Functional limit theorems for elephant random walks on general periodic structures

Shuhei Shibata · 2025

Read Functional limit theorems for elephant random walks on general periodic structures by Shuhei Shibata (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-2511-10347.txt.

Introduction

Shuhei Shibata, writing alone, studies elephant random walks on general periodic grids, building on limit theorems earlier known on the integer lattice. An elephant random walk remembers its past: at each step it recalls an earlier step and copies or reverses it with fixed chances, so the walk's future depends on its whole history. Periodic structures generalize the regular grid to repeating patterned spaces.

Shibata proves functional limit theorems describing the walk's large-scale behavior on such grids. New quantities appear that have no counterpart in the classical lattice setting, exposing how the underlying geometry shapes the end behavior. The paper is a probability theory contribution, broadening a memory-driven walk model from flat grids to structured environments.

Catalog record

CREATORShuhei Shibata
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE40 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-10347

Browse the full library · About the Tower of Babel

Full text

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

Functional limit theorems for elephant random walks on general periodic structures Shuhei Shibata Abstract This paper investigates functional limit theorems for the Elephant Random Walk (ERW) on general periodic structures, extending the Bertenghi’s results on ℤ d \mathbb{Z}^{d} . Our results reveal new structure-dependent quantities that do not appear in the classical setting ℤ d \mathbb{Z}^{d} , highlighting how the underlying structure affects the asymptotic behavior of the walk. 1 Introduction The asymptotic behavior of random walks with long-range memory has been extensively studied in recent years. In particular, Elephant Random Walk (ERW) has attracted considerable attention. The ERW model was introduced by Schütz and Trimper [ 23 ] in 2004 to study memory effects in a one-dimensional discrete-time random walk with a complete memory of its past, and it exhibits a phase transition from diffusive to superdiffusive behavior. One natural generalization of the one-dimensional ERW is to consider higher-dimensional settings, particularly random walks on the standard lattice \mathbb{Z}^{d} . A number of works have addressed this extension (cf. [ 3 , 4 , 7 , 16 , 14 , 13 , 6 , 11 , 21 ] ). Bercu and Laulin [ 3 ] investigated the asymptotic behavior of the Multi-dimensional Elephant Random Walk (MERW) via a martingale approach. Later, Bertenghi [ 7 ] established functional limit theorems for MERW by employing a Pólya-type urn analysis, following the approach initiated by Baur and Bertoin [ 2 ] . This connection between the (M)ERW and Pólya urns enables the derivation of functional limit theorems for the (M)ERW within the general framework developed by Janson [ 17 ] . In addition to studying the asymptotic behavior of a single walker, another natural extension is to investigate the interaction of multiple ERWs, particularly the collision problem—that is, whether two independent elephant random walks on the same lattice meet infinitely often or only finitely many times. Roy, Takei and Tanemura [ 22 ] studied the case where both ERWs on \mathbb{Z} have the same memory parameter. Later, Shibata and Shirai [ 24 ] extended their results to the case of different memory parameters and also obtained asymptotic results for the distance between them. Although this paper does not address the collision problem, we mention it here to emphasize the diversity of recent developments in the study of ERW. Most of the works mentioned above focus on the ERWs on the standard lattice \mathbb{Z}^{d} . In this paper, we extend this framework to more general structures, including \mathbb{Z}^{d} , as well as the triangular lattice , the hexagonal lattice and the brick wall , see Fig. 1 . Inspired by the work of Bertenghi [ 7 ] , who established functional limit theorems for the ERW on \mathbb{Z}^{d} , we derive the corresponding asymptotic results for the ERW on more general state spaces such as these lattices, using a Pólya-type urn approach. Our results reveal new underlying structure-dependent quantities which are not observed in the classical \mathbb{Z}^{d} case, demonstrating that the underlying structure significantly influences the asymptotic behavior of the walk. Figure 1 : Figure 1 . Left: Triangular lattice. All vertices (black dots) are structurally equivalent. Center: Hexagonal