Next Article in Journal
On State Occupancies, First Passage Times and Duration in Non-Homogeneous Semi-Markov Chains
Previous Article in Journal
A Chaotic Krill Herd Optimization Algorithm for Global Numerical Estimation of the Attraction Domain for Nonlinear Systems
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Multiple Dedekind Type Sums and Their Related Zeta Functions

1
Laboratoire de Mathématiques et Modélisation ďÉvry (LAMME), Université Paris-Saclay, CNRS (UMR 8071), Bâtiment I.B.G.B.I., 23 Boulevard de France, CEDEX, 91037 Evry, France
2
Department of Mathematics, Faculty of Arts and Science, University of Akdeniz, Antalya 07058, Turkey
*
Author to whom correspondence should be addressed.
Mathematics 2021, 9(15), 1744; https://doi.org/10.3390/math9151744
Submission received: 26 April 2021 / Revised: 20 July 2021 / Accepted: 22 July 2021 / Published: 24 July 2021
(This article belongs to the Section Algebra, Geometry and Topology)

Abstract

:
The main purpose of this paper is to use the multiple twisted Bernoulli polynomials and their interpolation functions to construct multiple twisted Dedekind type sums. We investigate some properties of these sums. By use of the properties of multiple twisted zeta functions and the Bernoulli functions involving the Bernoulli polynomials, we derive reciprocity laws of these sums. Further developments and observations on these new Dedekind type sums are given.

