Abstract

The purpose of this paper is to give some properties of the modified 𝑞-Bernoulli numbers and polynomials of higher order with weight. In particular, by using the bosonic 𝑝-adic 𝑞-integral on 𝑝, we derive new identities of 𝑞-Bernoulli numbers and polynomials with weight.

1. Introduction

Let 𝑝 be a fixed odd prime number. Throughout this paper 𝑝,𝑝, and 𝑝 will, respectively, denote the ring of 𝑝-adic rational integers, the field of 𝑝-adic rational numbers, and the completion of the algebraic closure of 𝑝. Let be the set of natural numbers and +={0}. The 𝑝-adic norm of 𝑝 is defined by |𝑝|𝑝=1/𝑝. When one talks of a 𝑞-extension, 𝑞 can be considered as an indeterminate, a complex number 𝑞, or a 𝑝-adic number 𝑞𝑝. Throughout this paper we assume that 𝛼 and 𝑞𝑝 with |1𝑞|𝑝<𝑝1/(𝑝1) so that 𝑞𝑥=exp(𝑥log𝑞).

Let 𝑈𝐷(𝑝) be the space of uniformly differentiable functions on 𝑝. For 𝑓𝑈𝐷(𝑝), the 𝑝-adic 𝑞-integral on 𝑝 is defined by Kim (see [13]) as follows:𝐼𝑞(𝑓)=𝑝𝑓(𝑥)𝑑𝜇𝑞(𝑥)=lim𝑁1𝑝𝑁𝑞𝑝𝑁1𝑥=0𝑓(𝑥)𝑞𝑥,(1.1) where [𝑥]𝑞 is the 𝑞-number of 𝑥 which is defined by [𝑥]𝑞=(1𝑞𝑥)/(1𝑞).

From (1.1), we have𝑞𝑛𝐼𝑞𝑓𝑛𝐼𝑞(𝑓)=(𝑞1)𝑛1𝑙=0𝑞𝑙𝑓(𝑙)+𝑞1log𝑞𝑛1𝑙=0𝑞𝑙𝑓(𝑙),(1.2) where 𝑓𝑛(𝑥)=𝑓(𝑥+𝑛) (see [24]).

As is well known, Bernoulli numbers are inductively defined by 𝐵0=1,(𝐵+1)𝑛𝐵𝑛=1if𝑛=1,0if𝑛>1,(1.3) with the usual convention about replacing 𝐵𝑛 by 𝐵𝑛 (see [3, 5]).

In [2, 5, 6], the 𝑞-Bernoulli numbers are defined by𝐵0,𝑞=𝑞1log𝑞,𝑞𝐵𝑞+1𝑛𝐵𝑛,𝑞=1if𝑛=1,0if𝑛>1,(1.4) with the usual convention about replacing 𝐵𝑛𝑞 by 𝐵𝑛,𝑞. Note that lim𝑞1𝐵𝑛,𝑞=𝐵𝑛. In the viewpoint of (1.4), we consider the modified 𝑞-Bernoulli numbers with weight.

In this paper we study families of the modified 𝑞-Bernoulli numbers and polynomials of higher order with weight. In particular, by using the multivariate 𝑝-adic 𝑞-integral on 𝑝, we give new identities of the higher-order 𝑞-Bernoulli numbers and polynomials with weight.

2. Modified 𝑞-Bernoulli Numbers with Weight of Higher Order

For 𝑛+, let us consider the following modified 𝑞-Bernoulli numbers with weight 𝛼 (see [1, 3]):𝐵(𝛼)𝑛,𝑞=𝑝[𝑥]𝑛𝑞𝛼𝑞𝑥𝑑𝜇𝑞(𝑥)=1(1𝑞)𝑛[𝛼]𝑛𝑞𝑛𝑙=0𝑛𝑙(1)𝑙𝛼𝑙[𝛼𝑙]𝑞,𝐵(𝛼)𝑛,𝑞(𝑥)=𝑝[𝑥+𝑦]𝑛𝑞𝛼𝑞𝑦𝑑𝜇𝑞(𝑦)=1(1𝑞)𝑛[𝛼]𝑛𝑞𝑛𝑙=0𝑛𝑙(1)𝑙𝑞𝛼𝑙𝑥𝛼𝑙[𝛼𝑙]𝑞.(2.1) From (2.1), we note that𝐵(𝛼)𝑛,𝑞(𝑥)=𝑛𝑙=0𝑛𝑙[𝑥]𝑛𝑙𝑞𝛼𝑞𝛼𝑙𝑥𝐵(𝛼)𝑙,𝑞(2.2) (see [1, 3]).

For 𝑘 and 𝑛+, by making use of the multivariate 𝑝-adic 𝑞-integral on 𝑝, we consider the following modified 𝑞-Bernoulli numbers with weight 𝛼 of order 𝑘, 𝐵(𝑘,𝛼)𝑛,𝑞:𝐵(𝑘,𝛼)𝑛,𝑞=𝑝𝑝𝑥1++𝑥𝑘𝑛𝑞𝛼𝑞𝑥1𝑥𝑘𝑑𝜇𝑞𝑥1𝑑𝜇𝑞𝑥𝑘.(2.3) Note that 𝐵(1,𝛼)𝑛,𝑞=𝐵(𝛼)𝑛,𝑞 and lim𝑞1𝐵(𝑘,𝛼)𝑛,𝑞=𝐵(𝑘)𝑛, where 𝐵(𝑘)𝑛 are the 𝑛th ordinary Bernoulli numbers of order 𝑘.

