 =  ( + b1) ( + b2 ) − z.
0,2 = S n+1 (S + b1) (S + b2 ) − z (S + 1)n = 0.

zn −
n (z) = 0 F2 ;z .
(b1 + 1)n (b2 + 1)n b1 + 1 + n, b2 + 1 + n
0,2 = S (S + n + b1) (S + n + b2) − z.

Linear functional

(a)x zx
L1,2 [u] =  u (x) (b1 + 1)x (b2 + 1)x x!
.
x=0
1,2 =  ( + b1) ( + b2 ) − z ( + a). (45)

1,2 = S n+1 (S + b1 ) (S + b2) − z (S + 1)n (S + a).
zn (a)n a+n
n (z) = 1 F2 ;z .
(b1 + 1)n (b2 + 1)n b 1 + 1 + n, b2 + 1 + n
1,2 = S (S + n + 1 + b1) (S + n + 1 + b2) − z (S + n + a).
110 D. D OMINICI
3.3.2.1. Christoffel Generalized Charlier functional

Linear functional

1 zx
LC0,1 [u] =  (x −  )u (x) (b + 1)x x!
.
x=0

C0,1 =  ( + b)( −  − 1) − z ( −  + 1),
a = − + 1, b1 = b, b2 = − − 1.
C0,1 = S n+1 (S + b)(S −  − 1) − z (S + 1)n (S −  + 1).

n− n n−+1
nC (z) = z 1 F2 ;z .
(b + 1)n b + 1 + n, n − 
C0,1 = S (S + n + b)(S + n −  − 1) − z (S + n −  + 1).
3.3.2.2. Geronimus Generalized Charlier functional

Linear functional

u (x) 1 zx
LG
0,1 [u] =  x− (b + 1)x x!
, ∈
/ N0 .
x=0

G
0,1 = 0,1 ( −  ) =  ( + b)( −  ) − z ( −  ),

a = − , b1 = b, b2 = − .
G
0,1 = 0,1 (S −  ) = S
n+1
(S + b) − z (S + 1)n (S −  ).

1 zn n−
nG (z) = 1 F2 ;z .
n −  (b + 1)n b + 1 + n, n −  + 1
G
0,1 = 0,1 (S + n −  ) = [S (S + n + b)− z] (S + n −  ).
3.3.2.3. Reduced-Uvarov Generalized Charlier functional

Since for the Generalized Charlier functional we have
 ( ) =  ( + b),
 = 0, −b.
i)  = 0 Linear functional

1 zx
 u (x)
U(0)
L0,1 [u] = +  u (0) .
x=0 (b + 1)x x!

U(0)
0,1 = ( − 1)0,1 =  ( + b)( − 1) − z ,
a = 0, b1 = b, b2 = −1.
U(0)
0,1 = (S − 1)0,1 = S n+1 (S + b)(S − 1) − z (S + 1)n S .
U(0)
0,1 = (S + n)0,1 = S (S + n + b)(S + n − 1)− z (S + n).
ii)  = −b Linear functional


1 zx
 u (x)
U(−b)
L0,1 [u] = +  u (−b) .
x=0 (b + 1)x x!

U(−b)
0,1 = ( + b − 1)0,1 =  ( + b)( + b − 1)− z ( + b),
a = b, b1 = b, b2 = b − 1.

U(−b)
0,1 = (S + b − 1)0,1 = S n+1 (S + b − 1)− z (S + 1)n (S + b).
U(−b)
0,1 = (S + n + b)0,1 = [S (S + n + b − 1)− z] (S + n + b).
112 D. D OMINICI
3.3.2.4. Truncated Generalized Charlier functional

Linear functional
N
1 zx
LT0,1 [u] =  u (x) (b + 1)x x!
, N ∈ N0 .
x=0
T0,1 = ( − N − 1)0,1 =  ( + b)( − N − 1) − z ( − N) ,
a = −N, b1 = b, b2 = −N − 1.
T0,1 = (S − N − 1)0,1 = S n+1 (S + b)(S − N − 1) − z (S + 1)n (S − N) .
T0,1 = (S + n − N)0,1 = S (S + n + b)(S + n − N − 1) − z (S + n − N).

Linear functional

(a1 )x (a2 )x zx
L2,2 [u] =  u (x) (b1 + 1)x (b2 + 1)x x!
.
x=0
2,2 =  ( + b1 ) ( + b2) − z ( + a1 ) ( + a2) . (46)

2,2 = S n+1 (S + b1 ) (S + b2) − z (S + 1)n (S + a1 ) (S + a2) .
zn (a1 )n (a2 )n a1 + n, a2 + n
n (z) = 2 F2 ;z .
(b1 + 1)n (b2 + 1)n b1 + 1 + n, b2 + 1 + n
2,2 = S (S + n + b1) (S + n + b2) − z (S + n + a1) (S + n + a2) .

3.3.3.1. Uvarov Charlier functional

Linear functional

zx
LU0,0 [u] =  u (x) x!
+  u ( ) z ,  = 0.
x=0
U0,0 = ( −  ) ( −  − 1)0,0
=  ( −  ) ( −  − 1) − z ( −  + 1)( −  ) ,
a1 = − + 1, a2 = − , b1 = − , b2 = − − 1.
U0,0 = (S −  )(S −  − 1)0,0

(47)
= S n+1 (S −  − 1) − z (S + 1)n (S −  + 1) (S −  ).
U0,0 = (S + n −  )(S + n −  + 1)0,0

(48)
= [S (S + n −  − 1) − z (S + n −  + 1)] (S + n −  ).
3.3.3.2. Double Christoffel Charlier functional

Linear functional

zx
 (x − 1) (x − 2) u (x)
2
LC0,0 [u] = .
x=0 x!

2
C0,0 =  ( − 1 − 1)( − 2 − 1) − z ( − 1 + 1)( − 2 + 1),
a1 = −1 + 1, a2 = −2 + 1, b1 = −1 − 1, b2 = −2 − 1.

2
C0,0 = S n+1 (S − 1 − 1)(S − 2 − 1) − z (S + 1)n (S − 1 + 1)(S − 2 + 1).
Modified moments
2
nC = n+1
C
+ (n − 2) nC
= n+2 + (2n + 1 − 1 − 2 ) n+1 + (n − 1) (n − 2) n
= z2 + (2n + 1 − 1 − 2 ) z + (n − 2) (n − 1) zn ez .
114 D. D OMINICI
2 n − 1 + 1, n − 2 + 1
nC (z) = (n − 1) (n − 2) zn 2 F2 ;z .
n − 1 , n − 2

2
C0,0 = S (S + n − 1 − 1)(S + n − 2 − 1)
− z (S + n − 1 + 1)(S + n − 2 + 1).
3.3.3.3. Geronimus Christoffel Charlier functional

Linear functional

x − 1 zx
LG,C
0,0 [u] =  x − 2 u (x) x!
, 1 = 2 , 2 ∈
/ N0 .
x=0
G,C C
0,0 = 0,0 ( − 2 ) =  ( − 1 − 1)( − 2 ) − z ( − 1 + 1)( − 2 ) ,
a1 = −1 + 1, a 2 = − 2 , b1 = −1 − 1, b 2 = − 2 .
G,C C
0,0 = 0,0 (S − 2 )
= S n+1 (S − 1 − 1) − z (S + 1)n (S − 1 + 1) (S − 2) .
n − 1 n n − 1 + 1, n − 2
nG,C (z) = z 2 F2 ;z .
n − 2 n − 1 , n − 2 + 1
G,C C
0,0 = 0,0 (S + n − 2)
= [S (S + n − 1 − 1) − z (S + n − 1 + 1)] (S + n − 2) .
3.3.3.4. Reduced-Uvarov Christoffel Charlier functional

Since for the Christoffel Charlier functional we have
 ( ) =  ( −  − 1),  ( ) =  −  + 1,
 = 0,  + 1,  − 1.
i)  = 0
Linear functional

zx
 (x −  )u (x)
U(0),C
L0,0 [u] = +  u (0) .
x=0 x!

U(0),C
0,0 = ( − 1)C0,0 =  ( −  − 1)( − 1) − z ( −  + 1)  ,
a1 = − + 1, a2 = 0, b1 = − − 1, b2 = −1.

U(0),C
0,0 = (S − 1)C0,0 = S n+1 (S −  − 1)(S − 1)
− z (S + 1)n (S −  + 1)S .

U(0),C
0,0 = (S + n)C0,0 = S (S + n −  − 1)(S + n − 1)
− z (S + n −  + 1)(S + n).
ii)  =  + 1
Linear functional

U( +1),C zx
L0,0 [u] =  (x −  )u (x) x!
+  u ( + 1)z +1 .
x=0

U( +1),C
0,0 = ( −  − 2)C0,0 =  ( −  − 1)( −  − 2)
− z ( −  + 1)( −  − 1),
a1 = − + 1, a2 = − − 1, b1 = − − 1, b2 = − − 2.

U( +1),C
0,0 = (S −  − 2)C0,0
= S n+1 (S −  − 2) − z (S + 1)n (S −  + 1) (S −  − 1).

