1 Introduction

Let { A n , k } n , k = 0 be a bounded double sequence of real or complex numbers, let [a] denote the greatest integer in aR, and let ( λ ) ν , ( λ ) 0 1, denote the Pochhammer symbol defined by

( λ ) ν := Γ ( λ + ν ) Γ ( λ )

by means of familiar gamma functions. In 1972, Srivastava [1] introduced the following family of polynomials:

S n N (z):= k = 0 [ n N ] ( n ) N k k ! A n , k z k ( n N 0 = N { 0 } ; N N ) ,
(1)

where ℕ is the set of positive integers.

Afterward, González et al. [2] extended the Srivastava polynomials S n N (z) as follows:

S n , m N (z):= k = 0 [ n N ] ( n ) N k k ! A n + m , k z k (m,n N 0 ;NN)
(2)

and investigated their properties extensively. Motivated essentially by the definitions (1) and (2), scientists investigated and studied various classes of Srivastava polynomials in one and more variables.

In [3], the following family of bivariate polynomials was introduced:

S n m , N (x,y):= k = 0 [ n N ] A m + n , k x n N k ( n N k ) ! y k k ! (n,m N 0 ,NN),

and it was shown that the polynomials S n m , N (x,y) include many well-known polynomials such as Lagrange-Hermite polynomials, Lagrange polynomials and Hermite-Kampé de Feriét polynomials.

In [4], Srivastava et al. introduced the three-variable polynomials

(3)

where { A m , n , k } is a triple sequence of complex numbers. Suitable choices of { A m , n , k } in equation (3) give a three-variable version of well-known polynomials (see also [5]). Recently, in [6], the multivariable extension of the Srivastava polynomials in r-variable was introduced

(4)

where { A m , k r 1 , k 1 , k 2 , , k r 2 } is a sequence of complex numbers.

In this paper we introduce the two-parameter Srivastava polynomials in one and more variables by inserting new indices. These polynomials include the family of polynomials which were introduced and/or investigated in [14, 6, 7] and [8]. We prove several two-sided linear generating relations and obtain various series identities for these polynomials. Furthermore, we exhibit some illustrative consequences of the main results for some well-known special polynomials which are contained by the two-parameter Srivastava polynomials.

2 Two-parameter one-variable Srivastava polynomials

In this section we introduce the following family of two-parameter one-variable polynomials:

S n m 1 , m 2 (x):= k = 0 n ( n ) k k ! A m 1 + m 2 + n , m 2 + k x k ( m 1 , m 2 ,n,k N 0 ),
(5)

where { A n , k } is a bounded double sequence of real or complex numbers. Note that appropriate choices of the sequence A n , k give one-variable versions of the well-known polynomials.

Remark 2.1 Choosing A m , n = ( α m ) n (m,n N 0 ) in (5), we get

S n m 1 , m 2 ( 1 x ) = ( 1 ) m 2 ( α + m 1 + n + 1 ) m 2 n ! ( x ) n L n ( α + m 1 ) (x),

where L n ( α ) (x) are the classical Laguerre polynomials given by

L n ( α ) (x)= ( x ) n n ! 2 F 0 ( n , α n ; ; 1 x ) .

Remark 2.2

Setting

A m , n = ( α + β + 1 ) 2 m ( β m ) n ( α + β + 1 ) m ( α β 2 m ) n (m,n N 0 )

in (5), we obtain

where P n ( α , β ) (x) are the classical Jacobi polynomials.

Remark 2.3 If we set A m , n = ( α + m 1 ) n (m,n N 0 ) in (5), then we get

S n m 1 , m 2 ( x β ) = ( α + m 1 + m 2 + n 1 ) m 2 y n (x,α+ m 1 +2 m 2 ,β)(β0),

where y n (x,α,β) are the classical Bessel polynomials given by

y n (x,α,β) = 2 F 0 ( n , α + n 1 ; ; x β ) .

Theorem 2.4 Let { f ( n ) } n = 0 be a bounded sequence of complex numbers. Then

(6)

provided each member of the series identity (6) exists.