Contents
 
1. Introduction    1
 
2. Twisted Barnes’ Type (h,q)-Zeta Functions    9
 
3. Twisted (h,q)-Dedekind Type Sums    13
 
References    19

1. Introduction

Throughout this paper, we use the following notations:
  • Z denotes the ring of integers;
  • N : = 1 , 2 , 3 , ;
  • N 0 : = N 0 ;
  • for q C , we denote
    x = x : q = 1 q x 1 q , q 1 x , q = 1
  • ξ is an r-th root of 1 with r N .
The purpose of this paper is not only to study the different types of higher-order twisted  ( h , q ) -Bernoulli numbers and polynomials, which generalize those of [1], but also to study the relations between these numbers, polynomials, and Dedekind type sums and related areas.
The main motivation is the study of a q-analogue of the generalized Barnes’ multiple zeta function (i.e., q-Barnes’ multiple zeta function in twisted version) and to introduce a new type of Dedekind sum. We prove a Dedekind’s type reciprocity law for these new sums. Our resulting generalized reciprocity formulas recover the results of Apostol [2] and Ota [3].
Many mathematicians have studied the so-called twisted Bernoulli numbers and polynomials and their interpolation functions. For more details, one can see Cenkci et al. [4], Hu-Min Soo [5], Bayad [6], Kim [7], Koblitz [8], and Simsek [9].
Let h N and q , ξ C , we assume that q < 1 . By using p-adic q-integral theory, Simsek [9] defined the generating function of the twisted ( h , q ) -Bernoulli numbers B n , ξ ( h ) ( q ) using the following generating function:
F ξ , q ( h ) ( t ) : = log q h + t ξ q h e t 1 = n = 0 B n , ξ ( h ) ( q ) t n n ! .
From the above equation, we have
B 0 , ξ ( h ) ( q ) = log q h ξ q h 1 , ξ q h ( B ξ ( h ) ( q ) + 1 ) n B n , ξ ( h ) ( q ) = δ 1 , n , n 1 ,
where δ 1 , n denotes the Kronecker symbol and the usual convention of symbolically replacing ( B ξ ( h ) ( q ) ) n by B n , ξ ( h ) ( q ) (cf. [9,10,11,12]). The link between the twisted ( h , q ) -extension of Bernoulli numbers and Frobenius–Euler numbers is given in [11] by the relation
B n , ξ ( h ) ( q ) = log q h H n ( ξ 1 q h ) + n H n 1 ( ξ 1 q h ) ξ q h 1 ,
where H n ( u ) denotes the Frobenius–Euler numbers, which are defined as follows. For u C with | u | > 1 , the Frobenius–Euler numbers H n ( u ) are defined by using of the following generating function:
1 u e t u = n = 0 H n ( u ) t n n ! ;
H n ( u ) are rational fractions of polynomials and were studied in great detail by Frobenius [13], who was particularly interested in their relationship to Bernoulli numbers [14] and relation (2.7) in [15].
One can observe that for ξ = 1 , B n , ξ ( h ) ( q ) = B n ( h ) ( q ) (cf. [16]) and lim q 1 B n ( h ) ( q ) = B n , B n are the Bernoulli numbers (cf. [2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32]).
The twisted ( h , q ) -Bernoulli polynomials are defined in [9] by using the generating function
F ξ , q ( h ) ( t , z ) = ( t + log q h ) e t z ξ q h e t 1 = n = 0 B n , ξ ( h ) ( z , q ) t n n ! .
We easily see that B n , ξ ( h ) ( 0 , q ) = B n , ξ ( h ) ( q ) . From (2), we have
B n , ξ ( h ) ( z , q ) = k = 0 n n k z n k B k , ξ ( h ) ( q ) .
The twisted ( h , q ) -Bernoulli numbers and polynomials of order v are given in [1,10,12]) by their generating functions:
F ξ , q ( v ) ( t a ) = ( log q h + t ) v j = 1 v a j ( ξ q h e t ) a j 1 = n = 0 B n , ξ ( h , v ) ( q a ) t n n ! ,
and
F ξ , q ( v ) ( t , x a ) = F ξ , q ( v ) ( t a ) e t x = n = 0 B n , ξ ( h , v ) ( x , q a ) t n n !
where a = ( a 1 , , a v ) .
By using (4), we obtain
B n , ξ ( h , v ) ( z , q a ) = j = 0 n n j z n j B j , ξ ( h , v ) ( q a ) .
Substituting a = ( a , , a ) (4), we see that
F ξ , q ( v ) ( t , x a ) = a v ( log q h + t ) v 1 ( ξ q h e t ) a 1 v e t x = n = 0 B n , ξ ( h , v ) ( x , q a ) t n n ! .
From the above, we arrive at the equality
n = 0 B n , ξ ( h , v ) ( x , q a ) t n n ! = a v ( log q h + t ) v j = 0 s + j 1 j ( ξ q h ) a j e t ( x + a j ) .
From [10,12], we quote the distribution formula.
Theorem 1.
[Distribution relation]. Let v, n, j N . We have
B n , ξ ( h , v ) ( z , q a ) = j n 1 b 1 , , b v = 0 j 1 ( ξ q h ) a . b B n , ξ ( h , v ) a . b + z j , q a
where a . b = i = 1 v a i b i , a = ( a 1 , , a v ) , b = ( b 1 , , b v ) .
It is well-known that the distribution relation is useful for the construction of distribution on the ring of p-adic integers Z p . For more details, see [8], Chap. II. On the other hand, the above Theorem 1 and the results of Chapter II of Koblitz’s book [8] illustrate that the twisted ( h , q ) -Bernoulli polynomials are p-adic in essence and have profound connections with the special values of certain zeta functions.