lattice. The vertices are partitioned into two structurally distinct classes, represented by red and blue dots. Right: Brick wall. As in the hexagonal lattice, the vertices are divided into two structurally distinct classes (red and blue dots). The remainder of the paper is organized as follows. In Section 2 and Section 3 , we provide the mathematical definition of the state spaces we consider and the ERW on them. In Section 4 , we introduce discrete-time urn models that serves as a framework for describing the position of the ERW. In Section 5 , we present the main asymptotic results for ERW based on the urn analysis. Finally, in Section 6 , for the reader’s convenience, we summarize the key quantities derived in this paper for several typical examples. 2 Settings To set the stage for our results, we first introduce the mathematical framework considered in this paper. Given two finite sets of distinct vectors \mathcal{U}:=\{u_{1},\dots,u_{m}\}\subset\mathbb{R}^{d} and \mathcal{W}:=\{w_{1},\dots,w_{m^{\prime}}\}\subset\mathbb{R}^{d} for fixed integers m,m^{\prime}\geq 2 . We call the elements of \mathcal{U} and \mathcal{W} step vectors since, in Section 3 , we will define a random walk whose steps are drawn from these sets. We will consider two settings depending on whether the step sets coincide, namely the case \mathcal{U}=\mathcal{W} and \mathcal{U}\neq\mathcal{W} . • Case \mathcal{U}=\mathcal{W} : In this case, we consider the semigroup generated by them, that is, \Gamma=\{\sum_{i=1}^{m}k_{i}u_{i};\ k_{i}\in\mathbb{N}\cup\{0\}\}\subset\mathbb{R}^{d}. (2.1) Here, we assume that this semigroup \Gamma is a lattice in \mathbb{R}^{d} , i.e., it is a discrete additive subgroup that spans \mathbb{R}^{d} as a vector space over \mathbb{R} . For this case, \Gamma is analyzed as a single-colored vertex set, as in the triangular lattice on the left in Fig. 1 , owing to the structural indistinguishability of all vertices. • Case \mathcal{U}\neq\mathcal{W} : In this case, we first define the semigroup \Gamma_{0}=\{\sum_{i=1}^{m}k_{i}u_{i}+\sum_{j=1}^{m^{\prime}}l_{j}w_{j};\ \sum_{i=1}^{m}k_{i}=\sum_{j=1}^{m^{\prime}}l_{j},\ k_{i},l_{j}\in\mathbb{N}\cup\{0\}\}. Here, we assume that this semigroup is a lattice in \mathbb{R}^{d} . We then consider the following state space \Gamma that can be expressed as the disjoint union of \Gamma_{0} and the translated set \Gamma_{0}+\mathcal{U} : \Gamma=\Gamma_{0}\sqcup(\Gamma_{0}+\mathcal{U})\subset\mathbb{R}^{d}. (2.2) We define the vertex classes by Z_{\mathcal{U}}:=\Gamma_{0} and Z_{\mathcal{W}}:=\Gamma_{0}+\mathcal{U} . Under these settings, we enforce the Alternating Rule : – For any z\in Z_{\mathcal{U}} and i=1,\dots,m , z+u_{i}\in Z_{\mathcal{W}} . – For any z\in Z_{\mathcal{W}} and j=1,\dots,m^{\prime} , z+w_{j}\in Z_{\mathcal{U}} . This rule ensures that the walk defined in Section 3 always remains within \Gamma . \Gamma is therefore a bipartite graph with vertices classified according to the decomposition in ( 2.2 ), and is analyzed as a two-colored vertex set, as in the hexagonal lattice and brick wall shown in the center and right panels of Fig. 1 . Remark 2.1 . When \mathcal{U}=\mathcal{W}=\{\pm e_{1},\pm e_{2},\dots,\pm e_{d}\} where e_{i} denotes the i -th standard basis vector of \mathbb{R}^{d} , \Gamma coincides with the standard integer lattice \mathbb{Z}^{d} . Note that there are also many other ways to choose the vectors that generate \mathbb{Z}^{d} . Various well-known structures, such as the triangular lattice, the hexagonal lattice and the brick wall, see Fig. 1 , can be realized as \Gamma , by choosing suitable families of step vectors \mathcal{U} and \mathcal{W} . For clarity, in Section 6 , we list the representative choices of \mathcal{U} and \mathcal{W} for several typical examples. Remark 2.2 . We introduce the \Gamma given by ( 2.2 ) to represent a setting in which each vertex is assigned a specific type (or color), as in the hexagonal lattice. Although the following example does not satisfy the decomposition of ( 2.2 ), one could construct an ERW model and a corresponding setting that covers it: \mathcal{U}=\{\pm e_{1},\pm e_{2},(1,1)^{\top}\},\quad\mathcal{W}=\{\pm e_{1},\pm e_{2}\}. We also note that the trivial case m=1 can be analyzed within the same framework, but it is excluded from our study. 3 Definition of the ERW on \Gamma Next, we introduce the mathematical definitions of the ERW \{S_{n}\}_{n=0}^{\infty} on \Gamma . For a given dimension d\geq 1 and the sets \mathcal{U} and \mathcal{W} introduced in Section 2 , let \{\sigma_{i}\}_{i=1}^{\infty} and \{\tau_{j}\}_{j=1}^{\infty} be random sequences taking values

Read the complete text