Open Access

Some new identities of Bernoulli, Euler and Hermite polynomials arising from umbral calculus

  • Dae San Kim1,
  • Taekyun Kim2Email author,
  • Dmitry V Dolgy3 and
  • Seog-Hoon Rim4
Advances in Difference Equations20132013:73

DOI: 10.1186/1687-1847-2013-73

Received: 3 January 2013

Accepted: 6 March 2013

Published: 26 March 2013

Abstract

In this paper, we derive the identities of higher-order Bernoulli, Euler and Frobenius-Euler polynomials from the orthogonality of Hermite polynomials. Finally, we give some interesting and new identities of several special polynomials arising from umbral calculus.

MSC: 05A10, 05A19.

Keywords

Bernoulli polynomial Euler polynomial Abel polynomial

1 Introduction

The Hermite polynomials are defined by the generating function to be
e 2 x t t 2 = e H ( x ) t = n = 0 H n ( x ) t n n !
(1.1)
with the usual convention about replacing H n ( x ) by H n ( x ) (see [1]). In the special case, x = 0 , H n ( 0 ) = H n are called the nth Hermite numbers. From (1.1) we have
H n ( x ) = ( H + 2 x ) n = l = 0 n ( n l ) H n l x l 2 l .
(1.2)
Thus, by (1.2), we get
d k d x k H n ( x ) = 2 k ( n ) k H n k ( x ) = 2 k n ! ( n k ) ! H n k ( x ) ,
(1.3)

where ( x ) k = x ( x 1 ) ( x k + 1 ) .

As is well known, the Bernoulli polynomials of order r are defined by the generating function to be
( t e t 1 ) r e x t = n = 0 B n ( r ) ( x ) t n n ! ( r R ) .
(1.4)

In the special case, x = 0 , B n ( r ) ( 0 ) = B n ( r ) are called the n th Bernoulli numbers of order r (see [14]).

The Euler polynomials of order r are also defined by the generating function to be
( 2 e t + 1 ) r e x t = n = 0 E n ( r ) ( x ) t n n ! ( r R ) .
(1.5)

In the special case, x = 0 , E n ( r ) ( 0 ) = E n ( r ) are called the nth Euler numbers of order r.

For λ ( 1 ) C , the Frobenius-Euler polynomials of order r are given by
( 1 λ e t λ ) r e x t = n = 0 H n ( r ) ( x | λ ) t n n ! ( r R ) .
(1.6)

In the special case, x = 0 , H n ( r ) ( 0 | λ ) = H n ( r ) ( λ ) are called the nth Frobenius-Euler numbers of order r (see [116]).

The Stirling numbers of the first kind are defined by the generating function to be
( x ) n = k = 0 n S 1 ( n , k ) x k (see [11, 14]) ,
(1.7)
and the Stirling numbers of the second kind are given by
( e t 1 ) n = n ! l = n S 2 ( l , n ) t l l ! (see [14]) .
(1.8)
In [1] it is known that H 0 ( x ) , H 1 ( x ) , , H n ( x ) from an orthogonal basis for the space
P n = { p ( x ) Q [ x ] | deg p ( x ) n }
(1.9)
with respect to the inner product
p 1 ( x ) , p 2 ( x ) = e x 2 p 1 ( x ) p 2 ( x ) d x (see [1]) .
(1.10)
For p ( x ) P n , let us assume that
p ( x ) = k = 0 n C k H k ( x ) .
(1.11)
Then, from the orthogonality of Hermite polynomials and Rodrigues’ formula, we have
C k = 1 2 k k ! π e x 2 H k ( x ) p ( x ) d x = ( 1 ) k 2 k k ! π ( d k d x k e x 2 ) p ( x ) d x (see [1]) .
(1.12)
In particular, for p ( x ) = x m ( m 0 ), we easily get
( d n d x n e x 2 ) x m d x = { 0 if  n > m  or  n m  with  m n 0 ( mod 2 ) , ( 1 ) n m ! π 2 m n ( m n 2 ) ! if  n m  with  m n 0 ( mod 2 ) .
(1.13)
Let be the set of all formal power series in the variable t over with
F = { f ( t ) = k = 0 a k k ! t k | a k C } .
(1.14)
Let us assume that is the algebra of polynomials in the variable x over and that P is the vector space of all linear functionals on . L | p ( x ) denotes the action of the linear functional L on polynomials p ( x ) , and we remind that the vector space structure on P is defined by
L + M | p ( x ) = L | p ( x ) + M | p ( x ) , c L | p ( x ) = c L | p ( x ) ,

where c is a complex constant (see [2, 11, 14]).

