Random walks and quadratic number fields

2025

Read Random walks and quadratic number fields (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-2512-03884.txt.

Introduction

A walker on a circle who jumps by random whole multiples of a fixed irrational step will, over time, visit every arc fairly. Bence Borda asks how fast that fairness arrives in a 2025 probability preprint. The answer lives in pure number theory: the step is a quadratic irrational, satisfying a quadratic equation with integer coefficients.

Borda measures convergence to uniformity with the quadratic Wasserstein metric, also read as a periodic discrepancy. He proves the rate is governed by deep arithmetic features of the number field the step generates. Fundamental units of its ring of algebraic integers and special values of zeta functions both enter the story. The result bridges random walks and analytic number theory. A question about shuffling points on a circle is answered by invariants invented for algebraic numbers. Such links often prove useful in both directions.

Catalog record

YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE127 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-03884

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Random walks and quadratic number fields Abstract We establish a novel type of connection between random walks and analytic number theory. Working with a random walk on the circle group ℝ / ℤ \mathbb{R}/\mathbb{Z} in which each step is a random integer multiple of a given quadratic irrational α \alpha , we show that the rate of convergence to uniformity in the quadratic Wasserstein metric (also known as the periodic L 2 L^{2} discrepancy) is governed by deep arithmetic invariants of the ring of algebraic integers of the real quadratic field ℚ ⁡ ( α ) \mathbb{Q}(\alpha) , such as fundamental units and special values of zeta functions. Bence Borda University of Sussex Brighton, BN1 9RH, United Kingdom Email: [email protected] Keywords: Wasserstein metric, L^{2} discrepancy, quadratic irrational, zeta function, Diophantine approximation Mathematics Subject Classification (2020): 60G50, 11K38, 11E45, 11R42 1 Introduction The behavior of a random walk on the circle group is governed by a delicate interplay between classical probability, Fourier analysis and Diophantine approximation. The main goal of this paper is to establish a further, rather surprising connection to analytic number theory. As a prototypical example, we work with a random walk in which each step is a random integer multiple of a given quadratic irrational \alpha . We show that the rate of convergence of such a random walk to uniformity is inextricably linked to the arithmetic of the real quadratic field \mathbb{Q}(\alpha) . In particular, we demonstrate that the random walk is able to detect deep arithmetic invariants such as fundamental units in the ring of algebraic integers, and special values of zeta functions. To state our main result, let X_{1},X_{2},\ldots be independent, identically distributed (i.i.d.) integer-valued random variables, and set S_{n}=\sum_{k=1}^{n}X_{k} . Given \alpha\in\mathbb{R} , the sequence S_{n}\alpha\pmod{\mathbb{Z}} is then a random walk on the circle group \mathbb{T}=\mathbb{R}/\mathbb{Z} . The empirical measure \mu_{N}=N^{-1}\sum_{n=1}^{N}\delta_{S_{n}\alpha\pmod{\mathbb{Z}}} converges to the normalized Haar measure \lambda on \mathbb{T} under very general assumptions. Our main result is the precise asymptotics of the expected value and the variance of the distance squared from the empirical measure to uniformity in the quadratic Wasserstein metric W_{\mathbb{T},2} on \mathbb{T} . See