A polynomial approach to Carlitz’s q q -Bernoulli numbers
Mohamed Mouzaia, Bakir FARHI · 2025
Read A polynomial approach to Carlitz’s q q -Bernoulli numbers by Mohamed Mouzaia, Bakir FARHI (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-20384.txt.
Introduction
Mohamed Mouzaia and Bakir Farhi take a polynomial approach to Carlitz's q-Bernoulli numbers in two thousand twenty-five. Bernoulli numbers anchor many sums and series in classic math. Their q analogues twist the classic forms by a parameter. The pair define a new polynomial sequence via the Jackson integral. At zero, its values match Carlitz's numbers exactly. The polynomials act as true q analogues of the classic ones. That restatement casts Carlitz's numbers in classic polynomial form. Fresh views of their traits follow from the bridge.
Catalog record
| CREATOR | Mohamed Mouzaia, Bakir FARHI |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 18 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-20384
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A polynomial approach to Carlitz’s q q -Bernoulli numbers Mohamed Mouzaia, Bakir FARHI Abstract This paper investigates q q -analogues of the classical Bernoulli polynomials and numbers. We introduce a new polynomial sequence ( B n , q ( X ) ) n ∈ ℕ 0 (B_{n,q}(X))_{n\in{\mathbb{N}}_{0}} , defined via the Jackson integral, and explore its connections with Carlitz’s q q -Bernoulli polynomials and numbers. Specifically, we prove that the numbers B n , q ( 0 ) B_{n,q}(0) are exactly the Carlitz q q -Bernoulli numbers and that the polynomials B n , q ( X ) B_{n,q}(X) are genuine q q -analogues of the classical Bernoulli polynomials. This approach leverages the Jackson integral to reformulate Carlitz’s q q -Bernoulli numbers in terms of classical polynomial structures, offering new insights into their properties. † † footnotetext: E-mail addresses : [email protected] (M. Mouzaia), [email protected] (B. Farhi). MSC 2020: Primary 05A30; Secondary 11B68. Keywords: Bernoulli polynomials and numbers, Carlitz’s q -Bernoulli polynomials and numbers, q -calculus, q -derivative, q -antiderivative, Jackson integral. 1 Introduction and Notation Throughout this paper, we let \mathbb{N} and \mathbb{N}_{0} denote the sets of positive and nonnegative integers, respectively. For n\in\mathbb{N}_{0} , we let \mathbb{R}_{n}[X] denote the \mathbb{R} -vector subspace of \mathbb{R}[X] consisting of polynomials of degree at most n . The Bernoulli polynomials and numbers, denoted B_{n}(X) and B_{n} ( n\in\mathbb{N}_{0} ), are defined via their exponential generating functions: \frac{t}{e^{t}-1}e^{Xt}=\sum_{n=0}^{\infty}B_{n}(X)\frac{t^{n}}{n!},\quad\frac{t}{e^{t}-1}=\sum_{n=0}^{\infty}B_{n}\frac{t^{n}}{n!}, so that B_{n}=B_{n}(0) for all n\in\mathbb{N}_{0} . These generating functions yield several key properties of the Bernoulli polynomials and numbers, including: \displaystyle B_{n}(X) \displaystyle=\sum_{k=0}^{n}\binom{n}{k}B_{k}X^{n-k}, (1.1) \displaystyle\int_{X}^{X+1}B_{n}(t)\,dt \displaystyle=X^{n}, (1.2) for all n\in\mathbb{N}_{0} . Notably, the second formula ( 1.2 ) characterizes the n -th Bernoulli polynomial: for all n\in\mathbb{N}_{0} , the unique real polynomial P(X) satisfying \int_{X}^{X+1}P(t)\,dt=X^{n} is P(X)=B_{n}(X) . As is well known, Bernoulli polynomials and numbers play a central role in number theory, combinatorics, and mathematical analysis, owing to their connections with sums of powers, zeta functions, and the Euler-Maclaurin formula [ 1 ] . For q>0 , q\neq 1 , the q -calculus generalizes classical calculus