The formal power series
f ( t ) = k = 0 a k k ! t k F
(1.15)
defines a linear functional on by setting
f ( t ) | x n = a n for all  n Z + = N { 0 } .
(1.16)
Thus, by (1.15) and (1.16), we get
t k | x n = n ! δ n , k ( n , k 0 ) ,
(1.17)

where δ n , k is the Kronecker symbol (see [2, 11, 14]).

Let f L ( t ) = k = 0 L | x k k ! t k . By (1.16), we get
f L ( t ) | x n = L | x n , n 0 .
(1.18)

Thus, by (1.18), we see that f L ( t ) = L . The map L f L ( t ) is a vector space isomorphism from P onto . Henceforth, will be thought of as both a formal power series and a linear functional. We call the umbral algebra. The umbral calculus is the study of umbral algebra (see [2, 11, 14]).

The order o ( f ( t ) ) of the nonzero power series f ( t ) is the smallest integer k for which the coefficient of t k does not vanish. A series f ( t ) having o ( f ( t ) ) = 1 is called a delta series, and a series f ( t ) having o ( f ( t ) ) = 0 is called an invertible series (see [2, 11, 14]). By (1.16) and (1.17), we see that e y t | p ( x ) = p ( y ) . For f ( t ) F and p ( x ) P , we have
f ( t ) = k = 0 f ( t ) | x k k ! t k , p ( x ) = k = 0 t k | p ( x ) k ! x k .
(1.19)
Let f ( t ) , g ( t ) F and p ( x ) P . Then we easily see that
f ( t ) g ( t ) | p ( x ) = f ( t ) | g ( t ) p ( x ) = g ( t ) | f ( t ) p ( x ) .
(1.20)
From (1.19), we can derive the following equation:
p ( k ) ( 0 ) = t k | p ( x ) and 1 | p ( k ) ( x ) = p ( k ) ( 0 ) .
(1.21)
Thus, by (1.21), we get
t k p ( x ) = p ( k ) ( x ) = d k p ( x ) d x k (see [2, 11, 14]) .
(1.22)
Let f ( t ) be a delta series, and let g ( t ) be an invertible series. Then there exists a unique sequence S n ( x ) of polynomials with g ( t ) f ( t ) k | S n ( x ) = n ! δ n , k , where n , k 0 (see [2, 11, 14]). The sequence S n ( x ) is called Sheffer sequence for ( g ( t ) , f ( t ) ) , which is denoted by S n ( x ) ( g ( t ) , f ( t ) ) . For f ( t ) F and p ( x ) P , we have
e y t 1 t | p ( x ) = 0 y p ( u ) d u , e y t 1 | p ( x ) = p ( y ) p ( 0 ) ,
(1.23)
and
f ( t ) | x p ( x ) = f ( t ) | p ( x ) .
(1.24)

In this paper, we introduce the identities of several special polynomials which are derived from the orthogonality of Hermite polynomials. Finally, we give some new and interesting identities of the higher-order Bernoulli, Euler and Frobenius-Euler polynomials arising from umbral calculus.