U( +1),C
0,0 = (S + n −  − 1)C0,0
= [S (S + n −  − 2) − z (S + n −  + 1)] (S + n −  − 1).
116 D. D OMINICI
iii)  =  − 1
Linear functional

U( −1),C zx
L0,0 [u] =  (x −  )u (x) x!
+  u ( − 1)z −1 .
x=0

U( −1),C
0,0 = ( −  + 1)C0,0
=  ( −  − 1)( −  + 1) − z ( −  + 1)( −  + 2),
a1 = − + 1, a2 = − + 2, b1 = − − 1, b2 = − + 1.
U( −1),C
0,0 = (S −  + 1)C0,0
= S n+1 (S −  − 1) − z (S + 1)n (S −  + 2) (S −  + 1).
U( −1),C
0,0 = (S + n −  + 2)C0,0
= [S (S + n −  − 1) − z (S + n −  + 2)] (S + n −  + 1).
3.3.3.5. Truncated Christoffel Charlier functional

Linear functional
N
zx
LT,C
0,0 [u] =  (x −  )u (x) x!
, N ∈ N0 .
x=0

T,C C
0,0 = ( − N − 1)0,0 =  ( −  − 1)( − N − 1)
− z ( −  + 1)( − N) ,
a1 = − + 1, a2 = −N, b1 = − − 1, b2 = −N − 1.
T,C C
0,0 = (S − N − 1)0,0 = S
n+1
(S −  − 1)(S − N − 1)
− z (S + 1)n (S −  + 1)(S − N) .
T,C C
0,0 = (S + n − N)0,0 = S (S + n −  − 1)(S + n − N − 1)
− z (S + n −  + 1)(S + n − N).
3.3.3.6. Double Geronimus Charlier functional

Linear functional

1 zx
 (x − 1) (x − 2) u (x)
2
LG
0,0 [u] = , 1 , 2 ∈
/ N0 .
x=0 x!

2
G G
0,0 = 0,0 ( − 2 ) = 0,0 ( − 1 ) ( − 2 )
=  ( − 1 ) ( − 2 ) − z ( − 1 ) ( − 2 ) ,
a 1 = − 1 , a 2 = − 2 , b 1 = − 1 , b 2 = − 2 .

2
G G
0,0 = 0,0 (S − 2 ) = 0,0 (S − 1 ) (S − 2 )
= S n+1 − z (S + 1)n (S − 1 ) (S − 2 ) .
2 zn n − 1 , n − 2
nG (z) = 2 F2 ;z .
(n − 1) (n − 2) n − 1 + 1, n − 2 + 1

2
G G
0,0 = 0,0 (S + n − 2) = 0,0 (S + n − 1) (S + n − 2)
= (S − z) (S + n − 1) (S + n − 2) .
3.3.3.7. Reduced-Uvarov Geronimus Charlier functional

Since for the Geronimus Charlier functional we have
 ( ) =  ( −  ),  ( ) =  −  , ∈
/ N0 ,
we will have a reduced case for its Uvarov transformation U ( ) if  = 0.

Linear functional

u (x) zx
 x−
U(0),G
L0,0 [u] = +  u (0) , ∈
/ N0 .
x=0 x!

U(0),G
0,0 = ( − 1)G
0,0 = ( − 1)0,0 ( −  )
=  ( − 1)( −  ) − z ( −  ),
118 D. D OMINICI
a1 = 0, a2 = − , b1 = −1, b2 = − .

U(0),G
0,0 = (S − 1)G
0,0 = (S − 1)0,0 (S −  )
= S n+1 (S − 1) − z (S + 1)n S (S −  ).

U(0),G
0,0 = (S + n)G
0,0 = (S + n)0,0 (S + n −  )
= [S (S + n − 1)− z (S + n)] (S + n −  ).
3.3.3.8. Truncated Geronimus Charlier functional

Linear functional
N
u (x) zx
LT,G
0,0 [u] =  x− x!
, N ∈ N0 , ∈
/ N0 .
x=0
T,G G
0,0 = ( − N − 1)0,0 = ( − N − 1)0,0 ( −  )
=  ( − N − 1)( −  ) − z ( − N) ( −  ),
a1 = −N, a2 = − , b1 = −N − 1, b2 = − .
T,G G
0,0 = (S − N − 1)0,0 = (S − N − 1)0,0 (S −  )
= S n+1 (S − N − 1) − z (S + 1)n (S − N) (S −  ).
T,G G
0,0 = (S + n − N)0,0 = (S + n − N)0,0 (S + n −  )
= [S (S + n − N − 1) − z (S + n − N)] (S + n −  ).
3.3.3.9. Double Uvarov Charlier functional

Since for the Reduced Uvarov Charlier functional we have
U(0)
0,0 ( ) =  ( − 1),
we will have a reduced case for its Uvarov transformation U ( ) if  = 1.

Linear functional

U 2 (1,0) zx
L0,0 [u] =  u (x) x!
+ 1 u (0) + 2 u (1) z.
x=0
U 2 (1,0) U(0)
0,0 = ( − 2)0,0 = ( − 2)( − 1)0,0
=  ( − 1)( − 2) − z ( − 1),
which is a special case of (47) with  = 1.

U 2 (1,0) U(0)
0,0 = (S − 2)0,0 = (S − 2)(S − 1)0,0
= S n+1 (S − 2) − z (S + 1)n S (S − 1).
U 2 (1,0) U(0)
0,0 = (S + n − 1)0,0 = (S + n − 1)(S + n)0,0
(49)
= [S (S + n − 2)− z (S + n)] (S + n − 1).
3.3.3.10. Reduced-Uvarov Truncated Charlier functional

Since for the Truncated Charlier functional we have
 ( ) =  ( − N − 1),  ( ) =  − N, N ∈ N0 ,
 = 0, N + 1, N.
i)  = 0
Linear functional
N
zx
 u (x)
U(0),T
L0,0 [u] = +  u (0) , N ∈ N0 .
x=0 x!

U(0),T
0,0 = ( − 1)T0,0 = ( − 1)( − N − 1)0,0
=  ( − N − 1)( − 1) − z ( − N)  ,
a1 = −N, a2 = 0, b1 = −N − 1, b2 = −1.
120 D. D OMINICI

U(0),T
0,0 = (S − 1)T0,0 = (S − 1)(S − N − 1)0,0
= S n+1 (S − 1)(S − N − 1) − z (S + 1)n S (S − N) .

U(0),T
0,0 = (S + n)T0,0 = (S + n)(S + n − N)0,0
= S (S + n − 1)(S + n − N − 1) − z (S + n)(S + n − N).
ii)  = N + 1
Linear functional
N
zx
 u (x)
U(N+1),T
L0,0 [u] = +  u (N + 1)zN+1 , N ∈ N0 .
x=0 x!

U(N+1),T
0,0 = ( − N − 2)T0,0 =  ( − N − 1)( − N − 2)
− z ( − N)( − N − 1),
which is a special case of (47) with  = N + 1.

U(N+1),T
0,0 = (S − N − 2)T0,0 = (S − N − 2)(S − N − 1)0,0
= S n+1 (S − N − 2) − z (S + 1)n (S − N) (S − N − 1).

U(N+1),T
0,0 = (S + n − N − 1)T0,0 = (S + n − N − 1)(S + n − N)0,0
= [S (S + n − N − 2) − z (S + n − N)] (S + n − N − 1).
iii)  = N
Linear functional
N
zx
 u (x)
U(N),T
L0,0 [u] = +  u (N) , N ∈ N0 .
x=0 x!

U(N),T
0,0 = ( − N) T0,0 = ( − N) ( − N − 1)0,0
=  ( − N − 1)( − N) − z ( − N) ( − N + 1),
which is a special case of (47) with  = N.


U(N),T
0,0 = (S − N) T0,0 = (S − N) (S − N − 1)0,0
= S n+1 (S − N − 1) − z (S + 1)n (S − N + 1) (S − N) .

U(N),T
0,0 = (S + n − N + 1)T0,0 = (S + n − N + 1)(S + n − N)0,0
= [S (S + n − N − 1) − z (S + n − N + 1)] (S + n − N).
3.3.3.11. Christoffel Generalized Meixner functional

Linear functional

(a)x zx
LC1,1 [u] =  (x −  )u (x) (b + 1)x x!
.
x=0
C0,1 =  ( + b)( −  − 1) − z ( + a)( −  + 1),
a1 = a, a2 = − + 1, b1 = b, b2 = − − 1.
C1,1 = S n+1 (S + b)(S −  − 1) − z (S + 1)n (S + a)(S −  + 1).

C (a)n n a + n, n −  + 1
n (z) = (n −  ) z 2 F2 ;z .
(b + 1)n b + 1 + n, n − 
C1,1 = S (S + n + b)(S + n −  − 1) − z (S + n + a)(S + n −  − 1).
3.3.3.12. Geronimus Generalized Meixner functional

Linear functional

u (x) (a)x zx
LG
1,1 [u] =  x− (b + 1)x x!
, ∈
/ N0 .
x=0
G
1,1 = 1,1 ( −  ) =  ( + b)( −  ) − z ( + a)( −  ) ,
122 D. D OMINICI
a1 = a, a2 = − , b1 = b, b2 = − .
G
1,1 = 1,1 (S −  ) = S
n+1
(S + b) − z (S + 1)n (S + a) (S −  )

G 1 (a)n n a + n, n − 
n (z) = z 2 F2 ;z .
n −  (b + 1)n b + 1 + n, n −  + 1
G
1,1 = 1,1 (S + n −  ) = [S (S + n + b)− z (S + n + a)] (S + n −  ).
3.3.3.13. Reduced-Uvarov Generalized Meixner functional

Since for the generalized Meixner functional we have
 ( ) =  ( + b),  ( ) =  + a,
 = 0, −b, −a.
i)  = 0
Linear functional

(a)x zx
 u (x)
U(0)
L1,1 [u] = +  u (0) .
x=0 (b + 1)x x!

U(0)
1,1 = ( − 1)1,1 =  ( + b)( − 1) − z ( + a)  ,
a1 = a, a2 = 0, b1 = b, b2 = −1.

U(0)
1,1 = (S − 1)1,1 = S n+1 (S + b)(S − 1) − z (S + 1)n (S + a)S .