through the parameter q , recovering standard results as q\to 1 . The q -analogue of a number or indeterminate X is defined as [X]_{q}:=\frac{q^{X}-1}{q-1} . Fundamental concepts, introduced by Jackson [ 4 ] , are the q -derivative and q -integral of a function f (satisfying suitable regularity conditions), defined as: \displaystyle D_{q}f(X) \displaystyle:=\frac{f(qX)-f(X)}{(q-1)X}, \displaystyle\int_{a}^{b}f(t)\,d_{q}t \displaystyle:=\begin{cases}(1-q)b\sum_{n=0}^{\infty}q^{n}f(q^{n}b)-(1-q)a\sum_{n=0}^{\infty}q^{n}f(q^{n}a),&\text{if }0<q<1,\\ (q-1)b\sum_{n=0}^{\infty}q^{-n}f(q^{-n}b)-(q-1)a\sum_{n=0}^{\infty}q^{-n}f(q^{-n}a),&\text{if }q>1,\end{cases} for all a,b in a suitable domain ensuring the well-definition of the expressions. A q -antiderivative of a function f is a function F such that D_{q}F=f . The fundamental theorem of q -calculus [ 5 ] states that if f is Jackson-integrable on [a,b]\subset\mathbb{R} and F is a q -antiderivative of f , continuous at 0 , then \int_{a}^{b}f(t)\,d_{q}t=F(b)-F(a) . Next, we define a real q -polynomial as any polynomial in [X]_{q} with real coefficients, equivalently a real polynomial in q^{X} . The degree of a q -polynomial is its degree as a polynomial in [X]_{q} (or q^{X} ). The set of all real q -polynomials forms an \mathbb{R} -vector space with principal bases ([X]_{q}^{k})_{k\in\mathbb{N}_{0}} and (q^{kX})_{k\in\mathbb{N}_{0}} . For n\in\mathbb{N}_{0} , the set of real q -polynomials of degree at most n is an \mathbb{R} -vector subspace with principal bases ([X]_{q}^{k})_{0\leq k\leq n} , (q^{kX})_{0\leq k\leq n} , and (q^{kX}[X]_{q}^{n-k})_{0\leq k\leq n} , the latter being particularly important for our purposes. Like classical polynomials, two real q -polynomials that coincide on an infinite subset of \mathbb{R} are identical (for q>0 , q\neq 1 ). To distinguish polynomials (resp. q -polynomials) from numbers, we consistently indicate the indeterminate X for polynomials (resp. q -polynomials). In q -calculus, some authors investigate q -analogues of well-known polynomial sequences, such as Bernoulli, Euler, and Chebyshev polynomials, in the form of q -polynomial sequences. For example, in [ 2 ] , Carlitz introduced the q -polynomial sequences (\eta_{n,q}(X))_{n\in\mathbb{N}_{0}} and (\beta_{n,q}(X))_{n\in\mathbb{N}_{0}} , defined by: \displaystyle\eta_{n,q}(X) \displaystyle=(q-1)^{-n}\sum_{k=0}^{n}(-1)^{n-k}\binom{n}{k}\frac{k}{[k]_{q}}q^{kX}, (1.3) \displaystyle\beta_{n,q}(X) \displaystyle=(q-1)^{-n}\sum_{k=0}^{n}(-1)^{n-k}\binom{n}{k}\frac{k+1}{[k+1]_{q}}q^{kX}, (1.4) for all n\in\mathbb{N}_{0} , with the convention \frac{k}{[k]_{q}}=1 for k=0 . Carlitz showed that the second q -polynomial sequence (\beta_{n,q}(X))_{n\in{\mathbb{N}}_{0}} converges to the classical Bernoulli polynomial sequence as q\to 1 , i.e., \lim_{q\to 1}\beta_{n,q}(X)=B_{n}(X) for all n\in\mathbb{N}_{0} . He also considered the values of \eta_{n,q}(X) and \beta_{n,q}(X) at X=0 , respectively denoted by \eta_{n,q} and \beta_{n,q} , and proved the important formulas: \displaystyle\eta_{n,q}(X) \displaystyle=\sum_{k=0}^{n}\binom{n}{k}\eta_{k,q}q^{kX}[X]_{q}^{n-k}, (1.5) \displaystyle\beta_{n,q}(X) \displaystyle=\sum_{k=0}^{n}\binom{n}{k}\beta_{k,q}q^{kX}[X]_{q}^{n-k}, (1.6) for all n\in\mathbb{N}_{0} . Thus, \beta_{n,q} are q -analogues of the Bernoulli numbers B_{n} , and Formula ( 1.6 ) is a q -analogue of Formula ( 1.1 ). For instance, the first four Carlitz q -Bernoulli numbers are \beta_{0,q}=1~,~\beta_{1,q}=-\dfrac{1}{[2]_{q}}~,~\beta_{2,q}=\dfrac{q}{[2]_{q}[3]_{q}}~,~\text{ and }\beta_{3,q}=-\dfrac{q(q-1)}{[3]_{q}[4]_{q}}. (1.7) For further reading on Carlitz’s q -Bernoulli numbers and polynomials and their extensions, see, e.g., [ 3 ] . This paper introduces, via the Jackson integral, a new sequence of polynomials, (B_{n,q}(X))_{n\in{\mathbb{N}}_{0}} , and studies its relationship with Carlitz’s q -polynomial sequences (\eta_{n,q}(X))_{n\in\mathbb{N}_{0}} and (\beta_{n,q}(X))_{n\in\mathbb{N}_{0}} . Our objectives are twofold: first, to establish that B_{n,q}(0)=\beta_{n,q} for all n\in\mathbb{N}_{0} ; second, to prove that B_{n,q}(X) are genuine q -analogues of the Bernoulli polynomials, that is, \lim_{q\to 1}B_{n,q}(X)=B_{n}(X) for all n\in{\mathbb{N}}_{0} . This approach provides a novel perspective on Carlitz’s q -Bernoulli numbers \beta_{n,q} by leveraging the Jackson integral to reformulate their associated q -polynomial sequences in terms of classical polynomial structures. 2 The results and the proofs Our q -Bernoulli polynomials B_{n,q}(X) are defined through the following proposition. Proposition 2.1 . For all n\in{\mathbb{N}}_{0} , there exists a unique real polynomial B_{n,q}(X) satisfying: \int_{X}^{qX+1}B_{n,q}(t)\,d_{q}t=(q-1)X^{n+1}+X^{n}. (2.1) Proof. Fix n\in{\mathbb{N}}_{0} . Comparing the degrees of the polynomials on both sides of ( 2.1 ), we see that if B_{n,q}(X) exists, it must be of degree n ; so B_{n,q}(X)\in{\mathbb{R}}_{n}[X] . To establish the existence and the uniqueness of B_{n,q}(X) in {\mathbb{R}}_{n}[X] , consider the {\mathbb{R}} -vector subspace E_{n} of {\mathbb{R}}_{n+1}[X] defined by: E_{n}:=\left\{P\in{\mathbb{R}}_{n+1}[X]:~P\left(\frac{1}{1-q}\right)=0\right\} and the linear mapping \varphi_{n} defined by: \begin{array}[]{rcl}\varphi_{n}:~{\mathbb{R}}_{n}[X]&\longrightarrow&E_{n}\\ P&\longmapsto&\int_{X}^{qX+1}P(t)\,d_{q}t.\end{array} We claim that \varphi_{n} is an isomorphism of {\mathbb{R}} -vector spaces. First, we show that \varphi_{n} is injective. Suppose P\in\ker{\varphi_{n}} , so that \varphi_{n}(P)=\int_{X}^{qX+1}P(t)\,d_{q}t=0_{{\mathbb{R}}[X]}. Evaluating at X=[k]_{q} ( k\in{\mathbb{N}}_{0} ), we have q[k]_{q}+1=[k+1]_{q} , so \int_{[k]_{q}}^{[k+1]_{q}}P(t)\,d_{q}t=0. Summing these integrals from k=0 to k=N-1 for N\in{\mathbb{N}} , we obtain \int_{0}^{[N]_{q}}P(t)\,d_{q}t=0\quad(\forall N\in{\mathbb{N}}). Let Q be the polynomial q -antiderivative of P such that Q(0)=0 , i.e., Q(X)=\int_{0}^{X}P(t)\,d_{q}t . Then, the above implies Q([N]_{q})=0 for all N\in{\mathbb{N}} . Since the points [N]_{q} ( N\in{\mathbb{N}} ) are pairwise distinct (because q>0 and q\neq 1 ), it follows that Q=0_{{\mathbb{R}}[X]} . Thus P=D_{q}Q=0_{{\mathbb{R}}[X]} . Hence \ker{\varphi_{n}}=\{0_{{\mathbb{R}}_{n}[X]}\} , and \varphi_{n} is injective. Furthermore, since \dim{{\mathbb{R}}_{n}[X]}=\dim{E_{n}}=n+1 , then \varphi_{n} is an isomorphism, as we claimed it to be. Consequently, the polynomial (q-1)X^{n+1}+X^{n}=X^{n}\left(1-(1-q)X\right) (which clearly belongs to E_{n} ) has a unique preimage by \varphi_{n} . This ensures the existence and the uniqueness of B_{n,q}(X) in {\mathbb{R}}_{n}[X] and completes the proof. ∎ Remark 2.2 . Assuming the existence of the limit \lim_{q\to 1}B_{n,q}(t) , which we provisionally denote by \widetilde{B}_{n}(t) , for n\in{\mathbb{N}}_{0} , by setting q\to 1 in Formula ( 2.1 ), we obtain \int_{X}^{X+1}\widetilde{B}_{n}(t)\,dt=X^{n}, which is one of the characterizing properties of the n th Bernoulli polynomial B_{n}(t) . Thus, \lim_{q\to 1}B_{n,q}(t)=B_{n}(t) , justifying the terminology “ q -Bernoulli polynomials” for the B_{n,q}(t) ’s. However, establishing