反向Rauzy-Veech算法和水平鞍连接

Przemysław Berk
{"title":"反向Rauzy-Veech算法和水平鞍连接","authors":"Przemysław Berk","doi":"10.2422/2036-2145.202110_010","DOIUrl":null,"url":null,"abstract":"We study the combinatorial and dynamical properties of translations surfaces with horizontal saddle connections from the point of view of backward Rauzy-Veech induction. Namely, we prove that although the horizontal saddle connections do not rule out existence of the infinite orbit under backward Rauzy-Veech algorithm, they disallow the ∞completeness of such orbit. Furthermore, we prove that if an orbit under backward Rauzy-Veech algorithm is infinite, then the minimality of the horizontal translation flow is equivalent to the eventual appearance of all horizontal saddle connections as sides of the polygonal represenation of a surface. The main goal of this note is to study the relations between horizontal saddle connections and the combinatorics of the inverse RauzyVeech algorithm for translation surfaces as well as dynamics of the horizontal translation flows. In [2] (Proposition 4.3) Marmi, Ulcigrai and Yoccoz prove that if a translation surface does not have horizontal saddle connections, then its backward Rauzy-Veech induction orbit is indefinitely well-defined and ∞-complete, that is every symbol is a backward winner infinitely many times. In the same article the authors pose a question, whether these two conditions are equivalent. We answer affirmatively to this question in Theorem 11. The proof utilizes only combinatorics and geometry of translation surfaces. However, before proving Theorem 11, we prove Proposition 7 which states that, typically, possessing horizontal saddle connections does not rule out that the backward orbit with respect to the inverse RauzyVeech algorithm is well defined. Moreover, in Theorem 12 we prove that appearance of horizontal connections as sides of polygonal representations of translations surfaces is closely tied to the minimality of the horizontal translation flow. More precisely, we show that the horizontal translation flow is minimal if and only if all (if any) horizontal saddle connections appear as sides of a polygonal representation of a surface after applying a finite number of backward Rauzy-Veech induction steps. MSC CLASSIFICATION: 37E05, 37E35 1 ar X iv :2 10 9. 13 69 1v 2 [ m at h. D S] 2 3 Fe b 20 22 P. BERK HORIZONTAL SADDLE CONNECTIONS Acknowledgments: The author would like to thank Corinna Ulcigrai for pointing out the problem and her continuous support and Frank Trujillo for many useful remarks. The research that lead to this result was supported by Swiss National Science Foundation Grant 200021 188617/1 and Narodowe Centrum Nauki Grant OPUS 142017/27/B/ST1/00078. 1. Interval exchange transformations and translations surfaces We recall first basic notions and properties related to IETs and translation surfaces. Let A be an alphabet of #A ≥ 2 elements. For more information and basic properties, including the ergodic properties of interval exchange transformations, translation surfaces and Rauzy-Veech algorithm we refer the reader e.g. to [4] and [5]. Let SA 0 :={π = (π0, π1) : A → {1, . . . ,#A} × {1, . . . ,#A}; π1 ◦ π−1 0 {1, . . . , k} = {1, . . . , k} ⇒ k = #A} be the set of irreducible permutations, where π0 and π1 are bijections. Let us also denote by R>0 the