2 Some identities of several special polynomials

From (1.5), we note that
( 2 e t + 1 ) r = ( 1 + e t 1 2 ) r = j = 0 ( r j ) ( e t 1 2 ) j .
(2.1)
By (2.1), we get
( 2 e t + 1 ) r e x t = j = 0 ( r j ) ( e t 1 2 ) j e x t = n = 0 ( j = 0 n ( r j ) ( e t 1 2 ) j x n ) t n n ! .
(2.2)
From (1.5) and (2.2), we have
E n ( r ) ( x ) = j = 0 n ( r j ) 2 j ( e t 1 ) j x n .
(2.3)
By (1.8) and (1.9), we get
( e t 1 ) j x n = k = 0 n j t k | ( e t 1 ) j x n k ! = k = 0 n j ( e t 1 ) j | t k x n k ! x k = j ! k = 0 n j ( n k ) ( e t 1 ) j | x n k j ! x k = j ! k = 0 n j ( n j ) S 2 ( n k , j ) x k = j ! k = j n ( n k ) S 2 ( k , j ) x n k .
(2.4)

Therefore, by (2.3) and (2.4), we obtain the following theorem.

Theorem 2.1 For n 0 , we have
E n ( r ) ( x ) = 0 j n j k n ( n k ) ( r j ) j ! 2 j S 2 ( k , j ) x n k = 0 k n ( n k ) [ 0 j k ( r j ) j ! 2 j S 2 ( k , j ) ] x n k .
By (1.5), we easily see that
E n ( r ) ( x ) = k = 0 n ( n k ) E k ( r ) x n k .
(2.5)

Therefore, by Theorem 2.1 and (2.5), we obtain the following corollary.

Corollary 2.2 For k 0 , we have
E k ( r ) = j = 0 k ( r j ) j ! 2 j S 2 ( k , j ) .
Let us take p ( x ) = E n ( r ) ( x ) P n . Then, by (1.11), we get
E n ( r ) ( x ) = k = 0 n C k H k ( x ) .
(2.6)
From (1.12), we can derive the computation of C k as follows:
C k = ( 1 ) k 2 k k ! π ( d k e x 2 d x k ) E n ( r ) ( x ) d x ,
(2.7)
where
( d k e x 2 d x k ) E n ( r ) ( x ) d x = ( n ) ( ( n 1 ) ) ( ( n k + 1 ) ) e x 2 E n k ( r ) ( x ) d x = ( 1 ) k n ! ( n k ) ! e x 2 l = 0 n k ( n k l ) E n k l ( r ) x l d x = ( 1 ) k n ! ( n k ) ! l = 0 n k ( n k l ) E n k l ( r ) e x 2 x l d x = ( 1 ) k n ! π 0 l n k , l : even 1 ( n k l ) ! 2 l ( l 2 ) ! j = 0 n k l ( r j ) j ! 2 j S 2 ( n k l , j ) .
(2.8)
From (2.7) and (2.8), we can derive the following equation:
C k = n ! 0 l n k , l : even E n k l ( r ) k ! ( n k l ) ! 2 k + l ( l 2 ) ! = n ! 0 l n k , l : even j = 0 n k l ( r j ) j ! S 2 ( n k l , j ) k ! ( n k l ) ! 2 k + l + j ( l 2 ) ! .
(2.9)

Therefore, by Corollary 2.2, (2.6) and (2.9), we obtain the following theorem.