In our paper, we will construct these zeta functions, and their study will be further detailed in Section 2 and Section 3.
Let us now specify the definition of the twisted ( h , q ) -Bernoulli numbers of order v by using the following generating function:
F ξ , q ( h , v ) ( t a ) = ( log q h + t ) v j = 1 v a j ( ξ q h ) a j e a j t 1 = n = 0 B n , ξ ( h , v ) ( q a ) t n n ! ,
where a = ( a 1 , , a v ) and t + log ( ξ q h ) < min 2 π a 1 , , 2 π a v (cf. [10,12]).
By using (1) and (6), we are now ready to give the relation between the numbers B n , ξ ( h , v ) ( q a ) and B n , ξ ( h ) ( q ) as follows. From (6), we obtain
n = 0 B n , ξ ( h , v ) ( q a ) t n n ! = j = 1 v log q a j h + a j t ( ξ q h ) a j e a j t 1 ,
By substituting (1) into the above, we find that
n = 0 B n , ξ ( h , v ) ( q a ) t n n ! = j = 1 v n = 0 B n , ξ a j ( h ) ( q a j ) a j n t n n ! .
By substituting v = 1 into (7), we have
B n , ξ ( h , 1 ) ( q a 1 ) = a 1 B n , ξ a 1 ( h ) ( q a 1 ) .
Setting v = 2 and a = ( a 1 , a 2 ) in (7), we have
n = 0 B n , ξ ( h , 2 ) ( q ( a 1 , a 2 ) ) t n n ! = j = 0 B j , ξ a 1 ( h ) ( q a 1 ) a 1 j t j j ! l = 0 B l , ξ a 2 ( h ) q a 2 a 2 l t l l ! .
By using the Cauchy product in the above, we obtain
B n , ξ ( h , 2 ) ( q ( a 1 , a 2 ) ) = 1 B j , ξ a 1 ( h ) ( q a 1 ) a 1 + 2 B l , ξ a 2 ( h ) ( q a 2 ) a 2 n ,
where i B ξ a ( h ) ( q a ) j = B j , ξ a ( h ) ( q a ) and i B ξ a ( h ) ( q a ) j l B ξ a ( h ) ( q a ) d B j + d , ξ a ( h ) ( q a ) if i l .
By (6), using geometric series, we find that
F ξ , q ( h , v ) ( t a ) = ( 1 ) v ( log q h + t ) v j = 1 v a j y 1 , , y v = 0 ( ξ q h e t ) a 1 y 1 + + a v y v = n = 0 B n , ξ ( h , v ) q a t n n ! .
For a = ( a , , a ) , Relation (6) is reduced to
F ξ , q ( h , v ) ( t a ) = ( log q h + t ) v a v 1 ( ξ q h ) a e a t 1 v
where t + log ( ξ q h ) < 2 π a .
Observe that if a = ( 1 , , 1 ) , then (6) reduces to
F ξ , q ( h , v ) ( t ( 1 , , 1 ) ) = log q h + t ξ q h e t 1 v = n = 0 B n , ξ ( h , v ) ( q ) t n n !
which is studied in [1].
Without loss of generality and also for the simplicity of the calculations, in this section, we treat in detail only the case where a = ( a , , a ) . Here, our method can be extended to the general case.
By replacing x by x y in (4) with y N , we have
n = 0 B n , ξ ( h , v ) x y , q a t n n ! = log q h a + a t v e x y t ( ( ξ q h ) a e a t 1 ) v .
After some calculations, we obtain
n = 0 B n , ξ ( h , v ) x y , q a t n n ! = ( log q h a + a t ) v e x y t ( ( ξ q h ) a y e a y t 1 ) v j = 0 y 1 ( ξ q h e t ) j a v .
From the above, we find that
n = 0 B n , ξ ( h , v ) x y , q a t n n !
is equal to
( log q h a + a t ) v ( ( ξ q h ) a y e a y t 1 ) v 1 + ( ξ q h e t ) a + ( ξ q h e t ) 2 a + + ( ξ q h e t ) a ( y 1 ) v .
Thus, we have
( ( ξ q h e t ) a y 1 ) v ( log q h a + a t ) v n = 0 B n , ξ ( h , v ) x y , q a t n n ! = b 1 , , b y 1 0 b 1 + + b y 1 = v v b 1 , , b y 1 ( ξ a q h a ) b 1 + 2 b 2 + + ( y 1 ) b y 1 e x + b 1 + 2 b 2 + + ( y 1 ) b y 1 y y t .
By using (4), we have
n = 0 B n , ξ ( h , v ) x y , q a t n n ! = b 1 , , b y 1 0 b 1 + + b y 1 = v v b 1 , , b y 1 a v ξ a q h a k = 1 y 1 k b k ( log q h + t ) v e x + k = 1 y 1 k b k y y t ( ( ξ q h ) a y e a y t 1 ) v = n = 0 y n b 1 , , b y 1 0 b 1 + + b y 1 = v v b 1 , , b y 1 ( ξ a q h a ) e k = 1 y 1 k b k B n , ξ y ( h , v ) ( x + k = 1 y 1 k b k y , q y a t ) t n n ! .
By identifying the coefficients of t n n ! in the above formula, we obtain the q-Raabe multiplication formula for the polynomials B n , ξ ( h , v ) ( x , q a ) .
Proposition 1
([10,12]). Let a = ( a , , a ) , v, y N , q C with q < 1 and ξ r = 1 with ξ 1 . Then, we have
B n , ξ ( h , v ) x y , q a = y n b 1 , , b y 1 0 b 1 + + b y 1 = v v b 1 , , b y 1 ξ q h a k = 1 y 1 k b k × B n , ξ y ( h , v ) x + b 1 + 2 b 2 + 3 b 3 + + ( y 1 ) b y 1 y , q y a .
We can rewrite F ξ , q ( h , v ) ( t a ) , defined in (9), as follows:
F ξ , q ( h , v ) ( t a ) = ( 1 ) v j = 1 v a j k = 0 v v k ( log q h ) k t v k j = 1 v 1 ξ q h a j e a j t 1 = n = 0 B n , ξ ( h , v ) ( q a ) t n n ! .
Now, we give some identities related to our twisted Bernoulli polynomials of higher order.
Let
F ξ , q ( h , v ) ( t , x a ) = log q + t v a v e x t ( ( ξ q e t ) a 1 ) v = n = 0 B n , ξ ( v , h ) x , q a t n n ! ,
where a = ( a , , a ) . By the well-known identity
1 ( ξ q h ) a e a t 1 = 1 + ( ξ q h ) a e a t + + ( ξ q h ) a ( m 1 ) e a ( m 1 ) t ( ξ q h ) a m e a m t 1 ,
we obtain