set of all d-dimensional positive real vectors and for every λ ∈ R>0 let |λ| := ∑ α∈A λα. An interval exchange transformation on [0, |λ|) (IET) T = (π, λ) ∈ SA 0 × R>0 is a bijective piecewise translation, where the intervals Iα :=  ∑ β∈A;π0(β)<π0(α) λβ, ∑ β∈A;π0(β)≤π0(α) λβ  for α ∈ A are rearranged inside [0, |λ|) with respect to the permutation π. More precisely, for every α ∈ A, we have T (x) = x+ δα if x ∈ Iα, where δα = ∑ β∈A; π1(β)<π1(α) λα − ∑ β∈A; π0(β)<π0(α) λα. Note that T preserves Lebesgue measure. We denote by Ω = [ωαβ]α,β the associated translation matrix, with coefficients given by ωαβ :=  +1 if π0(α) < π0(β) and π1(α) > π1(β); −1 if π0(α) > π0(β) and π1(α) < π1(β); 0 otherwise. Then, if δ := [δα]α∈A, we get δ = Ωπ · λ. 2 P. BERK HORIZONTAL SADDLE CONNECTIONS On the space SA 0 × R>0 we consider an operator R called RauzyVeech induction, defined as R(π, λ) = (π, λ), where (π, λ) is the first return map of (π, λ) to the interval [0, |λ| −min{λπ−1 0 (d), λπ−1 1 (d)}). If λπ−1 0 (d) > λπ −1 1 (d) we say that R is of ”top” type and we say that it is of ”bottom” type if λπ−1 0 (d) < λπ −1 1 (d) . We denote the symbol corresponding to the longer interval as w (the winner) and to the shorter one as l (the loser). The map R(π, λ) is properly defined as an interval exchange transformation of d intervals if and only if λπ−1 0 (d) 6= λπ−1 1 (d). Keane [1] gave an equivalent condition on (π, λ), for the iterations of Rauzy-Veech induction to be defined indefinitely. More precisely, we say that IET T satisfies Keane’s condition if for every two discontinuities a and b of T equality T (a) = b for some n ∈ N implies n = 1, a = T−1(0) and b = 0. In particular, if the vector λ is rationally independent, that is for every choice of cα ∈ Z, α ∈ A we have ∑ α∈A cαλα = 0 ⇒ cα = 0 for every α ∈ A, then (π, λ) satisfies Keane’s condition. When it is well defined, we denote R(π, λ) = (π, λ) for every n ∈ N. We say that the orbit of (π, λ) via Rauzy-Veech induction is ∞-complete if every symbol in A appears infinitely many times in the sequence of winners {w}. Note that λ = A(π, λ)λ, where a matrix A(π, λ) is defined in the following way Aαβ =  1 if α = β; −1 if α = w and β = l; 0 otherwise. Inductively, for every n ∈ N we define A(π, λ) = A1(πn−1, λn−1)An−1(π, λ). Then λ = A(π, λ)λ. We will refer to A(π, λ) as Rauzy-Veech matrices. Note that for every n ∈ N, the matrix (A(π, λ))−1 is non-negative. For every π ∈ SA 0 let ΘA = ΘA(π) = { τ ∈ RA; ∑ α∈A;π0(α)≤k τα > 0 and ∑ α∈A;π1(α)≤k τα < 0 for every k ∈ {1, . . . , d− 1} } . Then every (π, λ, τ) ∈ SA 0 × ΛA × ΘA1 may be see as a translation surface as follows. More precisely, first we consider two broken line 1Note that this space is not really a product space since ΘA depends on π and thus SA 0 ×ΛA×ΘA = ⋃ π∈SA 0 {π}×ΛA×ΘA(π). However, we shall use this notation for simplicity. 