Theorem 2.3 For n 0 , we have
E n ( r ) ( x ) = n ! k = 0 n { 0 l n k , l : even E n k l ( r ) k ! ( n k l ) ! 2 k + l ( l 2 ) ! } H k ( x ) = n ! k = 0 n { 0 l n k , l : even j = 0 n k l ( r j ) j ! S 2 ( n k l , j ) k ! ( n k l ) ! 2 k + l + j ( l 2 ) ! } H k ( x ) .
By (1.4), we easily see that
( t e t 1 ) r = ( 1 + e t t 1 t ) r = j = 0 ( r j ) ( e t t 1 t ) j .
(2.10)
Thus, by (2.10), we get
( t e t 1 ) r e x t = n = 0 ( j = 0 n ( r j ) ( e t t 1 t ) j x n ) t n n ! .
(2.11)
From (1.4) and (2.11), we have
B n ( r ) ( x ) = j = 0 n ( r j ) ( e t t 1 t ) j x n .
(2.12)
By (1.19), we easily get
( e t t 1 t ) j x n = k = 0 n j t k | ( e t t 1 t ) j x n k ! x k = k = 0 n j ( e t t 1 t ) j | t k x n k ! x k = k = 0 n j ( n k ) l = 0 j ( j l ) ( 1 ) j l ( e t 1 t ) l | x n k x k = k = 0 n j ( n k ) l = 0 j ( j l ) ( 1 ) j l t 0 | ( e t 1 t ) l x n k x k .
(2.13)
From (1.8), (1.21) and (2.13), we have
( e t t 1 t ) j x n = k = 0 n j l = 0 j ( n k ) ( j l ) ( 1 ) j l ( n k ) ! l ! ( n k + l ) ! S 2 ( n k + l , l ) x k .
(2.14)
Thus, by (2.12) and (2.14), we get
B n ( r ) ( x ) = j = 0 n k = 0 n j l = 0 j ( r j ) ( n k ) ( j l ) ( 1 ) j l S 2 ( n k + l , l ) ( n k + l l ) x k = k = 0 n ( n k ) [ j = 0 k l = 0 j ( r j ) ( j l ) S 2 ( k + l , l ) ( k + l l ) ( 1 ) j l ] x n k .
(2.15)

Therefore, by (2.12) and (2.15), we obtain the following theorem.

Theorem 2.4 For n 0 , we have
B n ( r ) ( x ) = k = 0 n ( n k ) [ j = 0 k l = 0 j ( r j ) ( j l ) S 2 ( k + l , l ) ( k + l l ) ( 1 ) j l ] x n k .
By (1.4), we easily get
B n ( r ) ( x ) = k = 0 n ( n k ) B k ( r ) x n k .
(2.16)

Therefore, by Theorem 2.4 and (2.16), we obtain the following corollary.

Corollary 2.5 For k 0 , we have
B k ( r ) = j = 0 k l = 0 j ( 1 ) j l ( r j ) ( j l ) S 2 ( k + l , l ) ( k + l l ) .
Let us consider p ( x ) = B n ( r ) ( x ) P n . Then, by (1.11), B n ( r ) ( x ) can be written as
B n ( r ) ( x ) = k = 0 n C k H k ( x ) .
(2.17)
Now, we compute C k ’s for B k ( r ) ( x ) as follows:
C k = ( 1 ) k 2 k k ! π ( d k e x 2 d x k ) B n ( r ) ( x ) d x ,
(2.18)
where
( d k e x 2 d x k ) B n ( r ) ( x ) d x = ( n ) ( ( n 1 ) ) ( ( n k + 1 ) ) e x 2 B n k ( r ) ( x ) d x = ( 1 ) k n ! ( n k ) ! l = 0 n k ( n k l ) B n k l ( r ) e x 2 x l d x = ( 1 ) k n ! π 0 l n k , l : even B n k l ( r ) ( n k l ) ! 2 l ( l 2 ) ! .
(2.19)
By Corollary 2.5 and (2.19), we get
( d k e x 2 d x k ) B n ( r ) ( x ) d x = ( 1 ) k n ! π 0 l n k , l : even s j = 0 n k l m = 0 j ( 1 ) j m ( r j ) ( j m ) S 2 ( n k l + m , m ) ( n k l ) ! 2 l ( l 2 ) ! ( n k l + m m ) .
(2.20)
From (2.18) and (2.20), we have
C k = n ! 0 l n k , l : even B n k l ( r ) ( n k l ) ! k ! 2 k + l ( l 2 ) ! = n ! 0 l n k , l : even j = 0 n k l m = 0 j ( 1 ) j m ( r j ) ( j m ) S 2 ( n k l + m , m ) ( n k l ) ! k ! 2 k + l ( l 2 ) ! ( n k l + m m ) .
(2.21)

Therefore, by (2.17) and (2.21), we obtain the following theorem.