Section 2 for definitions, notation and further context. Theorem 1 . Assume \mathbb{E}(X_{1})=0 and 0<\mathbb{E}(X_{1}^{2})<\infty , and \alpha is a quadratic irrational. Then \mathbb{E}\left(W_{\mathbb{T},2}^{2}(\mu_{N},\lambda)\right)\sim\frac{L^{2}c_{1}(L\alpha)}{\sigma^{2}}\cdot\frac{\log N}{N}\quad\textrm{and}\quad\mathrm{Var}\left(W_{\mathbb{T},2}^{2}(\mu_{N},\lambda)\right)\sim\frac{L^{4}c_{2}(L\alpha)}{\sigma^{4}}\cdot\frac{\log N}{N^{2}} with L=\mathrm{gcd}(\mathrm{supp}(X_{1})) , \sigma^{2}=\mathbb{E}(X_{1}^{2}) and explicit positive constants c_{1}(L\alpha) , c_{2}(L\alpha) . If in addition \mathbb{E}(|X_{1}|^{p})<\infty with some real constant p>2 , then \begin{split}\mathbb{E}\left(W_{\mathbb{T},2}^{2}(\mu_{N},\lambda)\right)&=\frac{L^{2}c_{1}(L\alpha)}{\sigma^{2}}\cdot\frac{\log N}{N}+O\left(\frac{1}{N}\right),\\ \mathrm{Var}\left(W_{\mathbb{T},2}^{2}(\mu_{N},\lambda)\right)&=\frac{L^{4}c_{2}(L\alpha)}{\sigma^{4}}\cdot\frac{\log N}{N^{2}}+O\left(\frac{(\log\log N)^{2}}{N^{2}}\right)\end{split} with implied constants depending only on \alpha and the distribution of X_{1} . In the terminology of probability theory, the sequence S_{n}\alpha\pmod{\mathbb{Z}} is a Markov chain on \mathbb{T} with stationary measure \lambda , started from a nonstationary distribution, indeed from the Dirac measure at 0 . The most interesting feature of this sequence is that it is at the critical threshold of a phase transition between weakly and strongly dependent sequences, which explains why the rates of \mathbb{E}(W_{\mathbb{T},2}^{2}(\mu_{N},\lambda)) and \mathrm{Var}(W_{\mathbb{T},2}^{2}(\mu_{N},\lambda)) differ from uniformly distributed i.i.d. samples by a factor of \log N . The phase transition is discussed in more detail in Section 2.5 . The connection to number theory comes from the values of the constants c_{1} and c_{2} in Theorem 1 . For instance, their values at the golden ratio are c_{1}\left(\frac{1+\sqrt{5}}{2}\right)=\frac{\sqrt{5}}{150\log\frac{1+\sqrt{5}}{2}}\quad\textrm{and}\quad c_{2}\left(\frac{1+\sqrt{5}}{2}\right)=\frac{\sqrt{5}}{6750\log\frac{1+\sqrt{5}}{2}}. Several more examples are listed in Table 1 . Finding the explicit values of c_{1}(\alpha) and c_{2}(\alpha) is far from easy, and relies on the theory of binary quadratic forms due to Gauss, as well as arithmetic properties of rings of algebraic integers of real quadratic fields. A real number \alpha is called a quadratic irrational if it is the root of an irreducible polynomial of degree 2 with integer coefficients. There is a unique minimal polynomial ax^{2}+bx+c with a,b,c\in\mathbb{Z} , \mathrm{gcd}(a,b,c)=1 and a>0 . Then the discriminant D=b^{2}-4ac>0 is not a perfect square, and \alpha is an element of the real quadratic field \mathbb{Q}(\sqrt{D}) . Rather surprisingly, it turns out that the constants c_{1}(\alpha) and c_{2}(\alpha) can be expressed in terms of special values of zeta functions \zeta(A,s) of \mathbb{Z} -modules A in \mathbb{Q}(\sqrt{D}) . We refer to Section 4 for definitions and further context. Theorem 2 . Let \alpha be a quadratic irrational, and let D be the discriminant of its minimal polynomial. Let A=\mathbb{Z}+\mathbb{Z}\alpha , and let \varepsilon be the smallest totally positive unit in the order \{\xi\in\mathbb{Q}(\sqrt{D}):\xi A\subseteq A\} such that \varepsilon>1 . Then c_{1}(\alpha)=\frac{2\zeta(A,-1)}{D^{1/2}\log\varepsilon}\quad\textrm{and}\quad c_{2}(\alpha)=\frac{4\zeta(A,-3)}{9D^{3/2}\log\varepsilon}. \alpha \sqrt{2} \sqrt{3} \sqrt{5} \sqrt{6} \sqrt{7} c_{1}(\alpha) \frac{\sqrt{2}}{48\log(1+\sqrt{2})} \frac{\sqrt{3}}{36\log(2+\sqrt{3})} \frac{\sqrt{5}}{100\log\frac{1+\sqrt{5}}{2}} \frac{\sqrt{6}}{24\log(5+2\sqrt{6})} \frac{\sqrt{7}}{21\log(8+3\sqrt{7})} c_{2}(\alpha) \frac{11\sqrt{2}}{17280\log(1+\sqrt{2})} \frac{23\sqrt{3}}{19440\log(2+\sqrt{3})} \frac{43\sqrt{5}}{54000\log\frac{1+\sqrt{5}}{2}} \frac{29\sqrt{6}}{8640\log(5+2\sqrt{6})} \frac{113\sqrt{7}}{26460\log(8+3\sqrt{7})} \alpha \sqrt{10} \sqrt{11} \sqrt{13} \sqrt{14} c_{1}(\alpha) \frac{47\sqrt{10}}{1200\log(3+\sqrt{10})} \frac{7\sqrt{11}}{132\log(10+3\sqrt{11})} \frac{\sqrt{13}}{52\log\frac{3+\sqrt{13}}{2}} \frac{5\sqrt{14}}{84\log(15+4\sqrt{14})} c_{2}(\alpha) \frac{2897\sqrt{10}}{432000\log(3+\sqrt{10})} \frac{2153\sqrt{11}}{261360\log(10+3\sqrt{11})} \frac{1247\sqrt{13}}{365040\log\frac{3+\sqrt{13}}{2}} \frac{2503\sqrt{14}}{211680\log(15+4\sqrt{14})} Table 1: The values of c_{1}(\alpha) and c_{2}(\alpha) for some quadratic irrational \alpha Theorem 2 provides a way to explicitly compute the constants c_{1}(\alpha) and c_{2}(\alpha) for any quadratic irrational \alpha . Indeed, \varepsilon can be expressed in terms of the smallest positive solution of Pell’s equation x^{2}-Dy^{2}=4 , whereas the special values of \zeta(A,s) at negative integers s are rational, and were first found by Shintani [ 42 ] , and later by Zagier [ 50 ] using a backward continued fraction algorithm. As an immediate corollary, we see that if \alpha\in\mathbb{Q}(\sqrt{d}) with some squarefree integer d>1 and \eta>1 is the fundamental unit in the ring of algebraic integers, then c_{1}(\alpha),c_{2}(\alpha)\in(\sqrt{d}/\log\eta)\mathbb{Q} . We will also show that c_{1} and c_{2} , as functions defined on the set of quadratic irrationals, satisfy the functional equations c_{j}(\alpha)=c_{j}(\alpha+1)=c_{j}(-\alpha)=c_{j}(1/\alpha),\qquad j=1,2. (1) In particular, c_{1}(\alpha) and c_{2}(\alpha) are invariant under shifts of both the regular and the backward continued fraction expansion of \alpha . Quadratic irrationals play a special role in the ergodic theory of circle rotations as well, as shown by Beck [ 6 ] and Avila, Duryev, Dolgopyat and Sarig [ 4 ] . In particular, the same constant c_{1}(\alpha) appears in probabilistic limit theorems for certain Birkhoff sums known as temporal limit theorems, in lattice point counting in polygons [ 6 ] , and in the L^{2} discrepancy of 2-dimensional lattices [ 19 ] . A constant very similar to c_{1}(\alpha) appears in [ 4 ] , where a completely different method is given for finding its explicit value using dynamical systems and Markov chains. The main goal of the present paper is to extend the framework of [ 4 , 6 ] from ergodic theory to random walks. In particular, Theorems 1 and 2 are the first results in the literature connecting random walks and the arithmetic of real quadratic fields, and also the first distributional results where special zeta values at s=-3 play a role. The choice of the Wasserstein metric in our results is for convenience, we expect other statistics of the empirical measure \mu_{N} to be related to the arithmetic of \mathbb{Q}(\alpha) as well. The rest of the paper is organized as follows. Section 2 is an overview of the Wasserstein metric and random walks on \mathbb{T} for further context. Section 3 deals with the probability theory part of the proof, using exponential sums and Diophantine approximation. We will show that the estimates \mathbb{E}\left(W_{\mathbb{T},2}^{2}(\mu_{N},\lambda)\right)\asymp\frac{\log N}{N}\quad\textrm{and}\quad\mathrm{Var}\left(W_{\mathbb{T},2}^{2}(\mu_{N},\lambda)\right)\asymp\frac{\log N}{N^{2}} actually hold for any badly approximable \alpha , but in general they cannot be refined to precise asymptotic relations. In particular, in Theorem 1 the

Read the complete text