n = 0 B n , ξ ( h , v ) ( m x , q a ) t n n ! = log q a h + a t v e m x t ( ( ξ q h ) a e a t 1 ) v = log q a h + a t v e m x t j = 0 m 1 ( ξ a q a h ) j e a j t v 1 ( ξ q h ) a m e a m t 1 v = log q a h + a t v e m x t 1 ξ a m q a h m e a m t 1 v × 1 + ξ a q a h e a t + + ξ a ( m 1 ) q a h ( m 1 ) e a ( m 1 ) t = log q a h + a t v e m x t 1 ξ m a q a h m e m a t 1 v × α 1 , , α m 1 0 v α 1 , , α m 1 ( ξ a q a h ) α 1 e α 1 a t ( ξ a q a h ) 2 α 2 e 2 α 2 a t × ( ξ a q a h ) α m 1 a ( m 1 ) e α m 1 a ( m 1 ) t = log q a h + a t ( ξ q h ) m a e m a t 1 v e m x t α 1 , , α m 1 0 v α 1 , , α m 1 × ( ξ a q a h ) α 1 + 2 α 2 + + ( m 1 ) α m 1 e a t ( α 1 + 2 α 2 + + ( m 1 ) α m 1 ) .
From the above, we have
n = 0 B n , ξ ( h , v ) m x , q a t n n ! = m v a v log q a h m + a t m ( ξ q ) a m e a m t 1 v α 1 , , α m 1 0 v α 1 , , α m 1 × ( ξ a q a h ) α 1 + 2 α 2 + + ( m 1 ) α m 1 e a t m x a + α 1 + 2 α 2 + + ( m 1 ) α m 1 m = ( a m ) v α 1 , , α m 1 0 v α 1 , , α m 1 ( ξ a q a h ) α 1 + 2 α 2 + + ( m 1 ) α m 1 × n = 0 B n , ξ m a ( h , a ) x a + α 1 + 2 α 2 + + ( m 1 ) α m 1 m , q a m a n t n n ! .
By comparing the coefficients of t n on both sides of the above, we arrive at the following result.
Proposition 2.
Let a = ( a , , a ) , v, m N , q C with q < 1 and ξ r = 1 with ξ 1 . Then, we have
B n , ξ ( h , v ) m x , q a = ( m a ) n v α 1 , , α m 1 0 α 1 + + α m 1 = v v α 1 , , α m 1 × ( ξ q h ) a y B n , ξ m a ( h , v ) x a + y m , q m a a ,
where y = α 1 + 2 α 2 + + ( m 1 ) α m 1 .
Remark 1.
We give some particular cases with a = ( a , , a ) .
  • Taking q 1 , by Proposition 2, we obtain
    B n , ξ ( h , v ) ( m x a ) = ( a m ) n v α 1 , , α m 1 0 α 1 + + α m 1 = v v α 1 , , α m 1 ξ a y B n , ξ m a ( h , v ) x a + y m a .
  • In addition, if ξ = 1 in the above, we deduce the well-known Carlitz’s multiplication formula [17]:
    B n ( v ) ( m x ) = m n v α 1 , , α m 1 0 α 1 + + α m 1 = v v α 1 , , α m 1 B n ( v ) x a + y m .
  • If v = 1 in Proposition 2, then the Raabe formula for the usual Bernoulli polynomials is given by
    B n ( m x ) = m n 1 j = 0 m 1 B n x + j m .
  • In [23], Kim investigated several properties of symmetry for the p-adic invariant integral on Z p . By using symmetry for the p-adic invariant integral on Z p , Kim proved (12). If a = 1 and ξ = 1 , q a λ , then we arrive at Luo’s formula [26]:
    B n ( v ) ( m x , λ ) = m n v α 1 , , α m 1 0 α 1 + + α m 1 = v v α 1 , , α m 1 λ y B n ( v ) x + y m , λ m .
By using (12), we obtain
n = 0 B n , ξ ( h , v ) m x , q a t n n ! = ( log q a h m + a t m ) v m v a v e m x t ( ( ξ q h ) a m e a m t 1 ) v α 1 , , α m 1 0 v α 1 , , α m 1 ( ξ a q a h ) y e a t y .
From the above, we obtain
n = 0 B n , ξ ( h , v ) m x , q a t n n ! = ( a m ) v α 1 , , α m 1 0 v α 1 , , α m 1 ( ξ q h ) a y n = 0 B n , ξ m a ( h , v ) q m a a t n ( a m ) n n ! × n = 0 ( x a + y m ) ( a m t ) n n ! = ( a m ) v + n α 1 , , α m 1 0 v α 1 , , α m 1 ( ξ q h ) a y × n = 0 j = 0 n ( x a + y m ) n j ( n j ) ! B n , ξ m a ( h , v ) q m a a t n n ! .
Thus, we arrive at the following result.
Corollary 1.
Let v, m N , q C with q < 1 and ξ r = 1 with ξ 1 . We have
B n , ξ ( h , v ) m x , q a ( a m ) n v = α 1 , , α m 1 0 α 1 + + α m 1 = v j = 0 n n j v α 1 , , α m 1 × ( ξ q h ) a y x a + y m n j B n , ξ m a ( h , v ) q m a a ,
where y = α 1 + 2 α 2 + + ( m 1 ) α m 1 .
For m = 1 , v = 1 , from Corollary 1, we obtain
B n , ξ ( h , 1 ) ( x , q ) = B n , ξ ( h ) ( x , q ) = j = 0 n n j x n j B n , ξ ( h ) ( q ) ,
(cf. [9]).
Corollary 2.
Let v N , q C with q < 1 and ξ r = 1 with ξ 1 . Then, we have the reduction formula
B n , ξ ( h , v ) ( x , q ) = k = 0 n n k B n k , ξ ( h , v 1 ) ( q ) B k , ξ ( h ) ( x , q ) , n N 0
where B n , ξ ( h , v ) ( x , q ) : = B n , ξ ( h , v ) x , q a and B n , ξ ( h , v ) ( q ) : = B n , ξ ( h , v ) ( 0 , q a ) ; a = ( 1 , , 1 ) holds true between the twisted ( h , q ) -Bernoulli polynomials B n , ξ ( h ) ( x , q ) and the twisted ( h , q ) -Bernoulli numbers B n , ξ ( h , v 1 ) ( q ) of order ( h , v 1 ) .
Proof. 
By using (10), we have
n = 0 B n , ξ ( h , v ) ( x , q ) t n n ! = log q h + t q h ξ e t 1 v 1 log q h + t q h ξ e t 1 e x t = n = 0 B n , ξ ( h , v 1 ) ( q ) z n n ! n = 0 B n , ξ ( h ) ( z , q ) t n n ! = n = 0 k = 0 n B k , ξ ( h ) ( x , q ) t k k ! B n k , ξ ( h , v 1 ) ( q ) t n k ( n k ) ! = n = 0 k = 0 n B n k , ξ ( h , v 1 ) ( q ) B k , ξ ( h ) ( x , q ) n k t n n !
here, we use the Cauchy product. By comparing the coefficients of t n n ! on both sides of the above equation, we then arrive at the desired result. □