Theorem 2.6 For n 0 , we have
B n ( r ) ( x ) = n ! k = 0 n { 0 l n k , l : even B n k l ( r ) ( n k l ) ! k ! 2 k + l ( l 2 ) ! } H k ( x ) = n ! k = 0 n { 0 l n k , l : even j = 0 n k l m = 0 j ( 1 ) j m ( r j ) ( j m ) S 2 ( n k l + m , m ) ( n k l ) ! k ! 2 k + l ( l 2 ) ! ( n k l + m m ) } H k ( x ) .
It is easy to show that
( 1 λ e t λ ) r = ( 1 + e t 1 1 λ ) r = j = 0 ( r j ) ( 1 1 λ ) j ( e t 1 ) j .
(2.22)
From (1.6) and (2.22), we have
H n ( r ) ( x | λ ) = j = 0 n ( r j ) ( 1 λ ) j ( e t 1 ) j x n ,
(2.23)
where
( e t 1 ) j x n = j ! k = j S 2 ( k , j ) t k k ! x n = j ! k = j n ( n k ) S 2 ( k , j ) x n k .
(2.24)
Thus, by (2.24), we get
( e t 1 ) j x n = j ! k = j n ( n k ) S 2 ( k , j ) x n k .
(2.25)
From (2.23) and (2.25), we can derive the following equation:
H n ( r ) ( x | λ ) = j = 0 n k = j n ( n k ) ( r j ) j ! ( 1 λ ) j S 2 ( k , j ) x n k = k = 0 n j = 0 k ( n k ) ( r j ) j ! ( 1 λ ) j S 2 ( k , j ) x n k = k = 0 n ( n k ) [ j = 0 k ( r j ) j ! ( 1 λ ) j S 2 ( k , j ) ] x n k .
(2.26)
By (1.6), we easily see that
H n ( r ) ( x | λ ) = k = 0 n ( n k ) H k ( r ) ( λ ) x n k .
(2.27)

Therefore, by (2.26) and (2.27), we obtain the following theorem.

Theorem 2.7 For k 0 , we have
H k ( r ) ( λ ) = j = 0 k ( r j ) j ! ( 1 λ ) j S 2 ( k , j ) .
Let us take p ( x ) = H n ( r ) ( x | λ ) P n . Then, by (1.11), H n ( r ) ( x | λ ) is given by
H n ( r ) ( x | λ ) = k = 0 n C k H k ( x ) .
(2.28)
By (1.12), we get
C k = ( 1 ) k 2 k k ! π ( d k e x 2 d x k ) H n ( r ) ( x | λ ) d x ,
(2.29)
where
( d k e x 2 d x k ) H n ( r ) ( x | λ ) d x = ( 1 ) k n ! ( n k ) ! l = 0 n k ( n k l ) H n k l ( r ) ( λ ) e x 2 x l d x = ( 1 ) k n ! π 0 l n k , l : even H n k l ( r ) ( λ ) ( n k l ) ! 2 l ( l 2 ) ! = ( 1 ) k n ! π 0 l n k , l : even j = 0 n k l ( r j ) j ! S 2 ( n k l , j ) ( n k l ) ! 2 l ( 1 λ ) j ( l 2 ) ! .
(2.30)
By (2.29) and (2.30), we get
C k = n ! 0 l n k , l : even H n k l ( r ) ( λ ) ( n k l ) ! k ! 2 l + k ( l 2 ) ! = n ! 0 l n k , l : even j = 0 n k l ( r j ) j ! S 2 ( n k l , j ) ( n k l ) ! k ! 2 k + l ( 1 λ ) j ( l 2 ) ! .
(2.31)

Therefore, by (2.28) and (2.31), we obtain the following theorem.

Corollary 2.8 For n 0 , we have
H n ( r ) ( x | λ ) = n ! k = 0 n { 0 l n k , l : even H n k l ( r ) ( λ ) ( n k l ) ! k ! 2 l + k ( l 2 ) ! } H k ( x ) = n ! k = 0 n { 0 l n k , l : even j = 0 n k l ( r j ) j ! S 2 ( n k l , j ) ( n k l ) ! k ! 2 k + l ( 1 λ ) j ( l 2 ) ! } H k ( x ) .

Declarations

Acknowledgements

The authors would like to express their sincere gratitude to referees for their valuable comments. This paper is supported in part by the Research Grant of Kwangwoon University in 2013.