U(0)
1,1 = (S + n)1,1 = S (S + n + b)(S + n − 1) − z (S + n + a)(S + n).
ii)  = −a
Linear functional

(a)x zx
 u (x)
U(−a)
L1,1 [u] = +  u (−a)z−a .
x=0 (b + 1)x x!
U(−a)
1,1 = ( + a)1,1 =  ( + b)( + a) − z ( + a)( + a + 1),
a1 = a, a2 = a + 1, b1 = b, b2 = a.
U(−a)
1,1 = (S + a)1,1 = S n+1 (S + b) − z (S + 1)n (S + a + 1) (S + a).
U(−a)
1,1 = (S + n + a + 1)1,1 = [S (S + n + b) − z (S + n + a + 1)] (S + n + a).
iii)  = −b
Linear functional

(a)x zx
 u (x)
U(−b)
L1,1 [u] = +  u (−b)z−b .
x=0 (b + 1)x x!
U(−b)
1,1 = ( + b − 1)1,1 =  ( + b)( + b − 1) − z ( + a)( + b),
a1 = a, a2 = b, b1 = b, b2 = b − 1.
U(−b)
1,1 = (S + b − 1)1,1 = S n+1 (S + b − 1)− z (S + 1)n (S + a) (S + b).
U(−b)
1,1 = (S + n + b)1,1 = [S (S + n + b − 1)− z (S + n + a)] (S + n + b).
124 D. D OMINICI
3.3.3.14. Truncated Generalized Meixner functional

Linear functional
N
(a)x zx
LT1,1 [u] =  u (x) (b + 1)x x!
, N ∈ N0 .
x=0
T1,1 = ( − N − 1)1,1 =  ( + b)( − N − 1) − z ( + a)( − N),
a1 = a, a2 = −N, b1 = b, b2 = −N − 1.
T1,1 = (S − N − 1)1,1 = S n+1 (S + b)(S − N − 1)

− z (S + 1)n (S + a)(S − N) .
T1,1 = (S + n − N)1,1 = S (S + n + b)(S + n − N − 1)

− z (S + n + a)(S + n − N).
3.3.4. Functional of type (3,0;N)

Linear functional
N
zx
L3,0;N [u] =  u (x) (−N)x (a1)x (a2 )x x!
, N ∈ N0 ,
x=0
3,0;N =  − z ( − N) ( + a1 ) ( + a2) .
3,0;N = S n+1 − z (S + 1)n (S − N) (S + a1) (S + a2) .
n − N, a1 + n, a2 + n
n (z) = (−N)n (a1 )n (a2 )n zn 3 F0 ;z .
−
3,0;N = S − z (S + n − N)(S + n + a1) (S + n + a2) .

3.3.5. Functional of type (3,1;N)

Linear functional
N
(−N)x (a1 )x (a2 )x zx
L3,1;N [u] =  u (x) (b + 1)x x!
, N ∈ N0 .
x=0
First moment
−N, a1 , a2
0 (z) = 3 F1 ;z .
b+1
3,1;N =  ( + b) − z ( − N) ( + a1) ( + a2 ) . (50)

3,1;N = S n+1 (S + b) − z (S + 1)n (S − N)(S + a1) (S + a2 ) .
(−N)n (a1 )n (a2 )n n n − N, a1 + n, a2 + n
n (z) = z 3 F1 ;z .
(b + 1)n b+1+n
3,1;N = S (S + n + b)− z (S + n − N)(S + n + a1) (S + n + a2) .
3.3.5.1. Christoffel Generalized Krawtchouk functional

Linear functional
N
zx
LC2,0;N [u] =  (x −  )u (x) (−N)x (a)x x! , N ∈ N0 .
x=0

C2,0;N =  ( −  − 1) − z ( − N) ( + a)( −  + 1)
a1 = −N, a2 = a, a3 = − + 1, b = − − 1.
C2,0;N = S n+1 (S −  − 1) − z (S + 1)n (S − N) (S + a)(S −  + 1).

(−N)n (a)n n n − N, a + n, n −  + 1
nC (z) = (n −  ) z 3 F1 ;z .
(b + 1)n n−
C2,0;N = S (S + n −  − 1) − z (S + n − N)(S + n + a)(S + n −  − 1).
126 D. D OMINICI
3.3.5.2. Geronimus Generalized Krawtchouk functional

Linear functional
N
u (x) zx
LG
2,0;N [u] =  x− (−N)x (a)x
x!
, N ∈ N0 , ∈
/ N0 .
x=0
G
2,0;N = 2,0;N ( −  ) =  ( −  ) − z ( − N) ( + a)( −  )
a1 = −N, a2 = a, a3 = − , b = − .
G
2,0;N = 2,0;N (S −  ) = S
n+1
− z (S + 1)n (S − N) (S + a) (S −  )

G 1 (−N)n (a)n n n − N, a + n, n − 
n (z) = z 3 F1 ;z .
n −  (b + 1)n n−+1
G
2,0;N = 2,0;N (S + n −  ) = [S − z (S + n − N)(S + n + a)] (S + n −  ).
3.3.5.3. Reduced-Uvarov Generalized Krawtchouk functional

Since for the Generalized Krawtchouk functional we have
 ( ) =  ,  ( ) = ( − N) ( + a), N ∈ N0 ,
 = 0, N, −a.
i)  = 0
Linear functional
N
zx
 u (x) (−N)x (a)x x! +  u (0) ,
U(0)
L2,0;N [u] = N ∈ N0 .
x=0

U(0)
2,0;N = ( − 1)2,0;N =  ( − 1) − z ( − N) ( + a)  ,
a1 = −N, a2 = a, a3 = 0, b = −1.

U(0)
2,0;N = (S − 1)2,0;N = S n+1 (S − 1) − z (S + 1)n (S − N)(S + a)S .

U(0)
2,0;N = (S + n)2,0;N = S (S + n − 1) − z (S + n − N)(S + n + a)(S + n).
ii)  = N
Linear functional
N
zx
 u (x) (−N)x (a)x x! +  u (N) zN .
U(N)
L2,0;N [u] =
x=0

U(N)
2,0;N = ( − N) 2,0;N =  ( − N) − z ( − N) ( + a)( − N + 1),
a1 = −N, a2 = a, a3 = −N + 1, b = −N.

U(N)
2,0;N = (S − N) 2,0;N
= S n+1 − z (S + 1)n (S + a)(S − N + 1) (S − N) .

U(N)
2,0;N = (S + n − N + 1)2,0;N
= [S − z (S + n + a)(S + n − N + 1)] (S + n − N).
iii)  = −a
Linear functional
N
zx
 u (x) (−N)x (a)x x! +  u (−a)z−a .
U(−a)
L2,0;N [u] =
x=0

U(−a)
2,0;N = ( + a)2,0;N
=  ( + a) − z ( − N) ( + a)( + a + 1),
a1 = −N, a2 = a, a3 = a + 1, b = a.
128 D. D OMINICI

U(−a)
2,0;N = (S + a)2,0;N
= S n+1 − z (S + 1)n (S − N) (S + a + 1) (S + a).

U(−a)
2,0;N = (S + n + a + 1)2,0;N
= [S − z (S + n − N)(S + n + a + 1)] (S + n + a).

Linear functional

(a1 ) (a2 )x (a3 )x zx
L3,2 [u] =  u (x) (b1 +x1) .
x=0 x (b2 + 1)x x!
First moment
a1 , a2 , a3
0 (z) = 3 F2 ;z .
b1 + 1, b2 + 1
3,2 =  ( + b1) ( + b2 ) − z ( + a1) ( + a2 ) ( + a3) . (51)

3,2 = S n+1 (S + b1) (S + b2 )

− z (S + 1)n (S + a1) (S + a2) (S + a3 ) .
(−N)n (a1 )n (a2 )n n n − N, a1 + n, a2 + n

n (z) = z 3 F1 ;z .
(b1 + 1)n (b2 + 1)n b + 1 + n, b2 + 1 + n
3,2 = S (S + n + b1) (S + n + b2)

− z (S + n + a1) (S + n + a2) (S + n + a3) .
3.3.6.1. Uvarov Meixner functional

Linear functional

zx
LU1,0 [u] =  u (x) (a)x x!
+  u ( ) z ,  = 0, −a.
x=0
U1,0 = ( −  )( −  − 1)1,0

=  ( −  )( −  − 1) − z ( + a)( −  + 1)( −  ),
a1 = a, a2 = − + 1, a3 = − , b1 = − , b2 = − − 1.
U1,0 = (S −  )(S −  − 1)1,0

= S n+1 (S −  − 1) − z (S + 1)n (S + a)(S −  + 1) (S −  ).
U1,0 = (S + n −  )(S + n −  + 1)1,0

= [S (S + n −  − 1) − z (S + n + a)(S + n −  + 1)] (S + n −  ).
3.3.6.2. Geronimus Christoffel Meixner functional

Linear functional

x − 1 zx
LG,C
1,0 [u] =  x − 2 u (x) (a)x x!
, 1 = 2 , 2 ∈
/ N0 .
x=0
G,C C
1,0 = 1,0 ( − 2 ) =  ( − 1 − 1)( − 2 ) − z ( + a)( − 1 + 1)( − 2 ) ,
a1 = a, a2 = −1 + 1, a 3 = − 2 , b1 = −1 − 1, b 2 = − 2 .
G,C C
1,0 = 1,0 (S − 2 )
= S n+1 (S − 1 − 1) − z (S + 1)n (S + a)(S − 1 + 1) (S − 2 ) .

n − 1 a + n, n − 1 + 1, n − 2
nG,C (z) = (a)n zn 3 F2 ;z .
n − 2 n − 1 , n − 2 + 1
G,C C
1,0 = 1,0 (S + n − 2)
= [(S + n − 1 − 1)S − z (S + n + a)(S + n − 1 + 1)] (S + n − 2) .
130 D. D OMINICI
3.3.6.3. Double Christoffel Meixner functional

Linear functional

zx
 (x − 1) (x − 2) u (x) (a)x
2
LC1,0 [u] = .
x=0 x!