2. Twisted Barnes’ Type ( h , q ) -Zeta Functions

In this section, we construct interpolation functions of the twisted ( h , q ) -Bernoulli polynomials and numbers of higher order. We also give some interesting identities related to these functions. Throughout this section, we study the complex s-plane. Let q C with q < 1 and ξ r = 1 ( ξ is an rth root of 1) ξ 1 .
By (1) and (2), we define the following functions:
F 1 , ξ , q ( h ) ( t ) = t ξ q h e t 1 = n = 0 b n , ξ ( h ) ( q ) t n n ! .
Note that the numbers b n , ξ ( h ) ( q ) are related to the so-called Apostol-Bernoulli numbers [26,27] and twisted ( h , q ) Bernoulli numbers.
We decompose (2) as follows:
F ξ , q ( h ) ( t ) = log q h t F 1 , ξ , q ( h ) ( t ) + F 1 , ξ , q ( h ) ( t ) = n = 0 B n , ξ ( h ) ( q ) t n n ! .
From the above and the relation (13), we find that
B n , ξ ( h ) ( q ) = log q ( n + 1 ) h b n + 1 , ξ ( h ) ( q ) + b n , ξ ( h ) ( q ) .
Let us define
F ξ , q ( h ) ( t , z ) = F 1 , ξ , q ( h ) ( t ) e t z = n = 0 b n , ξ ( h ) ( z , q ) t n n !
observe that b n , ξ ( h ) ( 0 , q ) = b n , ξ ( h ) ( q ) .
For 0 < z 1 , by applying the Mellin transformation to the above equation, we obtain
G q , ξ ( h ) s , z = 1 Γ ( s ) 0 t s 2 F ξ , q ( h ) ( t , z ) d t = n = 0 ξ q h n ( n + z ) s ,
it is easy to see that the series s n = 0 ξ q h n ( n + z ) s converge for whole complex plane, because q < 1 and ξ r = 1 .
We easily see that
G q , ξ ( h ) s , 1 = G q , ξ ( h ) s = n = 1 ξ q h n n s .
Observe that for ( s ) > 1 , we have
lim q 1 G q , 1 ( h ) ( s ) = n = 1 1 n s ,
which denotes the Riemann zeta function.
The function G q , ξ ( h ) s , z interpolates the polynomials b n , ξ ( h ) ( z , q ) at negative integers as follows:
G q , ξ ( h ) 1 n , z = b n , ξ ( h ) ( z , q ) n , n Z + .
We easily see from the above that
G q , ξ ( h ) 1 n = b n , ξ ( h ) ( q ) n , n Z + .
We now construct higher-order interpolation functions.
The Mellin transform of (11) is given by
Z q , ξ ( h , v ) ( s a ) = 1 Γ ( s ) 0 t s v 1 F ξ , q ( h , v ) ( t a ) d t
and we obtain
Z q , ξ ( h , v ) ( s a ) = ( 1 ) v j = 1 v a j k = 0 v v k ( log q h ) k 1 Γ ( s ) 0 t s v 1 t v k j = 1 v d t ξ q h a j e a j t 1
Hence, from the above, we define the zeta functions Z q , ξ ( h , v ) ( s , a ) .
Definition 1.
Let s , q C with q < 1 and ξ r = 1 with ξ 1 . We define
Z q , ξ ( h , v ) ( s a ) j = 1 v a j = k = 0 v v k ( log 1 q h ) k 1 Γ ( s ) 0 t s k 1 d t j = 1 v ξ q h a j e a j t 1 ,
where ( s ) > v + 1 and ( a j ) > 0 , 1 j v .
In [20], the authors gave another kind of the twisted Barnes zeta function.
By substituting v = 1 into (19), we have
Z q , ξ ( h , 1 ) ( s a 1 ) = a 1 1 s n = 1 ( q a 1 ξ a 1 ) n n s log q h a 1 s 1 n = 0 ( q h a 1 ξ a 1 ) n n s 1 = a 1 1 s ζ q , ξ ( h ) ( s )
(cf. [9]).
From (19), we find that
Z q , ξ ( h , v ) ( s , a ) j = 1 v a j = v 0 n 1 , , n v = 1 ξ q h a 1 n 1 + + a v n v ( a 1 n 1 + + a v n v ) s + v 1 log ( 1 q ) s 1 n 1 , , n v = 1 ξ q h a 1 n 1 + + a v n v ( a 1 n 1 + + a v n v ) s 1 + + v v log ( 1 q h ) v ( s 1 ) ( s 2 ) ( s v ) n 1 , , n v = 1 ξ q h a 1 n 1 + + a v n v ( a 1 n 1 + + a v n v ) s v .
From the above, we arrive at the following main theorem.
Theorem 2.
[Explicit formula]. Let s , q C , a = ( a 1 , , a v ) , with q < 1 , ( s ) > v , v N and ξ r = 1 with ξ 1 . Then, we have
Z q , ξ ( h , v ) ( s a ) = j = 1 v a j k = 0 v v k log 1 q h k ζ q , ξ ( h , v ) ( s k a ) j = 0 k s s j ,
where
ζ q , ξ ( h , v ) ( s a ) = n 1 , , n v = 1 ξ q h a 1 n 1 + + a v n v ( a 1 n 1 + + a v n v ) s .
We now give some special values of the function ζ q , ξ ( h , v ) ( s , a ) as follows:
Corollary 3.
Let ( s ) > v . Then, we have
ζ q , ξ ( h , v ) ( s , x a ) = n 1 , , n v = 1 ξ q h a 1 n 1 + + a v n v ( x + a 1 n 1 + + a v n v ) s .
(i) Difference equation of the function ζ q , ξ ( h , v ) ( s , x a ) :
ζ q , ξ ( h , v ) ( s , x + 1 a ) = ζ q , ξ ( h , v ) ( s , x a ) 1 x s ,
(ii) Distribution relation of the function ζ q , ξ ( h , v ) ( s , x a ) :
b s ζ q , ξ ( h , v ) ( s , x a ) = j = 0 b 1 ξ q h j ζ q b , ξ b ( h , v ) ( s , x + j b a ) .
(iii) For real number y,
b s ζ q , ξ ( h , v ) ( s , b y a ) = j = 0 b 1 ξ q h j Z q b , ξ b ( h , v ) ( s , y + j b a ) ,
where
Z q , ξ ( h , v ) ( s , x a ) = ζ q , ξ ( h , v ) ( s , x a ) i f ( x ) > 0 ζ q , ξ ( h , v ) ( s a ) i f x = 0 .
We note that it is easy to prove Corollary 3 from the definition ζ q , ξ ( h , v ) ( s , x , a ) .
Observe that if v = 1 and a = 1 , then (24) reduces to (15) and (16), that is,
Z q , ξ ( h , 1 ) ( s , x 1 ) = ζ q , ξ ( h ) ( s , x ) if ( x ) > 0 ζ q , ξ ( h , v ) ( s ) if x = 0 .
Remark 2.
We give comments on some special cases.
  • Substituting a 1 = = a v = 1 into (20), for s , ξ , q C with Re ( s ) > v , q < 1 and ξ r = 1 , r Z , Theorem 2 recovers the results of [10] (p. 488).
  • If v = 1 , then Theorem 2 reduces to
    Z q , ξ ( h , 1 ) ( s a ) = a 1 1 s ζ q a 1 , ξ a 1 ( h ) ( s ) ,
    where
    ζ q , ξ ( h ) ( s ) = n = 1 q n h ξ n n s log q h s 1 n = 1 q n h ξ n n s 1 ,
  • In [29], Simsek defined and studied the ( ξ , q ) -Hurwitz zeta function which is related to the function ζ q , ξ ( h ) ( s ) :
    ζ ξ , q ( s , z ) = 1 q log q ζ q , ξ ( 2 ) ( s , z ) .
  • In addition, if ξ = 1 , then ζ q , ξ ( h ) ( s ) is reduced to ζ q ( h ) ( s ) (cf. [16]).
  • If v = 1 , and q 1 , then Theorem 2 reduces to the twisted zeta functions which interpolate the twisted Bernoulli numbers:
    ζ ξ ( s ) = n = 1 ξ n n s ,
    where ξ r = 1 , r Z .
  • The Lerch transcendent Φ ( z , s , a ) (cf. e.g., [32] p. 121, [21]) is the analytic continuation of the series
    Φ ( z , s , a ) = 1 a s + z ( a + 1 ) s + z ( a + 2 ) s + = n = 0 z n ( n + a ) s ,
    The above series converges for ( a C Z 0 , s C when z < 1 ; ( s ) > 1 when z = 1 ), where
    Z 0 = Z 0 , Z = 1 , 2 , 3 , .
    Hence, the function ζ q , ξ ( h ) ( s ) is related to Φ ( z , s , a ) as follows:
    ζ q , ξ ( h ) ( s ) = Φ ( q h ξ , s , 1 ) log q h s 1 Φ ( q h ξ , s 1 , 1 ) .
By substituting s = 1 n , n N into (19) and using Cauchy’s residue theorem for the Hankel contour, we obtain the following theorem:
Theorem 3.
[Values at negative integers]. Let n N and a = ( a 1 , , a v ) . Then, we have
Z q , ξ ( h , v ) ( 1 n a ) = n ! ( n + v ) ! ( 1 ) v B n + v , ξ ( h , v ) ( q a ) .
From the above theorem, we arrive at the following corollary:
Corollary 4.
Let n , v N . Thus, we have
B n + v , ξ ( h , v ) ( q a ) = ( n + v ) ! n ! j = 1 v a j v k = 0 v v k log 1 q h k × ζ q , ξ ( h , v ) ( 1 n k a ) j = 0 k 1 n 1 n j .
By substituting v = 2 into Theorem 2, we have
Z q , ξ ( h , 2 ) ( s a ) j = 1 2 a j = k = 0 2 2 k log 1 q h k ζ q , ξ ( h , 2 ) ( s k a ) j = 0 k s s j .
Thus, we find that
Z q , ξ ( h , 2 ) ( s a ) = a 1 a 2 ζ q , ξ ( 2 ) ( s a ) + 2 a 1 a 2 s s 1 log 1 q h ζ q , ξ ( h , 2 ) ( s 1 a ) + a 1 a 2 s 2 ( s 1 ) ( s 2 ) log 1 q h 2 ζ q , ξ ( h , 2 ) ( s 2 a ) .