3 P. BERK HORIZONTAL SADDLE CONNECTIONS Figure 1. A translation surface and one step of backward Rauzy-Veech induction. The parallel segments are identified via translation. The winning segment is the one which is first crossed by a rightward separatrix starting from (0, 0).","PeriodicalId":8132,"journal":{"name":"ANNALI SCUOLA NORMALE SUPERIORE - CLASSE DI SCIENZE","volume":"37 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2021-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Backward Rauzy-Veech algorithm and horizontal saddle connections\",\"authors\":\"Przemysław Berk\",\"doi\":\"10.2422/2036-2145.202110_010\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We study the combinatorial and dynamical properties of translations surfaces with horizontal saddle connections from the point of view of backward Rauzy-Veech induction. Namely, we prove that although the horizontal saddle connections do not rule out existence of the infinite orbit under backward Rauzy-Veech algorithm, they disallow the ∞completeness of such orbit. Furthermore, we prove that if an orbit under backward Rauzy-Veech algorithm is infinite, then the minimality of the horizontal translation flow is equivalent to the eventual appearance of all horizontal saddle connections as sides of the polygonal represenation of a surface. The main goal of this note is to study the relations between horizontal saddle connections and the combinatorics of the inverse RauzyVeech algorithm for translation surfaces as well as dynamics of the horizontal translation flows. In [2] (Proposition 4.3) Marmi, Ulcigrai and Yoccoz prove that if a translation surface does not have horizontal saddle connections, then its backward Rauzy-Veech induction orbit is indefinitely well-defined and ∞-complete, that is every symbol is a backward winner infinitely many times. In the same article the authors pose a question, whether these two conditions are equivalent. We answer affirmatively to this question in Theorem 11. The proof utilizes only combinatorics and geometry of translation surfaces. However, before proving Theorem 11, we prove Proposition 7 which states that, typically, possessing horizontal saddle connections does not rule out that the backward orbit with respect to the inverse RauzyVeech algorithm is well defined. Moreover, in Theorem 12 we prove that appearance of horizontal connections as sides of polygonal representations of translations surfaces is closely tied to the minimality of the horizontal translation flow. More precisely, we show that the horizontal translation flow is minimal if and only if all (if any) horizontal saddle connections appear as sides of a polygonal representation of a surface after applying a finite number of backward Rauzy-Veech induction steps. MSC CLASSIFICATION: 37E05, 37E35 1 ar X iv :2 10 9. 13 69 1v 2 [ m at h. D S] 2 3 Fe b 20 22 P. BERK HORIZONTAL SADDLE CONNECTIONS Acknowledgments: The author would like to thank Corinna Ulcigrai for pointing out the problem and her continuous support and Frank Trujillo for many useful remarks. The research that lead to this result was supported by Swiss National Science Foundation Grant 200021 188617/1 and Narodowe Centrum Nauki Grant OPUS 142017/27/B/ST1/00078. 1. Interval exchange transformations and translations surfaces We recall first basic notions and properties related to IETs and translation surfaces. Let A be an alphabet of #A ≥ 2 elements. For more information and basic properties, including the ergodic properties of interval exchange transformations, translation surfaces and Rauzy-Veech algorithm we refer the reader e.g. to [4] and [5]. Let SA 0 :={π = (π0, π1) : A → {1, . . . ,#A} × {1, . . . ,#A}; π1 ◦ π−1 0 {1, . . . , k} = {1, . . . , k} ⇒ k = #A} be the set of irreducible permutations, where π0 and π1 are bijections. Let us