2
C1,0 =  ( − 1 − 1)( − 2 − 1) − z ( + a)( − 1 + 1)( − 2 + 1),
a1 = a, a2 = −1 + 1, a3 = −2 + 1, b1 = −1 − 1, b2 = −2 − 1.

2
C1,0 = S n+1 (S − 1 − 1)(S − 2 − 1)
− z (S + 1)n (S + a)(S − 1 + 1)(S − 2 + 1).
Modified moments
2
nC = n+1
C
+ (n − 2) nC
= n+2 + (2n + 1 − 1 − 2 ) n+1 + (n − 1) (n − 2) n

= 2 z2 + 1 z + 0 zn (a)n (1 − z)−a−n−2 ,
where
2 = (a + 1) (a + 2) ,
1 = a − (a + 2)1 − (a + 1)2 + n(1 + 2a + 1 + 2)
0 = (n − 1) (n − 2) .
2 a + n, n − 1 + 1, n − 2 + 1
nC (z) = (n − 1) (n − 2) (a)n zn 3 F2 ;z .
n − 1 , n − 2

2
C1,0 = S (S + n − 1 − 1)(S + n − 2 − 1)
− z (S + n + a)(S + n − 1 + 1)(S + n − 2 + 1).
3.3.6.4. Reduced-Uvarov Christoffel Meixner functional

Since for the Christoffel Meixner functional we have
 ( ) =  ( −  − 1),  ( ) = ( + a)( −  + 1),
 = 0,  + 1, −a,  − 1.
i)  = 0
Linear functional

zx
 (x −  )u (x) (a)x x! +  u (0) .
U(0),C
L1,0 [u] =
x=0

U(0),C
1,0 = ( − 1)C1,0 =  ( −  − 1)( − 1) − z ( + a)( −  + 1)  ,
a1 = a, a2 = − + 1, a3 = 0, b1 = − − 1, b2 = −1.

U(0),C
1,0 = (S − 1)C1,0
= S n+1 (S −  − 1)(S − 1) − z (S + 1)n (S + a)(S −  + 1)S .

U(0),C
1,0 = (S + n)C1,0 = S (S + n −  − 1)(S + n − 1)
− z (S + a)(S + n −  + 1)(S + n).
ii)  =  + 1
Linear functional

U( +1),C zx
L1,0 [u] =  (x −  )u (x) (a)x x! +  u ( + 1)z +1 .
x=0

U( +1),C
1,0 = ( −  − 2)C1,0
=  ( −  − 1)( −  − 2) − z ( + a)( −  + 1)( −  − 1),
a1 = a, a2 = − + 1, a3 = − − 1, b1 = − − 1, b2 = − − 2.
132 D. D OMINICI

U( +1),C
1,0 = (S −  − 2)C1,0
= S n+1 (S −  − 2) − z (S + 1)n (S + a)(S −  + 1) (S −  − 1).

U( +1),C
1,0 = (S + n −  − 1)C1,0
= [S (S + n −  − 2) − z (S + n + a)(S + n −  + 1)] (S + n −  − 1).
iii)  = −a
Linear functional

zx
 (x −  )u (x) (a)x x! +  u (−a)z−a .
U(−a),C
L0,0 [u] =
x=0

U(−a),C
1,0 = ( + a)C1,0
=  ( −  − 1)( + a) − z ( + a)( −  + 1)( + a + 1),
a1 = a, a2 = − + 1, a3 = a + 1, b1 = − − 1, b2 = a.

U(−a),C
1,0 = (S + a)C1,0
= S n+1 (S −  − 1) − z (S + 1)n (S −  + 1)(S + a + 1) (S + a).

U(−a),C
1,0 = (S + n + a + 1)C1,0
= [S (S + n −  − 1) − z (S + n −  + 1)(S + n + a + 1)] (S + n + a).
iv)  =  − 1
Linear functional

U( −1),C zx
L1,0 [u] =  (x −  )u (x) (a)x x! +  u ( − 1)z −1 .
x=0

U( −1),C
1,0 = ( −  + 1)C1,0
=  ( −  − 1)( −  + 1) − z ( + a)( −  + 1)( −  + 2),
a1 = a, a2 = − + 1, a3 = − + 2, b1 = − − 1, b2 = − + 1.

U( −1),C
1,0 = (S −  + 1)C1,0
= S n+1 (S −  − 1) − z (S + 1)n (S + a)(S −  + 2) (S −  + 1).

U( −1),C
1,0 = (S + n −  + 2)C1,0
= [S (S + n −  − 1) − z (S + n + a)(S + n −  + 2)] (S + n −  + 1).
3.3.6.5. Truncated Christoffel Meixner functional

Linear functional
N
zx
LT,C
1,0 [u] =  (x −  )u (x) (a)x x! , N ∈ N0 .
x=0
T,C C
1,0 = ( − N − 1)1,0
=  ( −  − 1)( − N − 1) − z ( + a)( −  + 1)( − N),
a1 = a, a2 = − + 1, a3 = −N, b1 = − − 1, b2 = −N − 1.
T,C C
1,0 = (S − N − 1)1,0 = S
n+1
(S −  − 1)(S − N − 1)
− z (S + 1)n (S + a)(S −  + 1)(S − N) .
T,C C
1,0 = (S + n − N)1,0 = S (S + n −  − 1)(S + n − N − 1)
− z (S + n + a)(S + n −  + 1)(S + n − N).
3.3.6.6. Double Geronimus Meixner functional

Linear functional

1 zx
 (x − 1) (x − 2) u (x) (a)x x! ,
2
LG
1,0 [u] = 1 , 2 ∈
/ N0 .
x=0
134 D. D OMINICI

2
G G
1,0 = 1,0 ( − 2 ) = 1,0 ( − 1 ) ( − 2 )
=  ( − 1 ) ( − 2 ) − z ( + a)( − 1 ) ( − 2 ) ,
a1 = a, a 2 = − 1 , a 3 = − 2 , b 1 = − 1 , b 2 = − 2 .

2
G G
1,0 = 1,0 (S − 2 ) = 1,0 (S − 1 ) (S − 2 )
= S n+1 − z (S + 1)n (S + a) (S − 1 ) (S − 2 ) .
2 (a)n zn a + n, n − 1, n − 2
nG (z) = 3 F2 ;z .
(n − 1) (n − 2) n − 1 + 1, n − 2 + 1

2
G G
1,0 = 1,0 (S + n − 2) = 1,0 (S + n − 1) (S + n − 2)
= [S − z (S + n + a)] (S + n − 1) (S + n − 2) .
3.3.6.7. Reduced-Uvarov Geronimus Meixner functional

Since for the Geronimus Meixner functional we have
 ( ) =  ( −  ) ,  ( ) = ( + a)( −  ), ∈
/ N0 ,
we will have a reduced case for its Uvarov transformation U ( ) if
 = 0, −a.
i)  = 0
Linear functional

u (x) zx
 x−
U(0),G
L1,0 [u] = (a)x +  u (0) , ∈
/ N0 .
x=0 x!

U(0),G
1,0 = ( − 1)G
1,0 = ( − 1)1,0 ( −  )
=  ( −  )( − 1) − z ( + a)( −  )  ,
a1 = a, a2 = − , a3 = 0, b1 = − , b2 = −1.

U(0),G
1,0 = (S − 1)G
1,0 = (S − 1)1,0 (S −  )
= S n+1 (S − 1) − z (S + 1)n (S + a)S (S −  ).

U(0),G
1,0 = (S + n)G
1,0 = (S + n)1,0 (S + n −  )
= [S (S + n − 1)− z (S + n + a)(S + n)] (S + n −  ).
ii)  = −a
Linear functional

u (x) zx
 x−
U(−a),G
L1,0 [u] = (a)x +  u (−a) z−a , ∈
/ N0 .
x=0 x!

U(−a),G
1,0 = ( + a)G
1,0 = ( + a)1,0 ( −  )
=  ( −  )( + a) − z ( + a)( −  )( + a + 1),
a1 = a, a2 = − , a3 = a + 1, b1 = − , b2 = a.

U(−a),G
1,0 = (S + a)G
1,0 = (S + a)1,0 (S −  )
= S n+1 − z (S + 1)n (S + a + 1) (S + a)(S −  ).

U(−a),G
1,0 = (S + a + n + 1)G
1,0 = (S + a + n + 1)1,0 (S + n −  )
= [S − z (S + a + n + 1)] (S + a + n)(S + n −  ).
3.3.6.8. Truncated Geronimus Meixner functional

Linear functional
N
u (x) zx
LT,G
1,0 [u] =  x− (a)x
x!
, N ∈ N0 , ∈
/ N0 .
x=0
T,G G
1,0 = ( − N − 1)1,0 = ( − N − 1)1,0 ( −  )
=  ( −  )( − N − 1) − z ( + a)( −  )( − N) ,
136 D. D OMINICI
a1 = a, a2 = − , a3 = −N b1 = − , b2 = −N − 1.
T,G G
1,0 = (S − N − 1)1,0 = (S − N − 1)1,0 (S −  )
= S n+1 (S − N − 1) − z (S + 1)n (S + a)(S − N) (S −  ).
T,G G
1,0 = (S + n − N)1,0 = (S + n − N)1,0 (S + n −  )
= [S (S + n − N − 1) − z (S + n + a)(S + n − N)] (S + n −  ).
3.3.6.9. Double Uvarov Meixner functional

Since for the Reduced-Uvarov Meixner functional we have
U(0) U(0)
1,0 ( ) =  ( − 1), 1,0 ( ) =  ( + a)
U(−a) U(−a)
1,0 ( ) =  ( + a), 1,0 ( ) = ( + a)( + a + 1),
we will have a reduced case for their Uvarov transformations U ( ) if
 = 1, −a, or  = 0, −a − 1.
i)  = 1
Linear functional

zx
 u (x) (a)x
U(1,0)
L1,0 [u] = + 1 u (0) + 2 u (1) z.
x=0 x!