Proof Let the left-hand side of (6) be denoted by Ψ(x). Then, using the definition of S n m 1 , m 2 (x) on the left-hand side of (6), we have

Ψ ( x ) = m 1 , m 2 , n = 0 f ( n + m 1 + m 2 ) k = 0 n ( n ) k k ! A m 1 + m 2 + n , m 2 + k x k w 1 m 1 m 1 ! w 2 m 2 m 2 ! t n n ! = m 1 , m 2 , n = 0 f ( n + m 1 + m 2 ) k = 0 n 1 k ! A m 1 + m 2 + n , m 2 + k ( x ) k w 1 m 1 m 1 ! w 2 m 2 m 2 ! t n ( n k ) ! = m 1 , m 2 , n , k = 0 f ( n + m 1 + m 2 + k ) ( x t ) k k ! A m 1 + m 2 + n + k , m 2 + k w 1 m 1 m 1 ! w 2 m 2 m 2 ! t n n ! .

Let m 1 m 1 n,

Ψ ( x ) = m 1 , m 2 , k = 0 f ( m 1 + m 2 + k ) ( x t ) k k ! A m 1 + m 2 + k , m 2 + k ( n = 0 m 1 w 1 m 1 n t n ( m 1 n ) ! n ! ) w 2 m 2 m 2 ! = m 1 , m 2 , k = 0 f ( m 1 + m 2 + k ) ( x t ) k k ! A m 1 + m 2 + k , m 2 + k ( n = 0 m 1 ( m 1 n ) w 1 m 1 n t n m 1 ! ) w 2 m 2 m 2 ! = m 1 , m 2 , k = 0 f ( m 1 + m 2 + k ) A m 1 + m 2 + k , m 2 + k ( w 1 + t ) m 1 m 1 ! w 2 m 2 m 2 ! ( x t ) k k ! .

Let m 2 m 2 k,

Ψ ( x ) = m 1 , m 2 = 0 f ( m 1 + m 2 ) A m 1 + m 2 , m 2 ( w 1 + t ) m 1 m 1 ! k = 0 m 2 w 2 m 2 k ( x t ) k ( m 2 k ) ! k ! = m 1 , m 2 = 0 f ( m 1 + m 2 ) A m 1 + m 2 , m 2 ( w 1 + t ) m 1 m 1 ! ( w 2 + ( x t ) ) m 2 m 2 ! .

 □

Remark 2.5 Choosing A m , n = ( α m ) n (m,n N 0 ) and x 1 x , then by Theorem 2.4, we get

If we set w 2 = t x and f=1 into the above equation, then we have

Remark 2.6

Setting

A m , n = ( α + β + 1 ) 2 m ( β m ) n ( α + β + 1 ) m ( α β 2 m ) n (m,n N 0 )

and x 2 1 + x in equation (6), we have

Remark 2.7 If we set A m , n = ( α + m 1 ) n (m,n N 0 ) and x x β in (6), then we can write

If we set f=1 and w 2 = x β t, then

Remark 2.8 If we consider Remarks 2.5 and 2.7, we get the following relation between Laguerre polynomials L n ( α ) (x) and Bessel polynomials y n (x,α,β):

3 Two-parameter two-variable Srivastava polynomials

In this section we introduce the following two-parameter family of bivariate polynomials:

S n m 1 , m 2 (x,y):= k = 0 n A m 1 + m 2 + n , m 2 + k x k k ! y n k ( n k ) ! ( m 1 , m 2 ,n,k N 0 ),
(7)

where { A n , k } is a bounded double sequence of real or complex numbers. Note that in the particular case these polynomials include the Lagrange polynomials.

Remark 3.1 Choosing A m , n = ( α ) m n ( β ) n (m,n N 0 ) in (7), we have

S n m 1 , m 2 (x,y)= ( α ) m 1 ( β ) m 2 g n ( α + m 1 , β + m 2 ) (x,y),

where g n ( α , β ) (x,y) are the Lagrange polynomials given by

g n ( α , β ) (x,y)= k = 0 n ( α ) n k ( β ) k ( n k ) ! k ! x k y n k .

Using similar techniques as in the proof of Theorem 2.4, we get the following theorem.

Theorem 3.2 Let { f ( n ) } n = 0 be a bounded sequence of complex numbers. Then

(8)

provided each member of the series identity (8) exists.

Remark 3.3 If we set A m , n = ( α ) m n ( β ) n (m,n N 0 ) in (8), we have

Choosing f=1 gives

Furthermore, since we have the relation between P n ( α , β ) (x,y) Jacobi polynomials and Lagrange polynomials [9] as

g n ( α , β ) (x,y)= ( y x ) n P n ( α n , β n ) ( x + y x y ) ,

we get the following generating relation for Jacobi polynomials P n ( α , β ) (x,y):

4 Two-parameter three-variable Srivastava polynomials

In this section we define two-parameter three-variable Srivastava polynomials as follows:

(9)

where { A n , k , l } n , k = 0 is a bounded triple sequence of real or complex numbers.