3. Twisted ( h , q ) -Dedekind Type Sums

In this section, we define twisted ( h , q ) -Dedekind type sums. We state and prove their reciprocity law. For more details on the elliptic analogue for the Dedekind reciprocity laws, see [6].
Let us recall the Apostol–Dedekind sums s n ( h , k ) :
s n ( h , k ) = a = 1 k 1 a k B ¯ n h a k ,
where h and k are coprime integers with k > 0 , n is a positive integer, and B ¯ n x is the nth Bernoulli function, which is defined as follows:
B ¯ n x = B n x x G , if n > 1 ,
and
B ¯ 1 x = B 1 x x G , if x Z 0 , if x Z ,
where B n x is the Bernoulli polynomial.
For even values of n, the sums s n ( h , k ) are relatively uninteresting. However, for odd values of n, these sums have a reciprocity law, first proved by Apostol [2]:
( n + 1 ) h k n s n ( h , k ) + k h n s n ( k , h ) = n B n + 1 + j = 1 n + 1 n + 1 j ( 1 ) j B j B n + 1 j h j k n + 1 j ,
where ( h , k ) = 1 . If n = 1 , then the Apostol–Dedekind sums reduce to the classical Dedekind sums.
We are ready to define the ( h , q ) -twisted Dedekind sums.
Definition 2.
Let h and k be coprime integers with k > 0 . Then, we define
S ξ , q , m ( h 1 ) ( h , k ) = j = 1 k 1 j k ξ q h 1 j b ¯ m , ξ ( h 1 ) h j k , q ,
where h 1 is an integer number and
b ¯ m , ξ ( h 1 ) ( x , q ) = b m , ξ ( h 1 ) ( x x G , q ) ,
where b m , ξ ( h 1 ) ( x , q ) is defined in (14).
Example 1.
By substituting ξ = 1 into (28), we have
S q , m ( h 1 ) ( h , k ) = j = 1 k 1 j k q h 1 j b ¯ m ( h 1 ) h j k , q .
by substituting h 1 = 1 , ξ = 1 and q 1 into (28), (28) reduces to (26).
By substituting h = 1 into (28) and using (13), we obtain
S ξ , q , m ( h 1 ) ( 1 , k ) = j = 1 k 1 j k ξ q h 1 j b m , ξ ( h 1 ) j k , q = l = 0 m m l b l , ξ ( h 1 ) q k m l + 1 j = 1 k 1 ξ q h 1 j j m l + 1
By using the well-known alternating sums of powers of consecutive ( h , q ) -integers for b m , ξ ( h 1 ) ( q ) ,
b = 0 k 1 ξ b q b h 1 b m 1 = ξ k q k h 1 b m , ξ ( h 1 ) ( k , q ) b m , ξ ( h 1 ) ( 0 , q ) m
(cf. [10]), we then obtain the following Theorem.
Theorem 4.
[Explicit evaluation for h = 1 ]. Let m and k be positive integers. Then, we have
S ξ , q , m ( h 1 ) ( 1 , k ) = l = 0 m m l b l , ξ ( h 1 ) q ξ k q h 1 k b m l + 2 , ξ ( h 1 ) ( k , q ) b m l + 2 , ξ ( h 1 ) ( q ) m l + 2 k m l + 1 .
In [3], Ota showed how to prove Apostol’s reciprocity law by using values at non-positive integers of Barnes’ double zeta function. In this section, we give a generalization of Apostol’s reciprocity law.
By substituting v = 2 and a = ( k , h ) into (21), and by using a method similar to that in, then we have
Z q , ξ ( h 1 , 2 ) ( s ( k , h ) ) = h k n 1 , n 2 = 0 n 1 , n 2 ( 0 , 0 ) ξ q h 1 k n 1 + h n 2 ( k n 1 + h n 2 ) s 2 h k log q h 1 s 1 n 1 , n 2 = 0 n 1 , n 2 ( 0 , 0 ) ξ q h 1 k n 1 + h n 2 ( k n 1 + h n 2 ) s 1 + h k log q h 1 2 s 1 s 2 n 1 , n 2 = 0 n 1 , n 2 ( 0 , 0 ) ξ q h 1 k n 1 + h n 2 ( k n 1 + h n 2 ) s 2 .
Setting n 1 = a + m h and n 2 = b + n k , where a = 0 , 1 , 2 , , h 1 , b = 0 , 1 , 2 , , k 1 and m, n = 0 , 1 , 2 , , ) in the above,
Z q , ξ ( h 1 , 2 ) ( s ( h , k ) ) = h k a mod h b mod k m , n = 0 ξ q h 1 a k + b h + h k ( m + n ) ( a k + b h + h k ( m + n ) ) s 2 h k log q h 1 s 1 a mod h b mod k m , n = 0 ξ q h 1 a k + b h + h k ( m + n ) ( a k + b h + h k ( m + n ) ) s 1 + h k log q h 1 2 s 1 s 2 a mod h b mod k m , n = 0 ξ q h 1 a k + b h + h k ( m + n ) ( a k + b h + h k ( m + n ) ) s 2 .
where means that the above sum take pairs on non-negative integers ( m , n ) with the exception of ( 0 , 0 ) when a k + b h = 0 . Putting m + n = j in the above, we then obtain
Z q , ξ ( h 1 , 2 ) ( s ( h , k ) ) = h k a mod h b mod k j = 0 ( j + 1 ) ξ q h 1 a k + b h + h k j ( a k + b h + h k j ) s 2 h k log q h 1 s 1 a mod h b mod k j = 0 ( j + 1 ) ξ q h 1 a k + b h + h k j ( a k + b h + h k j ) s 1 + h k log q h 1 2 s 1 s 2 a mod h b mod k j = 0 ( j + 1 ) ξ q h 1 a k + b h + h k j ( a k + b h + h k j ) s 2 .
After some elementary calculations in the above, we easily see that
ζ q , ξ ( h 1 , 2 ) ( s ( h , k ) ) = a mod h b mod k j = 0 ( a k + b h + h k j ) ξ q h 1 a k + b h + h k j ( a k + b h + h k j ) s + h k a mod h b mod k 1 a k + b h h k j = 0 ξ q h 1 a k + b h + h k j ( a k + b h + h k j ) s 2 log q h 1 s 1 a mod h b mod k j = 0 ( a k + b h + h k j ) ξ q h 1 a k + b h + h k j ( a k + b h + h k j ) s 1 2 h k log q h 1 s 1 a mod h b mod k 1 a k + b h h k j = 0 ξ q h 1 a k + b h + h k j ( a k + b h + h k j ) s 1 + log q h 1 2 s 1 s 2 a mod h b mod k j = 0 ( a k + b h + h k j ) ξ q h 1 a k + b h + h k j ( a k + b h + h k j ) s 2 + h k log q h 1 2 s 1 s 2 a mod h b mod k 1 a k + b h h k j = 0 ξ q h 1 a k + b h + h k j ( a k + b h + h k j ) s 2 .