also denote by R>0 the set of all d-dimensional positive real vectors and for every λ ∈ R>0 let |λ| := ∑ α∈A λα. An interval exchange transformation on [0, |λ|) (IET) T = (π, λ) ∈ SA 0 × R>0 is a bijective piecewise translation, where the intervals Iα :=  ∑ β∈A;π0(β)<π0(α) λβ, ∑ β∈A;π0(β)≤π0(α) λβ  for α ∈ A are rearranged inside [0, |λ|) with respect to the permutation π. More precisely, for every α ∈ A, we have T (x) = x+ δα if x ∈ Iα, where δα = ∑ β∈A; π1(β)<π1(α) λα − ∑ β∈A; π0(β)<π0(α) λα. Note that T preserves Lebesgue measure. We denote by Ω = [ωαβ]α,β the associated translation matrix, with coefficients given by ωαβ :=  +1 if π0(α) < π0(β) and π1(α) > π1(β); −1 if π0(α) > π0(β) and π1(α) < π1(β); 0 otherwise. Then, if δ := [δα]α∈A, we get δ = Ωπ · λ. 2 P. BERK HORIZONTAL SADDLE CONNECTIONS On the space SA 0 × R>0 we consider an operator R called RauzyVeech induction, defined as R(π, λ) = (π, λ), where (π, λ) is the first return map of (π, λ) to the interval [0, |λ| −min{λπ−1 0 (d), λπ−1 1 (d)}). If λπ−1 0 (d) > λπ −1 1 (d) we say that R is of ”top” type and we say that it is of ”bottom” type if λπ−1 0 (d) < λπ −1 1 (d) . We denote the symbol corresponding to the longer interval as w (the winner) and to the shorter one as l (the loser). The map R(π, λ) is properly defined as an interval exchange transformation of d intervals if and only if λπ−1 0 (d) 6= λπ−1 1 (d). Keane [1] gave an equivalent condition on (π, λ), for the iterations of Rauzy-Veech induction to be defined indefinitely. More precisely, we say that IET T satisfies Keane’s condition if for every two discontinuities a and b of T equality T (a) = b for some n ∈ N implies n = 1, a = T−1(0) and b = 0. In particular, if the vector λ is rationally independent, that is for every choice of cα ∈ Z, α ∈ A we have ∑ α∈A cαλα = 0 ⇒ cα = 0 for every α ∈ A, then (π, λ) satisfies Keane’s condition. When it is well defined, we denote R(π, λ) = (π, λ) for every n ∈ N. We say that the orbit of (π, λ) via Rauzy-Veech induction is ∞-complete if every symbol in A appears infinitely many times in the sequence of winners {w}. Note that λ = A(π, λ)λ, where a matrix A(π, λ) is defined in the following way Aαβ =  1 if α = β; −1 if α = w and β = l; 0 otherwise. Inductively, for every n ∈ N we define A(π, λ) = A1(πn−1, λn−1)An−1(π, λ). Then λ = A(π, λ)λ. We will refer to A(π, λ) as Rauzy-Veech matrices. Note that for every n ∈ N, the matrix (A(π, λ))−1 is non-negative. For every π ∈ SA 0 let ΘA = ΘA(π) = { τ ∈ RA; ∑ α∈A;π0(α)≤k τα > 0 and ∑ α∈A;π1(α)≤k τα < 0 for every k ∈ {1, . . . , d− 1} } . Then every (π, λ, τ) ∈ SA 0 × ΛA × ΘA1 may be see as a translation surface as follows. More precisely, first we consider two broken line 1Note that this space is not really a product space since ΘA depends on π and thus SA 0 ×ΛA×ΘA = ⋃ π∈SA 0 {π}×ΛA×ΘA(π). However, we shall use this notation for simplicity. 3 P. BERK HORIZONTAL SADDLE CONNECTIONS Figure 1. A translation surface and one step of backward Rauzy-Veech induction. The parallel segments are identified via translation. The winning segment is the one which is first crossed by a rightward separatrix starting from (0, 0).\",\"PeriodicalId\":8132,\"journal\":{\"name\":\"ANNALI SCUOLA NORMALE SUPERIORE - CLASSE DI SCIENZE\",\"volume\":\"37 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-09-28\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"ANNALI SCUOLA NORMALE SUPERIORE - CLASSE DI SCIENZE\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.2422/2036-2145.202110_010\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"ANNALI SCUOLA NORMALE SUPERIORE - CLASSE DI SCIENZE","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.2422/2036-2145.202110_010","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