U(1,0) U(0)
1,0 = ( − 2)1,0 = ( − 2)( − 1)1,0
=  ( − 1)( − 2) − z ( + a)  ( − 1),
a1 = a, a2 = 0, a3 = −1, b1 = −1, b2 = −2.

U(1,0) U(0)
1,0 = (S − 2)1,0 = (S − 2)(S − 1)1,0
= S n+1 (S − 2) − z (S + 1)n (S + a)S (S − 1).

U(1,0) U(0)
1,0 = (S + n − 1)1,0 = (S + n − 1)(S + n)1,0
= [S (S + n − 2) − z (S + n + a)(S + n)] (S + n − 1).
ii)  = −a
Linear functional

zx
 u (x) (a)x
U(−a,0)
L1,0 [u] = + 1 u (0) + 2 u (−a) z−a .
x=0 x!

U(−a,0) U(0)
1,0 = ( + a)1,0 = ( + a)( − 1)1,0
=  ( − 1)( + a) − z ( + a)  ( + a + 1),
a1 = a, a2 = 0, a3 = a + 1, b1 = −1, b2 = a.

U(−a,0) U(1)
1,0 = (S + a)1,0 = (S + a)(S − 1)1,0
= S n+1 (S − 1) − z (S + 1)n S (S + a + 1) (S + a).

U(−a,0) U(0)
1,0 = (S + n + a + 1)1,0 = (S + n + a + 1)(S + n)1,0
= [S (S + n − 1) − z (S + n)(S + n + a + 1)] (S + n + a).
iii)  = −a − 1
Linear functional

zx
 u (x) (a)x
U(−a−1,−a)
L1,0 [u] = + 1 u (−a) z−a + 2 u (−a − 1)z−a−1 .
x=0 x!

U(−a−1,−a) U(−a)
1,0 = ( + a + 1)1,0 = ( + a + 1)( + a)1,0
=  ( + a)( + a + 1) − z ( + a)( + a + 1)( + a + 2),
a1 = a, a2 = a + 1, a3 = a + 2, b1 = a, b2 = a + 1.
138 D. D OMINICI

U(−a−1,−a) U(−a)
1,0 = (S + a + 1)1,0 = (S + a + 1)(S + a)1,0
= S n+1 − z (S + 1)n (S + a + 2) (S + a)(S + a + 1).

U(−a−1,−a) U(−a)
1,0 = (S + n + a + 2)1,0 = (S + n + a + 2)(S + n + a + 1)1,0
= [S − z (S + n + a + 2)] (S + n + a)(S + n + a + 1).
3.3.6.10. Reduced-Uvarov Truncated Meixner functional

Since for the Truncated Meixner functional we have
 ( ) =  ( − N − 1),  ( ) = ( + a)( − N) , N ∈ N0 ,
 = 0, N + 1, −a, N.
i)  = 0
Linear functional
N
zx
 u (x) (a)x x! +  u (0) ,
U(0),T
L1,0 [u] = N ∈ N0 .
x=0

U(0),T
1,0 = ( − 1)T1,0 = ( − 1)( − N − 1)1,0
=  ( − N − 1)( − 1) − z ( + a)( − N)  ,
a1 = a, a2 = −N, a3 = 0, b1 = −N − 1, b2 = −1.

U(0),T
1,0 = (S − 1)T1,0 = (S − 1)(S − N − 1)1,0
= S n+1 (S − N − 1)(S − 1) − z (S + 1)n (S + a)(S − N)S .

U(0),T
1,0 = (S + n)T1,0 = (S + n)(S + n − N)1,0
= S (S + n − N − 1)(S + n − 1)− z (S + a)(S + n − N)(S + n).
ii)  = N + 1
Linear functional
N
zx
 u (x) (a)x x! +  u (N + 1)zN+1 ,
U(N+1),T
L1,0 [u] = N ∈ N0 .
x=0

U(N+1),T
1,0 = ( − N − 2)T1,0 = ( − N − 2)( − N − 1)1,0
=  ( − N − 1)( − N − 2) − z ( + a)( − N) ( − N − 1),
a1 = a, a2 = −N, a3 = −N − 1, b1 = −N − 1, b2 = −N − 2.

U(N+1),T
1,0 = (S − N − 2)T1,0 = (S − N − 2)(S − N − 1)1,0
= S n+1 (S − N − 2) − z (S + 1)n (S + a)(S − N) (S − N − 1).

U(N+1),T
1,0 = (S + n − N − 1)T1,0 = (S + n − N − 1)(S + n − N)1,0
= [S (S + n − N − 2) − z (S + n + a)(S + n − N)] (S + n − N − 1).
iii)  = −a
Linear functional
N
zx
 u (x) (a)x x! +  u (−a)z−a ,
U(−a),T
L1,0 [u] = N ∈ N0 .
x=0

U(−a),T
1,0 = ( + a)T1,0 = ( + a)( − N − 1)1,0
=  ( − N − 1)( + a) − z ( + a)( − N) ( + a + 1),
a1 = a, a2 = −N, a3 = a + 1, b1 = −N − 1, b2 = a.

U(−a),T
1,0 = (S + a)T1,0 = (S + a)(S − N − 1)1,0
= S n+1 (S − N − 1) − z (S + 1)n (S − N) (S + a + 1) (S + a).
140 D. D OMINICI

U(−a),T
1,0 = (S + n + a + 1)T1,0 = (S + n + a + 1)(S + n − N)1,0
= [S (S + n − N − 1) − z (S + n − N)(S + n + a + 1)] (S + n + a).
iv)  = N
Linear functional
N
zx
 u (x) (a)x x! +  u (N) zN ,
U(N),T
L1,0 [u] = N ∈ N0 .
x=0

U(N),T
1,0 = ( − N) T1,0 = ( − N) ( − N − 1)1,0
=  ( − N − 1)( − N) − z ( + a)( − N)( − N + 1),
a1 = a, a2 = −N, a3 = −N + 1, b1 = −N − 1, b2 = −N.

U(N),T
1,0 = (S − N) T1,0 = (S − N) (S − N − 1)1,0
= S n+1 (S − N − 1) − z (S + 1)n (S + a)(S − N + 1) (S − N) .

U(N),T
1,0 = (S + n − N + 1)T1,0 = (S + n − N + 1)(S + n − N)1,0
= [S (S + n − N − 1) − z (S + n + a)(S + n − N + 1)] (S + n − N).
3.3.6.11. Christoffel Generalized Hahn functional

Linear functional

(a1 )x (a2 )x zx
LC2,1 [u] =  (x −  )u (x) (b + 1)x x!
.
x=0
C2,1 =  ( + b)( −  − 1) − z ( + a1 ) ( + a2) ( −  + 1),
a3 = − + 1, b1 = b, b2 = − − 1.
C2,1 = S n+1 (S + b)(S −  − 1) − z (S + 1)n (S + a1 ) (S + a2) (S −  + 1).

(a1 )n (a2 )n n a + n, a2 + n, n −  + 1
nC (z) = (n −  ) z 3 F2 1 ;z .
(b + 1)n b + 1 + n, n − 
C2,1 = S (S + n + b)(S + n −  − 1)
− z (S + n + a1) (S + n + a2) (S + n −  + 1).
3.3.6.12. Geronimus Generalized Hahn functional

Linear functional

u (x) (a1 )x (a2 )x zx
LG
2,1 [u] =  x− (b + 1)x x!
, ∈
/ N0 .
x=0
G
2,1 = 2,1 ( −  ) =  ( + b)( −  ) − z ( + a1 ) ( + a2 ) ( −  ) ,
a3 = − , b1 = b, b2 = − .
G
2,1 = 2,1 (S −  ) = S
n+1
(S + b) − z (S + 1)n (S + a1 ) (S + a2) (S −  )

G 1 (a1 )n (a2 )n n a1 + n, a2 + n, n − 
n (z) = z 3 F2 ;z .
n −  (b + 1)n b + 1 + n, n −  + 1
G
2,1 = 2,1 (S + n −  )
= [S (S + n + b) − z (S + n + a1) (S + n + a2)] (S + n −  ).
3.3.6.13. Reduced-Uvarov Generalized Hahn functional

Since for the generalized Hahn functional we have
 ( ) =  ( + b),  ( ) = ( + a1 ) ( + a2) ,
 = 0, −b, −a1, −a2 .

142 D. D OMINICI
i)  = 0
Linear functional

(a1 )x (a2 )x zx
 u (x)
U(0)
L2,1 [u] = +  u (0) .
x=0 (b + 1)x x!

U(0)
2,1 = ( − 1)2,1 =  ( + b)( − 1) − z ( + a1) ( + a2)  ,
a3 = 0, b1 = b, b2 = −1.

U(0)
2,1 = (S − 1)2,1 = S n+1 (S + b)(S − 1)
− z (S + 1)n (S + a1) (S + a2 ) S .

U(0)
2,1 = (S + n)2,1 = S (S + n + b)(S + n − 1)
− z (S + n + a1) (S + n + a2) (S + n).
For the special case z = 1, see [5].

ii)  = −b
Linear functional

(a1 )x (a2 )x zx
 u (x)
U(−b)
L2,1 [u] = +  u (−b) z−b .
x=0 (b + 1)x x!

U(−b)
2,1 = ( + b − 1)2,1 =  ( + b)( + b − 1)
− z ( + a1) ( + a2 ) ( + b),
a3 = b, b1 = b, b2 = b − 1.