For 𝑘,𝑁, we have1𝑞1𝑞𝑝𝑁𝑘𝑝𝑁1𝑖1=0𝑝𝑁1𝑖𝑘=0𝑖1++𝑖𝑘𝑛𝑞𝛼=1𝑞1𝑞𝑝𝑁𝑘11𝑞𝛼𝑛𝑝𝑁1𝑖1,,𝑖𝑘=0𝑛𝑗=0𝑛𝑗(1)𝑗𝑞𝛼(𝑖1++𝑖𝑘)𝑗=1(1𝑞)𝑛[𝛼]𝑛𝑞𝑛𝑗=0𝑛𝑗(1)𝑗(1𝑞)𝑘1𝑞𝑝𝑁𝑘1𝑞𝛼𝑝𝑁𝑗1𝑞𝛼𝑗1𝑞𝛼𝑝𝑁𝑗1𝑞𝛼𝑗𝑘times.(2.4) By (1.1), (2.3), and (2.4), we get𝐵(𝑘,𝛼)𝑛,𝑞=1(1𝑞)𝑛[𝛼]𝑛𝑞𝑛𝑗=0𝑛𝑗(1)𝑗(𝛼𝑗)𝑘[𝛼𝑗]𝑘𝑞.(2.5) Therefore, by (2.5), we obtain the following theorem.

Theorem 2.1. For 𝑛0, one has 𝐵(𝑘,𝛼)𝑛,𝑞=1(1𝑞)𝑛[𝛼]𝑛𝑞𝑛𝑗=0𝑛𝑗(1)𝑗(𝛼𝑗)𝑘[𝛼𝑗]𝑘𝑞.(2.6)

Let us consider the modified 𝑞-Bernoulli and polynomials with weight 𝛼 of order 𝑘 as follows:𝐵(𝑘,𝛼)𝑛,𝑞(𝑥)=𝑝𝑝𝑥+𝑥1++𝑥𝑘𝑛𝑞𝛼𝑞𝑥1𝑥𝑘𝑑𝜇𝑞𝑥1𝑑𝜇𝑞𝑥𝑘.(2.7) By the same method of (2.5), we obtain the following theorem.

Theorem 2.2. For 𝑛+, one has 𝐵(𝑘,𝛼)𝑛,𝑞(𝑥)=1(1𝑞)𝑛[𝛼]𝑛𝑞𝑛𝑗=0𝑛𝑗(1)𝑗𝑞𝛼𝑥𝑗(𝛼𝑗)𝑘[𝛼𝑗]𝑘𝑞.(2.8)

By Theorem 2.2, we get𝐵(𝑘,𝛼)𝑛,𝑞1(𝑘𝑥)=1(1𝑞𝛼)𝑛𝑛𝑗=0𝑛𝑗(1)𝑗(𝛼𝑗)𝑘[𝛼𝑗]𝑘𝑞1𝑞𝛼𝑗(𝑘𝑥)=(1)𝑛𝑞𝛼𝑛(1𝑞𝛼)𝑛𝑛𝑗=0𝑛𝑗(1)𝑗𝑞1(𝑞1)𝛼𝑗(𝑞𝛼𝑗1)𝑞𝛼𝑗𝑘𝑞𝛼𝑗(𝑘𝑥)=(1)𝑛𝑞𝛼𝑛(1𝑞𝛼)𝑛𝑛𝑗=0𝑛𝑗(1)𝑗𝑞𝛼𝑗𝑥𝑞𝑘(𝛼𝑗)𝑘[𝛼𝑗]𝑘𝑞=(1)𝑛𝑞𝛼𝑛𝑘𝐵(𝑘,𝛼)𝑛,𝑞(𝑥).(2.9) Therefore, by (2.9), we obtain the following theorem.

Theorem 2.3. For 𝑛+, one has 𝐵(𝑘,𝛼)𝑛,𝑞1(𝑘𝑥)=(1)𝑛𝑞𝛼𝑛𝑘𝐵(𝑘,𝛼)𝑛,𝑞(𝑥),𝐵(𝑘,𝛼)𝑛,𝑞1(𝑘)=(1)𝑛𝑞𝛼𝑛𝑘𝐵(𝑘,𝛼)𝑛,𝑞.(2.10)

From Theorem 2.3, we note thatlim𝑞1𝐵(𝑘,𝛼)𝑛,𝑞1(𝑘𝑥)=𝐵(𝑘)𝑛(𝑘𝑥),lim𝑞1𝐵(𝑘,𝛼)𝑛,𝑞1(𝑘)=(1)𝑛𝐵(𝑘)𝑛.(2.11) Thus, we have 𝐵(𝑘)𝑛(𝑘)=(1)𝑛𝐵(𝑘)𝑛, where 𝐵(𝑘)𝑛 are the 𝑛th Bernoulli numbers of order 𝑘.