Authors’ Affiliations

(1)
Department of Mathematics, Sogang University
(2)
Department of Mathematics, Kwangwoon University
(3)
Hanrimwon, Kwangwoon University
(4)
Department of Mathematics Education, Kyungpook National University

References

  1. Kim DS, Kim T, Rim SH, Lee S-H: Hermite polynomials and their applications associated with Bernoulli and Euler numbers. Discrete Dyn. Nat. Soc. 2012., 2012: Article ID 974632, 13 pp. doi:10.1155/2012/974632Google Scholar
  2. Kim T: Symmetry p -adic invariant integral on Z p for Bernoulli and Euler polynomials. J. Differ. Equ. Appl. 2008, 14: 1267–1277. 10.1080/10236190801943220View ArticleGoogle Scholar
  3. Kim T, Choi J, Kim YH, Ryoo CS: On q -Bernstein and q -Hermite polynomials. Proc. Jangjeon Math. Soc. 2011, 14(2):215–221.MathSciNetGoogle Scholar
  4. Ryoo C: Some relations between twisted q -Euler numbers and Bernstein polynomials. Adv. Stud. Contemp. Math. 2011, 21(2):217–223.MathSciNetGoogle Scholar
  5. Araci S, Acikgoz M: A note on the Frobenius-Euler numbers and polynomials associated with Bernstein polynomials. Adv. Stud. Contemp. Math. 2012, 22(3):399–406.MathSciNetGoogle Scholar
  6. Araci S, Erdal D, Seo J: A study on the fermionic p -adic q -integral representation on Z p associated with weighted q -Bernstein and q -Genocchi polynomials. Abstr. Appl. Anal. 2011., 2011: Article ID 649248, 10 pp.Google Scholar
  7. Bayad A: Modular properties of elliptic Bernoulli and Euler functions. Adv. Stud. Contemp. Math. 2010, 20(3):389–401.MathSciNetGoogle Scholar
  8. Can M, Cenkci M, Kurt V, Simsek Y: Twisted Dedekind type sums associated with Barnes’ type multiple Frobenius-Euler-functions. Adv. Stud. Contemp. Math. 2009, 18(2):135–160.MathSciNetGoogle Scholar
  9. Carlitz L: The product of two Eulerian polynomials. Math. Mag. 1963, 368: 37–41.View ArticleGoogle Scholar
  10. Ding D, Yang J: Some identities related to the Apostol-Euler and Apostol-Bernoulli polynomials. Adv. Stud. Contemp. Math. 2010, 20(1):7–21.MathSciNetGoogle Scholar
  11. Kim DS, Kim T: Some identities of Frobenius-Euler polynomials arising from umbral calculus. Adv. Differ. Equ. 2012., 2012: Article ID 196. doi:10.1186/1687–1847–2012–196Google Scholar
  12. Ozden H, Cangul IN, Simsek Y: Remarks on q -Bernoulli numbers associated with Daehee numbers. Adv. Stud. Contemp. Math. 2009, 18(1):41–48.MathSciNetGoogle Scholar
  13. Rim S-H, Joung J, Jin J-H, Lee S-J: A note on the weighted Carlitz’s type q -Euler numbers and q -Bernstein polynomials. Proc. Jangjeon Math. Soc. 2012, 15: 195–201.MathSciNetGoogle Scholar
  14. Roman S: The Umbral Calculus. Dover, New York; 2005.Google Scholar
  15. Kim T: An identity of the symmetry for the Frobenius-Euler polynomials associated with the fermionic p -adic invariant q -integrals on Z p . Rocky Mt. J. Math. 2011, 41: 239–247. 10.1216/RMJ-2011-41-1-239View ArticleGoogle Scholar
  16. Simsek Y: Generating functions of the twisted Bernoulli numbers and polynomials associated with their interpolation functions. Adv. Stud. Contemp. Math. 2008, 16(2):251–278.MathSciNetGoogle Scholar

Copyright

© Kim et al.; licensee Springer. 2013

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.