Asymptotics of the number of geodesics in the discrete Heisenberg group (Q2307958)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Asymptotics of the number of geodesics in the discrete Heisenberg group |
scientific article |
Statements
Asymptotics of the number of geodesics in the discrete Heisenberg group (English)
0 references
25 March 2020
0 references
The absolute of a group with a fixed system of generators is defined as the set of ergodic Markov measures for which the system of cotransition probabilities is the same as for the simple (right) random walk generated by the uniform distribution on the generators. The absolute is a new boundary of a group, generated by random walks on the group. The absolute is divided into two parts, the Laplacian and the degenerate. A description of the absolute has been obtained for the free group and some other hyperbolic groups, for abelian groups and semigroups. The Laplacian part of the absolute has been described for arbitrary nilpotent groups (see, for example, [\textit{A. M. Vershik} and \textit{A. V. Malyutin}, Funct. Anal. Appl. 52, No. 3, 163--177 (2018; Zbl 1483.20125); translation from Funkts. Anal. Prilozh. 52, No. 3, 3--21 (2018); \textit{A. M. Vershik} and \textit{A. V. Malyutin}, Eur. J. Math. 4, No. 4, 1476--1490 (2018; Zbl 1403.28014)] and references therein). The aim of the authors is to give a complete description of the absolute for a noncommutative nilpotent group, the discrete Heisenberg group. As they claim, it is obtained in a paper under preparation. But it is needed ``something'' for this description. In the present paper, this ``something'' is given. In the Cayley graph of the discrete Heisenberg group with the standard set of generators, the number of geodesic paths between two given points is equal to the number of restricted partitions with some condition on the size and the number of parts, or, which is the same, the number of Young diagrams of given area fitting in a rectangle of given size. So one of the main steps in the problem of describing the absolute of the Heisenberg group is reduced to the problem of counting the number of geodesic paths, this problem restated in terms of Young diagrams, and it is solved using the known properties of \(q\)-binomial coefficients. Before stating the main result proved here, let us quote some notation and terminology. Let \(N_{2}=\langle x,y \mid [[x,y],x]=[[x,y],y]=1\rangle \) be the standard presentation of the discrete Heisenberg group. It is well known that for every element \(g\in N_{2}\) there exists a unique triple of integers \((m, n, k)\) such that \(g=[x, y]^{k}y^{n}x^{m}\). This allows one to obtain a convenient model for the Cayley graph of the Heisenberg group by placing an element \(g=[x, y]^{k}y^{n}x^{m}\) at the point \((m, n, k)\) of the three-dimensional Euclidean space. Let \(r(m,n,k)=\min(m,n,mn-k)\), for \(m,n,k \in\mathbb{N}\). Theorem. In the discrete Heisenberg group \(N_{2}=\langle x,y \mid [[x,y],x]=[[x,y],y]=1\rangle\), denote by \(\dim(g)\) the number of shortest (geodesic) paths in the Cayley graph of the pair \((N_{2},\{x,y\})\) leading to an element \(g\) from the group identity. Then \(\frac{\dim([x, y]^{k+1}y^{n}x^{m})}{\dim([x,y]^{k}y^{n}x^{m})} \longrightarrow 1\), when \(r(m,n,k)\longrightarrow\infty\). The authors use the standard geometric interpretation of elements of the discrete Heisenberg group as paths on the plane and invoke, from [\textit{M. Shapiro}, Math. Ann. 285, No. 4, 601--624 (1989; Zbl 0693.20037)], the: Proposition. In the integer lattice on the plane, two paths starting at the origin represent the same element of the Heisenberg group if and only if their endpoints coincide and the oriented area of the closed contour formed by these paths vanishes. Therefore, the theorem it is restated to: Lemma 1. Denote by \(P_{m,n}(Q)\) the number of Young diagrams of area \(Q\) fitting in a rectangle of width \(m\) and height \(n\). Then \(\frac{ P_{m,n}(Q+1)}{P_{m,n}(Q)}\longrightarrow 1\), as \(q(m,n,k)=\min(m,n,Q,mn-Q)\) for \(m,n,Q \in\mathbb{N}\) leads to \(\infty\). The proof of this lemma is reduced to the proof of: Lemma 2. Let \(t\in \mathbb{N}\) and \(\varepsilon >0\). Then for arbitrary positive integers \(Q\), \(m\), \(n\) such that \(Q \leq mn/2\) and \( \min\{m,n\sqrt{2Q}\}\geq \max\{36t^{2},1/\epsilon \}\) we have \(\frac{P_{m,n}(Q,t)}{P_{m,n}(Q)} <\varepsilon\).
0 references
discrete Heisenberg group
0 references
Gaussian \(q\)-binomial coefficients
0 references
Young diagrams
0 references
0 references
0 references