By using (22) in the above, we find that it is equal to
Z q , ξ ( h 1 , 2 ) ( s ( h , k ) ) = ( h k ) 1 s a mod h b mod k ξ q h 1 a k + b h Z q , ξ s 1 , a k + b h h k ( h k ) 1 s ( a , b ) B ξ q h 1 a k + b h a k + b h h k 1 1 s + ( h k ) 1 s a mod h b mod k ξ q h 1 a k + b h 1 a k + b h h k Z q , ξ s , a k + b h h k ( h k ) 1 s ( a , b ) B ξ q h 1 a k + b h 1 a k + b h h k a k + b h h k 1 s 2 ( h k ) 2 s log q h 1 s 1 a mod h b mod k ξ q h 1 a k + b h Z q , ξ s 2 , a k + b h h k + 2 ( h k ) 2 s log q h 1 s 1 ( a , b ) B ξ q h 1 a k + b h a k + b h h k 1 2 s 2 ( h k ) 2 s log q h 1 s 1 a mod h b mod k ξ q h 1 a k + b h 1 a k + b h h k Z q , ξ s 1 , a k + b h h k + 2 ( h k ) 2 s log q h 1 s 1 ( a , b ) B ξ q h 1 a k + b h 1 a k + b h h k a k + b h h k 1 1 s + ( h k ) 3 s log q h 1 2 s 1 s 2 a mod h b mod k ξ q h 1 a k + b h Z q , ξ s 3 , a k + b h h k ( h k ) 3 s log q h 1 2 s 1 s 2 ( a , b ) B ξ q h 1 a k + b h a k + b h h k 1 3 s + ( h k ) 3 s log q h 1 2 s 1 s 2 a mod h b mod k 1 a k + b h h k ξ q h 1 a k + b h Z q , ξ s 2 , a k + b h h k ( h k ) 3 s log q h 1 2 s 1 s 2 ( a , b ) B ξ q h 1 a k + b h 1 a k + b h h k a k + b h h k 1 2 s ,
where
B = ( a , b ) Z × Z : 0 a h 1 , 0 b k 1 , k a + h b k h .
Since k a + h b : 0 a h 1 , 0 b k 1 , a , b Z is a complete set of representatives modulo h k , by using (23), (15), and (16), we easily obtain
Z q , ξ ( h 1 , 2 ) ( s ( h , k ) ) = G q , ξ ( h 1 ) s 1 + h k G q , ξ ( h 1 ) s 2 log q h 1 s 1 G q , ξ ( h 1 ) s 2 + h k G q , ξ ( h 1 ) s 1 + log q h 1 2 s 1 s 2 G q , ξ ( h 1 ) s 3 + h k G q , ξ ( h 1 ) s 2 k h 1 s a mod h ξ q h 1 a k G q , ξ ( h 1 ) s , a k h h k 1 s b mod k ξ q h 1 b h 1 G q , ξ ( h 1 ) s , b h k + 2 log q h 1 s 1 k h 2 s a mod h ξ q h 1 a k 1 G q , ξ ( h 1 ) s 1 , a k h + 2 log q h 1 s 1 h k 2 s b mod k ξ q h 1 b h 1 G q , ξ ( h 1 ) s 1 , b h k + log q h 1 2 s 1 s 2 k h 3 s a mod h ξ q h 1 a k G q , ξ ( h 1 ) s 2 , a k h + log q h 1 2 s 1 s 2 h k 3 s b mod k ξ q h 1 b h G q , ξ ( h 1 ) s 2 , b h k .
By using the above, we arrive at the following theorem.
Theorem 5.
Let s , q C , with ( s ) > 4 and q < 1 and ξ < 1 . Then, we have
Z q , ξ ( h 1 , 2 ) ( s ( h , k ) ) = ζ q , ξ ( h 1 ) s 1 + h k ζ q , ξ ( h ) s log q h 1 s 1 ζ q , ξ ( h 1 ) s 2 + h k log q h 1 s 1 ζ q , ξ ( h 1 ) s 1 k h 1 s a mod h ξ q h 1 a k G q , ξ ( h 1 ) s , a k h h k 1 s b mod k ξ q h 1 b h G q , ξ ( h 1 ) s , b h k 2 log q h 1 s 1 k h 2 s a mod h ξ q h 1 a k G q , ξ ( h 1 ) s 1 , a k h 2 log q h 1 s 1 h k 2 s b mod k ξ q h 1 b h G q , ξ ( h 1 ) s 1 , b h k log q 2 s 1 s 2 k h 3 s a mod h ξ q h 1 a k G q , ξ ( h 1 ) s 2 , a k h log q 2 s 1 s 2 h k 3 s b mod k ξ q h 1 b h G q , ξ ( h 1 ) s 2 , b h k
where the function ζ q , ξ s is defined in (25).
By using Theorem 5, we prove the reciprocity law of the ( h , q ) -twisted Dedekind sums in the next theorem.
Theorem 6.
[Reciprocity law]. Let h and k be coprime integers with k > 0 . Let n N . Then, we have
k h n S ξ , q , n ( h 1 ) ( k , h ) + h k n S ξ , q , n ( h 1 ) ( h , k ) + 2 log q h 1 n + 1 k h n + 1 S ξ , q , n + 1 ( h 1 ) ( k , h ) + h k n + 1 S ξ , q , n + 1 ( h 1 ) ( h , k ) + log q h 1 2 n + 1 n + 2 k h n + 2 S ξ , q , n + 2 ( h 1 ) ( k , h ) + h k n + 2 S ξ , q , n + 2 ( h 1 ) ( h , k ) = n B n + 2 , ξ ( h 1 , 2 ) ( q ( h , k ) ) + B n + 1 , ξ ( h 1 ) ( q ) n + 1 + h k n B n , ξ ( h 1 ) ( q ) + log q h 1 n + 1 B n + 1 , ξ ( h 1 ) ( q ) + log q h 1 n + 2 B n + 2 , ξ ( h 1 ) ( q ) ,
where B n + 2 , ξ ( h 1 , 2 ) ( q ( h , k ) ) is defined in (8).
Proof. 
By substituting s = 1 n , ( n Z + ) into Theorem 5 and using (17) and (18), after some calculations, we arrive at the desired result. □
Remark 3.
We point out how we recover, from our Theorem 6, some known results.
  • By substituting q 1 and ξ = 1 into Theorem 6, we obtain the result of Ota ([3], [p. 8, Theorem B]):
    k h n S ξ , q , n ( h 1 ) ( k , h ) + h k n S ξ , q , n ( h 1 ) ( h , k ) = n n + 1 B n + 2 ( 2 ) ( ( h , k ) ) + B n + 1 + h k n B n .
    Consequently, Theorem 6 is a generalization of Ota’s theorem ([3], [p. 8, Theorem B]).
  • By substituting h 1 = 1 and ξ = 1 , q 1 into Theorem 6, then we easily arrive at (27). Thus, Theorem 6 is also a generalization of Apostol’s reciprocity theorem [2] for odd n.
  • In [29], Simsek constructed p-adic ( ξ , q ) -Dedekind sums and Hardy–Berndt type sums. In the future, we will study the properties of the twisted p-adic Dedekind sums associated with our objects of study here.