From (2.3) and (2.7), we can derive the following equations:𝐵(𝑙,𝛼)𝑘,𝑞(𝑥)=lim𝑁1[𝑚]𝑙𝑞𝑝𝑁𝑙𝑞𝑚𝑚1𝑖1,,𝑖𝑙=0𝑝𝑁1𝑛1,,𝑛𝑙=0𝑥+𝑖1++𝑖𝑙+𝑚𝑛1++𝑛𝑙𝑘𝑞𝛼=[𝑚]𝑘𝑞𝛼[𝑚]𝑙𝑞𝑚1𝑖1,,𝑖𝑙=0𝑝𝑝𝑥+𝑖1++𝑖𝑙𝑚+𝑥1++𝑥𝑙𝑘𝑞𝛼𝑚×𝑞𝑚𝑥1𝑚𝑥𝑙𝑑𝜇𝑞𝑚𝑥1𝑑𝜇𝑞𝑚𝑥𝑘=[𝑚]𝑘𝑞𝛼[𝑚]𝑙𝑞𝑚1𝑖1,,𝑖𝑙=0𝐵(𝑙,𝛼)𝑘,𝑞𝑚𝑥+𝑖1++𝑖𝑙𝑚.(2.12) Therefore, by (2.12), we obtain the following theorem.

Theorem 2.4. For 𝑘+ and 𝑙,𝑚, one has 𝐵(𝑙,𝛼)𝑘,𝑞(𝑥)=[𝑚]𝑘𝑞𝛼[𝑚]𝑙𝑞𝑚1𝑖1,,𝑖𝑙=0𝐵(𝑙,𝛼)𝑘,𝑞𝑚𝑥+𝑖1++𝑖𝑙𝑚.(2.13)

In particular,𝐵(𝑙,𝛼)𝑘,𝑞(𝑚𝑥)=[𝑚]𝑘𝑞𝛼[𝑚]𝑙𝑞𝑚1𝑖1,,𝑖𝑙=0𝐵(𝑙,𝛼)𝑘,𝑞𝑚𝑥+𝑖1++𝑖𝑙𝑚.(2.14)

From (1.2), we can derive the following integral:𝑝𝑓(𝑥+1)𝑞𝑥𝑑𝜇𝑞(𝑥)=𝑝𝑓(𝑥)𝑞𝑥𝑑𝜇𝑞(𝑥)+𝑞1log𝑞𝑓(0),𝑝𝑓(𝑥+2)𝑞𝑥𝑑𝜇𝑞(𝑥)=𝑝𝑓1(𝑥)𝑞𝑥𝑑𝜇𝑞(𝑥)+𝑞1log𝑞𝑓(1)=𝑝𝑓(𝑥)𝑞𝑥𝑑𝜇𝑞(𝑥)+𝑞1log𝑞𝑓(0)+𝑓(1).(2.15) Continuing this process, we obtain𝑝𝑓(𝑥+𝑛)𝑞𝑥𝑑𝜇𝑞(𝑥)=𝑝𝑓(𝑥)𝑞𝑥𝑑𝜇𝑞(𝑥)+𝑞1log𝑞𝑛1𝑙=0𝑓(𝑙).(2.16) By (2.16), we get𝑝[𝑥+𝑛]𝑚𝑞𝛼𝑞𝑥𝑑𝜇𝑞(𝑥)=𝑝[𝑥]𝑚𝑞𝛼𝑞𝑥𝑑𝜇𝑞(𝑥)+𝑚𝛼[𝛼]𝑞𝑛1𝑙=0[𝑙]𝑚1𝑞𝛼𝑞𝛼𝑙.(2.17) Therefore, by (2.1) and (2.17), we obtain the following theorem.

Theorem 2.5. For 𝑛 and 𝑚+, one has 𝐵(𝛼)𝑚,𝑞(𝑛)𝐵(𝛼)𝑚,𝑞=𝑚𝛼[𝛼]𝑞𝑛1𝑙=0[𝑙]𝑚𝑞𝛼𝑞𝛼𝑙.(2.18)

In an analogues manner as the previous investigation [710], we can define a further generalization of modified 𝑞-Bernoulli numbers with weight. Let 𝜒 be the Dirichlet character with conductor 𝑑. Then the generalized 𝑞-Bernoulli numbers with weight attached to 𝜒 can be defined as follows:𝐵(𝛼)𝑛,𝜒,𝑞=𝑋𝜒(𝑥)[𝑥]𝑛𝑞𝛼𝑞𝑥𝑑𝜇𝑞(𝑥)=[𝑑]𝑛𝑞𝛼[𝑑]𝑞𝑑1𝑎=0𝜒(𝑎)𝐵(𝛼)𝑛,𝑞𝑑𝑎𝑑.(2.19) We expect to investigate these objects in future papers. This definition 𝐵(𝛼)𝑛,𝑞 was also given in a previous paper (see [9]).

Acknowledgments

The authors express their sincere gratitude to the referees for their valuable suggestions and comments. This paper is supported in part by the Research Grant of Kwangwoon University in 2011.