U(−b)
2,1 = (S + b − 1)2,1
= S n+1 (S + b − 1)− z (S + 1)n (S + a1) (S + a2 ) (S + b).

U(0)
2,1 = (S + n + b)2,1
= [S (S + n + b − 1)− z (S + n + a1) (S + n + a2)] (S + n + b).
iii)  = −a1
Linear functional

(a1 )x (a2 )x zx
 u (x)
U(−a1 )
L2,1 [u] = +  u (−a1 ) z−a1 .
x=0 (b + 1)x x!
U(−a )
2,1 1 = ( + a1) 2,1 =  ( + b)( + a1) − z ( + a1) ( + a2) ( + a1 + 1),
a3 = a1 + 1, b1 = b, b2 = a1 .
U(−a1 )
2,1 = (S + a1 ) 2,1
= S n+1 (S + b) − z (S + 1)n (S + a2) (S + a1 + 1) (S + a1 ) .
U(−a1 )
2,1 = (S + n + a1 + 1)2,1
= [S (S + n + b) − z (S + n + a2) (S + n + a1 + 1)] (S + n + a1) .
3.3.6.14. Truncated Generalized Hahn functional

Linear functional
N
(a1 )x (a2 )x zx
LT2,1 [u] =  u (x) (b + 1)x x!
, N ∈ N0 .
x=0

T2,1 = ( − N − 1)2,1 =  ( + b)( − N − 1) − z ( + a1) ( + a2 ) ( − N) ,
a3 = −N, b1 = b, b2 = −N − 1.
T2,1 = (S − N − 1)2,1 = S n+1 (S + b)(S − N − 1)
− z (S + 1)n (S + a1 ) (S + a2) (S − N) .
T2,1 = (S + n − N)2,1 = S (S + n + b)(S + n − N − 1)
− z (S + n + a1) (S + n + a2) (S + n − N).
144 D. D OMINICI
4. Summary
In this section, we list all the functionals and their transformations. Note that we
have
s=0 s=1 s=2
Charlier (0, 0) 4 cases (1, 1) 15 cases (2, 2)
Meixner (1, 0) 5 cases (2, 1) 20 cases (3, 2)
Generalized Charlier (0, 1) 5 cases (1, 2)
Generalized Meixner (1, 1) 6 cases (2, 2)
Generalized Krawtchouk (2, 0; N) 5 cases (3, 1; N)
Generalized Hahn (2, 1) 7 cases (3, 2)
Charlier
Differential operator:
00 =  − z.
(i) Christoffel transformation (of type 1, 1):
C0,0 =  ( −  − 1) − z ( −  + 1).
Double transformations (of type 2, 2 ):

2
C0,0 =  ( − 1 − 1)( − 2 − 1) − z ( − 1 + 1)( − 2 + 1),
G,C
0,0 =  ( − 1 − 1)( − 2 ) − z ( − 1 + 1)( − 2 ) ,
U(0),C
0,0 =  ( −  − 1)( − 1) − z ( −  + 1)  ,
U( +1),C
0,0 =  ( −  − 1)( −  − 2) − z ( −  + 1)( −  − 1),
U( −1),C
0,0 =  ( −  − 1)( −  + 1) − z ( −  + 1)( −  + 2),
T,C
0,0 =  ( −  − 1)( − N − 1) − z ( −  + 1)( − N) .
(ii) Geronimus transformation (of type 1, 1):
G
0,0 =  ( −  ) − z ( −  ).
C,G G,C
0,0 = 0,0 ,
2
G
0,0 =  ( − 1 ) ( − 2 ) − z ( − 1 ) ( − 2 ) ,
U(0),G
0,0 =  ( −  ) ( − 1) − z ( −  )  ,
U( ),G
0,0 = G
0,0 ,
T,G
0,0 =  ( −  ) ( − N − 1) − z ( −  )( − N) .
(iii) Reduced-Uvarov transformation U (0) (of type 1, 1):

U(0)
0,0 =  ( − 1) − z .

C,U(0) U(0),C G,U(0) U(0),G
0,0 = 0,0 , 0,0 = 0,0 ,
U 2 (1,0)
0,0 =  ( − 1)( − 2) − z ( − 1),
T,U(0)
0,0 =  ( − 1)( − N − 1) − z ( − N).
(iv) Truncation transformation (of type 1, 1):
T0,0 =  ( − N − 1) − z ( − N).

U(0),T T,U(0)
C,T T,C
0,0 = 0,0 , G,T T,G
0,0 = 0,0 , 0,0 = 0,0 ,
U(N+1),T
0,0 =  ( − N − 1)( − N − 2) − z ( − N) ( − N − 1),
U(N),T
0,0 =  ( − N − 1)( − N) − z ( − N) ( − N + 1).
(v) Uvarov transformation (of type 2, 2):
U0,0 =  ( −  − 1)( −  ) − z ( −  )( −  + 1),  = 0.
Meixner
1,0 =  − z ( + a).
(i) Christoffel transformation (of type 2, 1):
C1,0 =  ( −  − 1) − z ( + a)( −  + 1).

2
C1,0 =  ( − 1 − 1)( − 2 − 1) − z ( + a)( − 1 + 1)( − 2 + 1),
146 D. D OMINICI
G,C
1,0 =  ( − 1 − 1)( − 2 ) − z ( + a)( − 1 + 1)( − 2 ) ,
U(0),C
1,0 =  ( −  − 1)( − 1) − z ( + a)( −  + 1)  ,
U( +1),C
1,0 =  ( −  − 1)( −  − 2) − z ( + a)( −  + 1)( −  − 1),
U(−a),C
1,0 =  ( −  − 1)( + a) − z ( + a)( −  + 1)( + a + 1),
U( −1),C
1,0 =  ( −  − 1)( −  + 1) − z ( + a)( −  + 1)( −  + 2),
T,C
1,0 =  ( −  − 1)( − N − 1) − z ( + a)( −  + 1)( − N) .
(ii) Geronimus transformation (of type 2, 1):
G
1,0 =  ( −  ) − z ( + a)( −  ) .
C,G G,C
1,0 = 1,0 ,
2
G
1,0 =  ( − 1 ) ( − 2 ) − z ( + a)( − 1 ) ( − 2 ) ,
U(0),G
1,0 =  ( −  )( − 1) − z ( + a)( −  )  ,
U(−a),G
1,0 =  ( −  )( + a) − z ( + a)( −  )( + a + 1),
U( ),G
1,0 = G
1,0 ,
G,T
1,0 =  ( −  ) ( − N − 1) − z ( + a)( −  ) ( − N) .
(iii) Reduced-Uvarov transformation U (0) (of type 2, 1):

U(0)
1,0 =  ( − 1) − z ( + a)  .

C,U(0) U(0),C G,U(0) U(0),G
1,0 = 1,0 , 1,0 = 1,0 ,
U 2 (1,0)
1,0 =  ( − 1)( − 2) − z ( + a)  ( − 1),
U 2 (−a,0)
1,0 =  ( − 1)( + a) − z ( + a)  ( + a + 1),
T,U(0)
1,0 =  ( − 1)( − N − 1) − z ( + a)  ( − N).
(iv) Reduced-Uvarov transformation U (−a) (of type 2, 1):

U(−a)
1,0 =  ( + a) − z ( + a)( + a + 1).
C,U(−a) U(−a),C G,U(−a) U(−a),G U 2 (0,−a) U 2 (−a,0)
1,0 = 1,0 , 1,0 = 1,0 , 1,0 = 1,0 ,
U 2 (−a−1,−a)
1,0 =  ( + a)( + a + 1) − z ( + a)( + a + 1)( + a + 2),
T,U(−a)
1,0 =  ( + a)( − N − 1) − z ( + a)( + a + 1)( − N) .
(v) Truncation transformation (of type 2, 1):

T1,0 =  ( − N − 1) − z ( + a)( − N) .
U(0),T T,U(0) U(−a),T T,U(−a)
C,T T,C
1,0 = 1,0 , G,T T,G
1,0 = 1,0 , 1,0 = 1,0 , 1,0 = 1,0 ,
U(N+1),T
1,0 =  ( − N − 1)( − N − 2) − z ( + a)( − N) ( − N − 1),
U(N),T
1,0 =  ( − N − 1)( − N) − z ( + a)( − N) ( − N + 1).
(vi) Uvarov transformation (of type 3, 2):