Author Contributions

Conceptualization, A.B. and Y.S. Both authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Acknowledgments

This paper was supported by the Scientific Research Project Administration, Akdeniz University.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Simsek, Y.; Kurt, V.; Kim, D. New approach to the complete sum of products of the twisted (h,q)-Bernoulli numbers and polynomials. J. Nonlinear Math. Phys. 2007, 14, 44–56. [Google Scholar] [CrossRef]
  2. Apostol, T.M. Generalized Dedekind sums and transformation formulae of certain Lambert series. Duke Math. J. 1950, 17, 147–157. [Google Scholar] [CrossRef]
  3. Ota, K. Derivatives of Dedekind sums and their reciprocity law. J. Number Theory 2003, 98, 280–309. [Google Scholar] [CrossRef] [Green Version]
  4. Cenkci, M.; Can, M.; Kurt, V.; Simsek, Y. Twisted Dedekind type sums associated with Barnes’ type multiple Frobenius-Euler l-Functions. Adv. Stud. Contemp. Math. 2009, 18, 135–160. [Google Scholar]
  5. Hu, S.; Kim, M.-S. The p-adic analytic Dedekind sums. J. Number Theory 2017, 171, 112–127. [Google Scholar] [CrossRef]
  6. Bayad, A. Jacobi forms in two variables: Multiple elliptic Dedekind sums, the Kummer-Von Staudt Clausen congruences for elliptic Bernoulli functions and values of Hecke L-functions. Montes Taurus J. Pure Appl. Math. 2019, 1, 58–129. [Google Scholar]
  7. Kim, T. An analogue of Bernoulli numbers and their congruences. Rep. Fac. Sci. Engrg. Saga Univ. Math. 1994, 22, 21–26. [Google Scholar]
  8. Koblitz, N. p-Adic Analysis: A Short Course on Recent Work; London Mathematical Society Lecture Note Series; 1980; Available online: https://www.cambridge.org/cn/academic/subjects/mathematics/number-theory/p-adic-analysis-short-course-recent-work?format=PB (accessed on 26 April 2021).
  9. Simsek, Y. Twisted (h,q)-Bernoulli numbers and polynomials related to twisted (h,q)-zeta function and L-function. J. Math. Anal. Appl. 2006, 324, 790–804. [Google Scholar] [CrossRef] [Green Version]
  10. Simsek, Y. Multiple interpolation functions of higher order (h,q)-Bernoulli numbers. AIP Conf. Proc. 2008, 1048, 486–489. [Google Scholar]
  11. Simsek, Y. Twisted p-adic (h,q)-L-functions. Comput. Math. Appl. 2010, 59, 2097–2110. [Google Scholar] [CrossRef] [Green Version]
  12. Simsek, Y. Analytic Theory, Approximation Theory, and Special Functions/Families of Twisted Bernoulli Numbers, Twisted Bernoulli Polynomials, and Their Applications; Milovanovic, G.V., Rassias, M.T., Eds.; Springer: Berlin/Heidelberg, Germany, 2014; Available online: http://www.springer.com/mathematics/numbers/book/978-1-4939-0257-6?otherVersion=978-1-4939-0258-3# (accessed on 26 April 2021).
  13. Frobenius, G. Uber die Bernoullischen Zahlen und die Eulerschen Polynome. Sitzungsberichte Der Preuss. Akad. Der Wiss. 1910, 39, 809–847. [Google Scholar]
  14. Apostol, T.M. Elementary Proof of the Transformation Formula for Lambert Series Involving Generalized Dedekind Sums. J. Number Theory 1982, 15, 14–24. [Google Scholar] [CrossRef] [Green Version]
  15. Carlitz, L. q-Bernoulli and Eulerian numbers. Trans. Am. Math. Soc. 1954, 76, 332–350. [Google Scholar]
  16. Kim, T. A new approach to q-zeta function. Adv. Stud. Contep. Math. 2005, 11, 157–162. [Google Scholar]
  17. Carlitz, L. The multiplication formulas for the Bernoulli and Euler polynomials. Maths. Mag. 1953, 27, 59–64. [Google Scholar] [CrossRef]
  18. Araci, S.; Şen, E.; Acikgoz, M. A Note on the Modified q-Dedekind Sums. Available online: https://arxiv.org/pdf/1212.5837.pdf (accessed on 26 April 2021).
  19. Barnes, W. The theory of the multiple gamma functions. Trans. Cambridge. Philos. Soc. 1904, 19, 374–425. [Google Scholar]
  20. Bayad, A.; Simsek, Y. Values of twisted Barnes zeta functions at negative integers. Russ. J. Math. Phys. 2013, 20, 129–137. [Google Scholar] [CrossRef]
  21. Guillera, J.; Sondow, J. Double integrals and infinite products for some classical constants via analytic continuations of Lerch’s transcendent. Ramanujan J. 2008, 16, 247–270. [Google Scholar] [CrossRef] [Green Version]
  22. Kim, T. Sums of products of q-Bernoulli numbers. Arch. Math. 2001, 76, 190–195. [Google Scholar] [CrossRef]
  23. Kim, T. Symmetry p-adic invariant integral on Z p for Bernoulli and Euler polynomials. J. Differ. Equ. Appl. 2008, 14, 1267–1277. [Google Scholar] [CrossRef]
  24. Kim, T.; Jang, L.C.; Rim, S.-H.; Pak, H. On the twisted q-zeta functions and q-Bernoulli polynomials. Far East J. Appl. Math. 2003, 13, 13–21. [Google Scholar]
  25. Kim, T.; Rim, S.-H.; Simsek, Y.; Kim, D. On the analogs of Bernoulli and Euler numbers, related identities and zeta and l-functions. J. Korean Math. Soc. 2008, 45, 435–453. [Google Scholar] [CrossRef] [Green Version]
  26. Luo, Q.-M. The Multiplication formula for the Apostol-Bernoulli and Apostol-Euler polynomials of higher order. Integral Transform. Spec. Funct. 2009, 20, 377–391. [Google Scholar] [CrossRef]
  27. Luo, Q.-M.; Srivastava, H.M. Some relationships between the Apostol-Bernoulli and Apostol-Euler polynomials. Comput. Math. Appl. 2006, 51, 631–642. [Google Scholar] [CrossRef] [Green Version]
  28. Shiratani, K.; Yokoyama, S. An Application of p-adic convolutions. Mem. Fac. Sci. Kyushu Univ. Ser. A Math. 1982, 36, 73–83. [Google Scholar] [CrossRef] [Green Version]
  29. Simsek, Y. p-adic Dedekind and Hardy-Berndt sums related to Volkenborn integral on Z p . In Proceedings of the 16th International Conference on Finite or Infinite Dimensional Complex, Gyeongju, Korea, 28 July–1 August 2008. [Google Scholar]
  30. Simsek, Y. On twisted q-Hurwitz zeta function and q-two-variable L-function. Appl. Math. Comput. 2007, 187, 466–473. [Google Scholar] [CrossRef]
  31. Srivastava, H.M.; Kim, T.; Simsek, Y. q-Bernoulli numbers and polynomials associated with multiple q-zeta functions and basic L-series. Russ. J. Math. Phys. 2005, 12, 241–268. [Google Scholar]
  32. Srivastava, H.M.; Choi, J. Series Associated with the Zeta and Related Functions; Kluwer Acedemic Publishers: Dordrecht, The Netherlands; Boston, MA, USA; London, UK, 2001. [Google Scholar]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Bayad, A.; Simsek, Y. Multiple Dedekind Type Sums and Their Related Zeta Functions. Mathematics 2021, 9, 1744. https://doi.org/10.3390/math9151744

AMA Style

Bayad A, Simsek Y. Multiple Dedekind Type Sums and Their Related Zeta Functions. Mathematics. 2021; 9(15):1744. https://doi.org/10.3390/math9151744

Chicago/Turabian Style

Bayad, Abdelmejid, and Yilmaz Simsek. 2021. "Multiple Dedekind Type Sums and Their Related Zeta Functions" Mathematics 9, no. 15: 1744. https://doi.org/10.3390/math9151744

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop