Quantum integers and cyclotomy

A.v. Borisov, Melvyn B. Nathanson, Yang WangJournal of Number Theory 2004 · DOI

TL;DR

This paper classifies all rational functions that satisfy a specific functional equation related to the multiplication of "quantum integers." The authors demonstrate that these functions can be uniquely represented in a structured form involving cyclotomic polynomials, providing a complete characterization of these solutions. This work offers a deeper understanding of algebraic structures arising from number theory and abstract algebra.

What Problem Does This Paper Address?

The paper tackles a question posed by M. B. Nathanson concerning the classification of sequences of polynomials and rational functions that satisfy a particular functional equation: $f_{mn}(q) = f_m(q)f_n(q^m)$. This equation is inspired by the properties of "quantum integers," which are defined as polynomials of the form $[n]_q = 1 + q + q^2 + \dots + q^{n-1}$. The core problem is to understand the structure of all possible sequences of rational functions $\{f_n(q)\}_{n=1}^\infty$ that adhere to this multiplication rule. The authors focus on "nontrivial" solutions where $f_1(q) = 1$, excluding the trivial case where all $f_n(q) = 0$.

Methods

The authors employ a combination of algebraic number theory and abstract algebra techniques. Key methods include:

  • Support of a Sequence: Defining the "support" of a sequence $F = \{f_n(q)\}_{n=1}^\infty$ as the set of integers $n$ for which $f_n(q) \neq 0$. They show this support is always a multiplicative semigroup generated by a set of prime numbers, $S(P)$.
  • Logarithm Group: Introducing the concept of a "logarithm group" $L(K) = K^*/\Gamma$, where $K$ is an algebraically closed field and $\Gamma$ is the group of roots of unity. This group is shown to be a vector space over the rational numbers $\mathbb{Q}$. This abstract structure is crucial for analyzing the behavior of zeros and poles.
  • Cyclotomic Polynomials: Utilizing cyclotomic polynomials $\Phi_k(q)$ and related polynomials $F_k(q) = q^k - 1$. These are fundamental building blocks in the representation of the solutions.
  • Multiset Representation: Representing each function $f_n(q)$ as a quotient of products of $F_k(q)$ polynomials, indexed by multisets of positive integers. This allows for a unique and structured representation of the solutions.
  • Induction: Employing inductive arguments to build up the general form of the solutions from simpler cases, particularly for prime numbers in the support set.

Datasets

Not discussed in the paper. The paper deals with theoretical structures and functional equations rather than empirical data or computational experiments.

Key Insights and Ideas

  • Structure of Support: The set of integers $n$ for which $f_n(q)$ is non-zero (the support) is always a multiplicative semigroup generated by a set of primes. This significantly constrains the possible solutions.
  • Zeros and Poles are Roots of Unity: A crucial finding is that for any nontrivial solution, the zeros and poles of $f_n(q)$ (for $n$ in the support) must be either 0 or roots of unity. This is a strong structural property derived from the functional equation.
  • Unique Representation: Every rational function solution can be uniquely expressed in a specific form involving a completely multiplicative arithmetic function $\lambda(n)$, a rational exponent $t_0$, and a finite product of cyclotomic polynomials (or related $F_k(q)$ terms). This provides a complete classification.
  • Connection to Cyclotomy: The solutions are deeply intertwined with cyclotomic polynomials, highlighting a connection between abstract functional equations and classical number theory objects.

Key Results

The paper provides a complete classification of all nontrivial rational function solutions to the functional equation $f_{mn}(q) = f_m(q)f_n(q^m)$. Specifically, for a solution $F = \{f_n(q)\}_{n=1}^\infty$ with support $S(P)$ (the semigroup generated by primes $P$), there exist:

  1. A completely multiplicative arithmetic function $\lambda(n)$ with support $S(P)$.
  2. A rational number $t_0$ such that $t_0(n-1)$ is an integer for all $n \in S(P)$.
  3. A finite set $R$ of positive integers and a set $\{t_r\}_{r \in R}$ of integers.

Such that for every $n \in S(P)$, $f_n(q)$ can be written as: $f_n(q) = \lambda(n) q^{t_0(n-1)} \prod_{r \in R} \left(\frac{q^r - 1}{q - 1}\right)^{t_r}$

The paper also proves that all zeros and poles of these functions must be roots of unity or zero.

Limitations and Future Work

The authors identify two main open problems:

  • Polynomial Solutions: They seek a simple criterion to determine when a rational function solution is actually a polynomial. While having non-negative exponents $t_r$ is sufficient, it's not necessary, as shown by an example.
  • General Field: They aim to extend their structure theorem for rational function and polynomial solutions to coefficients in an arbitrary field, not just the field of rational numbers $\mathbb{Q}$.

Summary updated

BibTeX
@article{borisov2004quantum,
  author = {A.V. Borisov and Melvyn B. Nathanson and Yang Wang},
  title = {{Quantum integers and cyclotomy}},
  journal = {Journal of Number Theory},
  year = {2004},
  doi = {10.1016/j.jnt.2004.06.015},
}

Loading related content...

Figure