Using similar techniques as in the proof of Theorem 2.4, we get the following theorem.

Theorem 4.1 Let { f ( n ) } n = 0 be a bounded sequence of complex numbers. Then

(10)

provided each member of the series identity (10) exists.

Theorem 4.2 Let { f ( n ) } n = 0 be a bounded sequence of complex numbers, and let S n m 1 , m 2 , M (x,y,z) be defined by (9). Suppose also that two-parameter two-variable polynomials P m 1 , m 2 M (x,y) are defined by

P m 1 , m 2 M (x,y)= l = 0 [ m 2 M ] A m 1 + m 2 , m 2 , l x m 2 M l ( m 2 M l ) ! y l l ! .
(11)

Then the family of two-sided linear generating relations holds true between the two-parameter three-variable Srivastava polynomials S n m 1 , m 2 , M (x,y,z) and P m 1 , m 2 M (x,y):

(12)

Suitable choices of A n , k , l in equations (9) and (11) give some known polynomials.

Remark 4.3 Choosing M=2 and A m , n , k = ( α ) m n ( γ ) n 2 k ( β ) k (m,n N 0 ) in (9), we get

S n m 1 , m 2 , 2 (x,y,z)= ( α ) m 1 ( γ ) m 2 u n ( α + m 1 , β , γ + m 2 ) (z,x,y),

where u n ( α , β , γ ) (x,y,z) is the polynomial given by

u n ( α , β , γ ) (x,y,z)= k = 0 n l = 0 [ k 2 ] ( β ) l ( γ ) k 2 l ( α ) n k y l l ! x n k ( n k ) ! z k 2 l ( k 2 l ) ! .

Now, by setting M=1 and A m , n , k = ( α ) k ( β ) n k ( γ ) m n (m,n N 0 ) in the definition (9), we obtain

S n m 1 , m 2 , 1 (x,y,z)= ( γ ) m 1 ( β ) m 2 g n ( α , β + m 2 , γ + m 1 ) (x,y,z),

where g n ( α , β , γ ) (x,y,z) are the Lagrange polynomials given by

g n ( α , β , γ ) (x,y,z)= k = 0 n l = 0 k ( α ) l ( β ) k l ( γ ) n k x l l ! y k l ( k l ) ! z n k ( n k ) ! .

Remark 4.4 If we set M=2 and A m , n , k = ( α ) m n ( γ ) n 2 k ( β ) k (m,n N 0 ) in (11), then

P m 1 , m 2 2 (x,y)= ( α ) m 1 h m 2 ( γ , β ) (x,y),

where h m 2 ( γ , β ) (x,y) denotes the Lagrange-Hermite polynomials given explicitly by

h m 2 ( γ , β ) (x,y)= l = 0 [ m 2 2 ] ( γ ) m 2 2 l ( β ) l x m 2 2 l ( m 2 2 l ) ! y l l ! .

Furthermore, choosing M=1 and A m , n , k = ( α ) k ( β ) n k ( γ ) m n (m,n N 0 ) in the definition (11), we have

P m 1 , m 2 1 (x,y)= ( γ ) m 1 g m 2 ( β , α ) (x,y),

where g m 2 ( α , β ) (x,y) are the Lagrange polynomials given by

g m 2 ( α , β ) (x,y)= l = 0 m 2 ( α ) m 2 l ( β ) l x m 2 l ( m 2 l ) ! y l l ! .

Remark 4.5 If we set w 1 zt and w 2 yt in Theorem 4.1, then we get

(13)

Now, if we set f=1, M=2 and A m , n , k = ( α ) m n ( γ ) n 2 k ( β ) k (m,n N 0 ) in (13), then

m 1 , m 2 , n = 0 ( α ) m 1 ( γ ) m 2 u n ( α + m 1 , β , γ + m 2 ) (z,x,y) ( z t ) m 1 m 1 ! ( y t ) m 2 m 2 ! t n = ( 1 x t 2 ) β .

Furthermore, if we set M=2 and A m , n , k = ( α ) m n ( γ ) n 2 k ( β ) k (m,n N 0 ) in (12), then