U1,0 =  ( −  − 1)( −  ) − z ( + a)( −  ) ( −  + 1),  = 0, −a.
Generalized Charlier
0,1 =  ( + b) − z.
Transformations (of type 1, 2):
(i) Christoffel
C0,1 =  ( + b)( −  − 1) − z ( −  + 1),
(ii) Geronimus
G
0,1 =  ( + b)( −  ) − z ( −  ) ,
(iii) Reduced-Uvarov
U(0)
0,1 =  ( + b)( − 1) − z ,
U(−b)
0,1 =  ( + b)( + b − 1)− z ( + b),
(iv) Truncation
T0,1 =  ( + b)( − N − 1) − z ( − N) .
148 D. D OMINICI
Generalized Meixner
1,1 =  ( + b) − z ( + a).
(i) Christoffel
C1,1 =  ( + b)( −  − 1) − z ( + a)( −  + 1),
(ii) Geronimus
G
1,1 =  ( + b)( −  ) − z ( + a)( −  ),
U(0)
1,1 =  ( + b)( − 1) − z ( + a)  ,
U(−b)
1,1 =  ( + b)( + b − 1) − z ( + a)( + b),
U(−a)
1,1 =  ( + b)( + a) − z ( + a)( + a + 1),
(iv) Truncation
T1,1 =  ( + b)( − N − 1) − z ( + a)( − N) .
Generalized Krawtchouk
2,0;N =  − z ( − N) ( + a)
Transformations (of type 3, 1; N):
(i) Christoffel
C2,0;N =  ( −  − 1) − z ( − N) ( + a)( −  + 1),
(ii) Geronimus
G
2,0;N =  ( −  ) − z ( − N) ( + a)( −  ),
U(0)
2,0;N =  ( − 1) − z ( − N) ( + a)  ,
U(N)
2,0;N =  ( − N) − z ( − N) ( + a)( − N + 1),
U(−a)
2,0;N =  ( + a) − z ( − N) ( + a)( + a + 1).
Generalized Hahn functional of type I
2,1 =  ( + b) − z ( + a1 ) ( + a2) .
(i) Christoffel
C2,1 =  ( + b)( −  − 1) − z ( + a1) ( + a2 ) ( −  + 1),
(ii) Geronimus
G
2,1 =  ( + b)( −  ) − z ( + a1 ) ( + a2 ) ( −  ),
U(0)
2,1 =  ( + b)( − 1) − z ( + a1 ) ( + a2)  ,
U(−b)
2,1 =  ( + b)( + b − 1)− z ( + a1 ) ( + a2) ( + b),
U(−a )
2,1 1 =  ( + b)( + a1) − z ( + a1 ) ( + a2) ( + a1 + 1),
U(−a )
2,1 2 =  ( + b)( + a2) − z ( + a1 ) ( + a2) ( + a2 + 1),
(iv) Truncation
T2,1 =  ( + b)( − N − 1) − z ( + a1 ) ( + a2) ( − N).
5. Conclusion
We have studied the discrete functionals (6) characterized by the hypergeometric

differential equation satisfied by their first moment 0 (z) = L [1] ,
[ ( + b) − z ( + a)] [0 ] = 0, a ∈ Kp, b ∈ Kq .
We obtained recurrence relations for the moments on the monomial and falling factorial
polynomial bases, and gave examples for all functionals of class s 2, where s =
max {p − 1, q}.
We are currently working on further applications of our results to study some prop-
erties of the orthogonal polynomials themselves (representations, recurrence-relation
coefficients, generating functions, etc.).
Acknowledgements. This work was supported by the strategic program “Innova-

tives OÖ–2010 plus” from the Upper Austrian Government. We wish to thank Professor
Veronika Pillwein for her generous sponsorship.
150 D. D OMINICI
REFERENCES
[1] M. A LFARO AND R. Á LVAREZ N ODARSE, A characterization of the classical orthogonal discrete
and q -polynomials, J. Comput. Appl. Math. 201 (1), 48–54 (2007).
[2] R. Á LVAREZ N ODARSE , J. A RVES Ú , AND F. M ARCELL ÁN, Modifications of quasi-definite linear
functionals via addition of delta and derivatives of delta Dirac functions, Indag. Math. (N.S.) 15 (1),
1–20 (2004).
[3] R. Á LVAREZ N ODARSE , A. G. G ARC ÍA , AND F. M ARCELL ÁN, On the properties for modifications
of classical orthogonal polynomials of discrete variables, In “Proceedings of the International Confer-
ence on Orthogonality, Moment Problems and Continued Fractions (Delft, 1994)”, vol. 65, pp. 3–18
(1995).
[4] R. Á LVAREZ N ODARSE AND F. M ARCELL ÁN, Difference equation for modifications of Meixner poly-
nomials, J. Math. Anal. Appl. 194 (1), 250–258 (1995).
[5] R. Á LVAREZ N ODARSE AND F. M ARCELL ÁN, The modification of classical Hahn polynomials of a
discrete variable, Integral Transform. Spec. Funct. 3 (4), 243–262 (1995).
[6] R. Á LVAREZ N ODARSE AND J. P ETRONILHO, On the Krall-type discrete polynomials, J. Math. Anal.
Appl. 295 (1), 55–69 (2004).
[7] H. BAVINCK AND R. K OEKOEK, On a difference equation for generalizations of Charlier polynomi-
als, J. Approx. Theory 81 (2), 195–206 (1995).
[8] H. BAVINCK AND H. VAN H AERINGEN, Difference equations for generalized Meixner polynomials,
J. Math. Anal. Appl. 184 (3), 453–463 (1994).
[9] S. B ELMEHDI , On semi-classical linear functionals of class s = 1 . Classification and integral repre-
sentations, Indag. Math. (N.S.) 3 (3), 253–275 (1992).
[10] L. B OELEN , G. F ILIPUK , AND W. VAN A SSCHE, Recurrence coefficients of generalized Meixner
polynomials and Painlevé equations, J. Phys. A 44 (3), 035202, 19 (2011).
[11] M. I. B UENO AND F. M ARCELL ÁN, Darboux transformation and perturbation of linear functionals,
Linear Algebra Appl. 384, 215–242 (2004).
[12] P. L. B UTZER AND T. H. K OORNWINDER, Josef Meixner: his life and his orthogonal polynomials,
Indag. Math. (N.S.) 30 (1), 250–264 (2019).
[13] K. C ASTILLO , F. R. R AFAELI , AND A. S UZUKI , Stieltjes’ theorem for classical discrete orthogonal
polynomials, J. Math. Phys. 61 (10), 103505, 16 (2020).
[14] C. V. L. C HARLIER, Über die Darstellung willkürlicher Funktionen, Ark. Mat., Astr. Fys. 2 (20), 35
(1905–1906).
[15] T. S. C HIHARA, Orthogonal polynomials and measures with end point masses, Rocky Mountain J.
Math. 15 (3), 705–719 (1985).
[16] P. A. C LARKSON, Recurrence coefficients for discrete orthonormal polynomials and the Painlevé
equations, J. Phys. A 46 (18), 185205, 18 (2013).
[17] D. D ICKINSON, On Lommel and Bessel polynomials, Proc. Amer. Math. Soc. 5, 946–956 (1954).
[18] D. D OMINICI , Asymptotic analysis of the Askey-scheme. I. From Krawtchouk to Charlier, Cent. Eur.
J. Math. 5 (2), 280–304 (2007).
[19] D. D OMINICI , Asymptotic analysis of the Krawtchouk polynomials by the WKB method, Ramanujan
J. 15 (3), 303–338 (2008).
[20] D. D OMINICI , Laguerre-Freud equations for generalized Hahn polynomials of type I, J. Difference
Equ. Appl. 24 (6), 916–940 (2018).
[21] D. D OMINICI , Recurrence relations for the moments of discrete semiclassical orthogonal polynomials,
J. Class. Anal. 20 (2), 143–180 (2022).
[22] D. D OMINICI AND F. M ARCELL ÁN, Discrete semiclassical orthogonal polynomials of class one,
Pacific J. Math. 268 (2), 389–411 (2014).
[23] D. D OMINICI AND F. M ARCELL ÁN, Discrete semiclassical orthogonal polynomials of class 2, In
“Orthogonal polynomials: current trends and applications”, vol. 22 of “SEMA SIMAI Springer Ser.”,
pp. 103–169, Springer, Cham ([2021] 2021).
c
[24] A. J. D UR ÁN, Orthogonal polynomials satisfying higher-order difference equations, Constr. Approx.
36 (3), 459–486 (2012).
[25] A. J. D UR ÁN, From Krall discrete orthogonal polynomials to Krall polynomials, J. Math. Anal. Appl.
450 (2), 888–900 (2017).
[26] A. J. D UR ÁN, The algebras of difference operators associated to Krall-Charlier orthogonal polyno-
mials, J. Approx. Theory 234, 64–81 (2018).
[27] A. J. D UR ÁN, Christoffel transform of classical discrete measures and invariance of determinants
of classical and classical discrete polynomials, J. Math. Anal. Appl. 503 (2), Paper No. 125306, 29
(2021).
[28] P. F EINSILVER , J. M C S ORLEY, AND R. S CHOTT, Combinatorial interpretation and operator calcu-
lus of Lommel polynomials, J. Combin. Theory Ser. A 75 (1), 163–171 (1996).
[29] G. F ILIPUK AND W. VAN A SSCHE, Recurrence coefficients of a new generalization of the Meixner
polynomials, SIGMA Symmetry Integrability Geom. Methods Appl. 7, Paper 068, 11 (2011).
[30] G. F ILIPUK AND W. VAN A SSCHE, Discrete orthogonal polynomials with hypergeometric weights
and Painlevé VI, SIGMA Symmetry Integrability Geom. Methods Appl. 14, Paper No. 088, 19 (2018).
[31] M. F OUPOUAGNIGNI , M. N. H OUNKONNOU , AND A. RONVEAUX, Laguerre-Freud equations for
the recurrence coefficients of D semi-classical orthogonal polynomials of class one, In “Proceedings
of the VIIIth Symposium on Orthogonal Polynomials and Their Applications (Seville, 1997)”, vol. 99,
pp. 143–154 (1998).
[32] A. G. G ARC ÍA , F. M ARCELL ÁN , AND L. S ALTO, A distributional study of discrete classical orthog-
onal polynomials, In “Proceedings of the Fourth International Symposium on Orthogonal Polynomials
and their Applications (Evian-Les-Bains, 1992)”, vol. 57, pp. 147–162 (1995).
[33] W. G AUTSCHI , On generating orthogonal polynomials, SIAM J. Sci. Statist. Comput. 3 (3), 289–317
(1982).
[34] E. G ODOY, F. M ARCELL ÁN , L. S ALTO , AND A. Z ARZO, Perturbations of discrete semiclassical
functionals by Dirac masses, Integral Transform. Spec. Funct. 5 (1–2), 19–46 (1997).
[35] M. N. H OUNKONNOU , C. H OUNGA , AND A. RONVEAUX, Discrete semi-classical orthogonal poly-
nomials: generalized Charlier, J. Comput. Appl. Math. 114 (2), 361–366 (2000).
[36] S. K ARLIN AND J. L. M C G REGOR, The Hahn polynomials, formulas and an application, Scripta
Math. 26, 33–46 (1961).
[37] M. P. K RAVCHUK, Sur une généralisation des polynômes d’Hermite, C. R. Acad. Sci. Paris 189,
620–622 (1929).
[38] K. F. L EE AND R. W ONG, Asymptotic expansion of the modified Lommel polynomials hn, (x) and
their zeros, Proc. Amer. Math. Soc. 142 (11), 3953–3964 (2014).
[39] M. M A ÑAS , Pearson equations for discrete orthogonal polynomials: III – Christoffel and Geronimus
transformations, Rev. R. Acad. Cienc. Exactas Fı́s. Nat. Ser. A Mat. RACSAM 116 (4), Paper No.
168, 23 (2022).
[40] M. M A ÑAS , I. F ERN ÁNDEZ -I RISARRI , AND O. F. G ONZ ÁLEZ -H ERN ÁNDEZ, Pearson equations for
discrete orthogonal polynomials: I. Generalized hypergeometric functions and Toda equations, Stud.
Appl. Math. 148 (3), 1141–1179 (2022).
[41] I. G. M ACDONALD, Symmetric functions and Hall polynomials, Oxford Classic Texts in the Physical
Sciences. The Clarendon Press, Oxford University Press, New York, second ed. (2015).
[42] D. M AKI , On constructing distribution functions: With applications to Lommel polynomials and
Bessel functions, Trans. Amer. Math. Soc. 130, 281–297 (1968).
[43] F. M ARCELL ÁN AND L. S ALTO, Discrete semi-classical orthogonal polynomials, J. Differ. Equations
Appl. 4 (5), 463–496 (1998).
[44] P. M ARONI , Sur la suite de polynômes orthogonaux associée à la forme u = c d +  (x − c)−1 L ,
Period. Math. Hungar. 21 (3), 223–248 (1990).
[45] P. M ARONI AND M. M EJRI , The symmetric D -semi-classical orthogonal polynomials of class one,
Numer. Algorithms 49 (1–4), 251–282 (2008).
[46] J. M EIXNER, Orthogonale Polynomsysteme Mit Einer Besonderen Gestalt Der Erzeugenden Funktion,
J. London Math. Soc. 9 (1), 6–13 (1934).
[47] A. F. N IKIFOROV, S. K. S USLOV, AND V. B. U VAROV, Classical orthogonal polynomials of a dis-
crete variable, Springer Series in Computational Physics. Springer-Verlag, Berlin (1991).
[48] P. N JIONOU S ADJANG , W. K OEPF, AND M. F OUPOUAGNIGNI , On moments of classical orthogonal
polynomials, J. Math. Anal. Appl. 424 (1), 122–151 (2015).
[49] F. W. J. O LVER , D. W. L OZIER , R. F. B OISVERT, AND C. W. C LARK (eds.), NIST handbook of
mathematical functions, U.S. Department of Commerce, National Institute of Standards and Technol-
ogy, Washington, DC; Cambridge University Press, Cambridge (2010).
152 D. D OMINICI
[50] A. RONVEAUX , S. B ELMEHDI , E. G ODOY, AND A. Z ARZO, Recurrence relation approach for con-
nection coefficients. Applications to classical discrete orthogonal polynomials, In “Symmetries and
integrability of difference equations (Estérel, PQ, 1994)”, vol. 9 of “CRM Proc. Lecture Notes”, pp.
319–335. Amer. Math. Soc., Providence, RI (1996).
[51] A. RONVEAUX AND L. S ALTO, Discrete orthogonal polynomials – polynomial modification of a
classical functional, J. Differ. Equations Appl. 7 (3), 323–344 (2001).
[52] M. S GHAIER AND M. Z AATRA, A large family of Dw -semiclassical polynomials of class one, Integral
Transforms Spec. Funct. 28 (5), 386–402 (2017).
[53] M. S GHAIER AND M. Z AATRA, A family of discrete semi-classical orthogonal polynomials of class
one, Period. Math. Hungar. 76 (1), 68–87 (2018).
[54] C. S MET AND W. VAN A SSCHE, Orthogonal polynomials on a bi-lattice, Constr. Approx. 36 (2),
215–242 (2012).
[55] L. T RUKSA, Hypergeometric orthogonal systems of polynomials. i, Aktuárské Vědy 2 (2), 65–84
(1931).
[56] L. T RUKSA, Hypergeometric orthogonal systems of polynomials. ii, Aktuárské Vědy 2 (3), 113–144
(1931).
[57] L. T RUKSA, Hypergeometric orthogonal systems of polynomials. iii, Aktuárské Vědy 2 (4), 177–203
(1931).
[58] W. VAN A SSCHE AND M. F OUPOUAGNIGNI, Analysis of non-linear recurrence relations for the
recurrence coefficients of generalized Charlier polynomials, J. Nonlinear Math. Phys. 10 (suppl. 2),
231–237 (2003).
[59] L. V ERDE -S TAR, Recurrence coefficients and difference equations of classical discrete orthogonal
and q -orthogonal polynomial sequences, Linear Algebra Appl. 440, 293–306 (2014).
[60] L. V ERDE -S TAR, A unified construction of all the hypergeometric and basic hypergeometric families
of orthogonal polynomial sequences, Linear Algebra Appl. 627, 242–274 (2021).
[61] L. V INET AND A. Z HEDANOV, Hypergeometric orthogonal polynomials with respect to Newtonian
bases, SIGMA Symmetry Integrability Geom. Methods Appl. 12, Paper No. 048, 14 (2016).
(Received March 10, 2022) Diego Dominici

Research Institute for Symbolic Computation
Johannes Kepler University Linz
Altenberger Straße 69, 4040 Linz, Austria
and
Department of Mathematics
State University of New York at New Paltz
1 Hawk Dr., New Paltz, NY, 12561-2443, USA
e-mail: ddominic@risc.jku.at
dominicd@newpaltz.edu
Journal of Classical Analysis

www.ele-math.com
jca@ele-math.com

Jca 21 08

Uploaded by

Copyright:

Available Formats

Jca 21 08

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Jca 21 08

Uploaded by

Copyright:

Available Formats

J ournal of

RECURRENCE RELATIONS FOR THE MOMENTS OF

Let K be a field (we mostly think of K as R or C) and N0 be the set of nonneg-

and  : F → F be the differential operator defined by [49, 16.8.2]

and for m = 0 we understand that K0 = 0/ and

We denote by k,n the Kronecker delta

Mathematics subject classification (2020): Primary 33C47; Secondary 33C20, 11B37.

and for c ∈ Km we will use the notation [49, 16.1]

(c)n = (c1 )n · · · (cm )n , (0)

and we assume that x and z are independent variables.

 (x) = x (x + b),  (x) = (x + a), (3)

then the class s of the functional L p,q is given by

s = max {deg ( ) − 1, deg( ) − 1} = o − 1, (4)

and semiclassical functionals of class s = 0 are called classical. In [22], we classified

it follows using linearity that for all u ∈ K [x] ,

u ( ) [zx f ] = zx u ( + x)[ f ] , f ∈ F. (5)

Let L : K[x] → F be a discrete functional

where  : N0 → K is a given function. If we set f = 1 in (5), we get

u ( ) [zx ] = u (x) zx , u ∈ K [x]

n = L [n ] = n ( ) L [1] = n ( ) [0 ] , n ∈ N0 . (8)

2.1. Newtonian bases

 [n ] = (S + n ) [n ] , (10)

where S denotes the shift operator in n

S [cn ] = cn+1 . (11)

R EMARK 1. From (10) and (11) we see that

 [n+1 ] = n+2 + n+1n+1 = S [n+1 + n n ] = S (S + n ) [n ] , (12)

which is different from

(S + n ) S [n ] = (S + n ) [n+1 ] = n+2 + n n+1,

so caution must be exercised when n depends on n .

n+1 (x) = (x − n) n (x) = xn (x − 1), n 0, (16)

n (z) = L p,q [xn ] , n ∈ N0 ,

n (z) = L p,q [n ] , n ∈ N0 .

we obtain the hypergeometric representation

Multiplying the ODE  p,q [y] = 0 by  n and using (5), we get

 n+1 ( + b) − z ( + 1)n ( + a) [0 ] = 0. (17)

 p,q = S n+1 (S + b) − z (S + 1)n (S + a). (18)

The polynomials (x + c) can be written in the monomial basis as

of order n + s + 1, where s is the class of the functional L p,q defined in (4).

 p,q = S (S + n + b) − z (S + n + a). (20)

and from (12) we conclude that

Using (21), we can rewrite  p,q in extended form

Using (9) and (22) we obtain

nC = LCp,q [n ] = L p,q [(x −  )n ] = n+1 + (n −  ) n ,

the first moment has the hypergeometric representation

Similarly, (20) and (25) give Cp,q nC = 0, with

Cp,q = S (S + n + b)(S + n −  − 1) − z (S + n + a)(S + n −  + 1).

The differential operator Up,q is defined by

Up,q = ( −  ) ( −  − 1) p,q, (29)

Up,q = G,C C,G

in the sense of applying a double transformation to the operator  p,q .

Up,q = (S + n −  )(S + n −  − 1) p,q .

R EMARK 2. Because  −  annihilates the function z , we have Gp,q [z ] = 0

If  ( ) = 0, then the differential operator

is called a reduced-Uvarov transformation, since

4) The truncation transformation at N , which we define by

The first moment admits the hypergeometric representation

Tp,q =  ( ) ( − N − 1) − z ( ) ( − N) = ( − N − 1) p,q . (30)

Similarly, (20) and (30) give Tp,q nT = 0, with

Tp,q = (S + n − N) p,q .