从后向Rauzy-Veech归纳法的观点出发,研究了具有水平鞍连接的平移曲面的组合性质和动力学性质。即,我们证明了水平鞍连接虽然不排除反向Rauzy-Veech算法下无限轨道的存在,但不允许无限轨道的∞完备性。进一步,我们证明了如果在反向Rauzy-Veech算法下的轨道是无限的,那么水平平移流的极小性等价于所有水平鞍连接作为曲面多边形表示的边的最终外观。本文的主要目的是研究水平鞍连接与平移曲面逆RauzyVeech算法组合之间的关系,以及水平平移流的动力学。在[2](命题4.3)中,Marmi, Ulcigrai和Yoccoz证明,如果平移曲面不存在水平鞍连接,则其后向Rauzy-Veech诱导轨道是无限期定义良好且∞完全的,即每个符号都是无限次的后向赢家。在同一篇文章中,作者提出了一个问题,这两个条件是否等价。在定理11中,我们肯定地回答了这个问题。证明只利用了组合学和平动曲面的几何学。然而,在证明定理11之前,我们证明命题7,该命题指出,通常情况下,拥有水平鞍连接并不排除相对于逆RauzyVeech算法的反向轨道是定义良好的。此外,在定理12中,我们证明了水平连接作为平移曲面多边形表示的边的外观与水平平移流的极小性密切相关。更准确地说,我们证明了当且仅当所有(如果有的话)水平鞍连接在应用有限数量的反向Rauzy-Veech归纳步骤后出现为曲面多边形表示的边时,水平平移流是最小的。MSC分类:37E05, 37E35 1 ar X iv:2 10 9。13 69 1v 2 [m at h. D S] 2 3 Fe b 22 P. BERK水平鞍座连接致谢:作者感谢Corinna Ulcigrai指出了问题并一直给予支持,感谢Frank Trujillo提供了许多有用的意见。导致这一结果的研究得到了瑞士国家科学基金会资助200021 188617/1和Narodowe Centrum Nauki资助OPUS 142017/27/B/ST1/00078的支持。1. 区间交换变换和平移曲面我们首先回顾与区间交换变换和平移曲面相关的基本概念和性质。设A为A≥2个元素的字母表。要了解更多的信息和基本性质,包括区间交换变换、平移曲面和Rauzy-Veech算法的遍历性质,请参见[4]和[5]。设SA 0:={π = (π0, π1): A→{1,…,# a} ×{1,…, #};π1◦π−10{1,…, k} ={1,…, k}⇒k = #A}是不可约置换的集合,其中π0和π1是双射。我们也用R>0表示所有d维正实向量的集合,并且对于每个λ∈R>0,令|λ|:=∑α∈A λα。在[0,|λ|) (IET) T = (π, λ)∈SA 0 × R>0上的区间交换变换是一个双射分段平移,其中区间Iα:=∑β∈a;π0(α) > π0(β)且π1(α) < π1(β)−1;0。然后,如果δ:= [δα]α∈A,则得到δ = Ωπ·λ。在空间sa0 × R>0上,我们考虑一个称为RauzyVeech归纳的算子R,定义为R(π, λ) = (π, λ),其中(π, λ)是(π, λ)到区间[0,|λ|−min{λπ−10 (d), λπ−11 (d)}的第一个返回映射。若λπ−10 0 (d) > λπ−11 1 (d),我们称R为“上”型,若λπ−10 0 (d) < λπ−11 1 (d),我们称R为“下”型。我们将较长间隔对应的符号表示为w(赢家),将较短间隔对应的符号表示为l(输家)。当且仅当λπ−10 0 (d) 6= λπ−11 1 (d)时,映射R(π, λ)被恰当地定义为d区间的区间交换变换。Keane[1]给出了(π, λ)的等价条件,使得Rauzy-Veech归纳迭代是无限定义的。更准确地说,我们说IET T满足Keane的条件,如果对于T等于T (a) = b的每两个不连续点a和b,对于某些n∈n意味着n = 1, a = T - 1(0)和b = 0。特别地,如果向量λ是理性独立的,即对于cα∈Z, α∈A的每一个选择,我们有∑α∈A cαλα = 0⇒cα = 0,则(π, λ)满足Keane的条件。当它被很好地定义时,对于每个n∈n,我们表示R(π, λ) = (π, λ)。我们说,如果A中的每个符号在胜利者序列{w}中出现无限多次,通过Rauzy-Veech归纳(π, λ)的轨道是∞-完全的。 注意,λ = A(π, λ)λ,其中矩阵A(π, λ)以以下方式定义:Aαβ =1如果α = β;−1 α = w, β = l;0。归纳地说,对于每个n∈n,我们定义A(π, λ) = A1(πn−1,λn−1)An−1(π, λ)。则λ = A(π, λ)λ。我们将把A(π, λ)称为Rauzy-Veech矩阵。注意,对于每个n∈n,矩阵(A(π, λ))−1是非负的。对于每一个π∈SA 0,令ΘA = ΘA(π) = {τ∈RA;∑α∈;π0(α)≤kτα> 0,∑α∈;π1(α)≤kτα< 0 k每∈{1,…, d−1}}。则每个(π, λ, τ)∈SA 0 × ΛA × ΘA1可以看作如下的平移曲面。更准确地说,首先我们考虑两条折线1注意,这个空间并不是一个真正的乘积空间,因为ΘA依赖于π,因此SA 0 ×ΛA×ΘA = <s:2> π∈SA 0 {π}×ΛA×ΘA(π)。但是,为了简单起见,我们将使用这种符号。P. BERK水平鞍座连接图1。一个平移曲面和一步逆向Rauzy-Veech归纳。通过平移确定平行段。获胜的段是从(0,0)开始的向右分离矩阵首先穿过的段。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
Backward Rauzy-Veech algorithm and horizontal saddle connections
We study the combinatorial and dynamical properties of translations surfaces with horizontal saddle connections from the point of view of backward Rauzy-Veech induction. Namely, we prove that although the horizontal saddle connections do not rule out existence of the infinite orbit under backward Rauzy-Veech algorithm, they disallow the ∞completeness of such orbit. Furthermore, we prove that if an orbit under backward Rauzy-Veech algorithm is infinite, then the minimality of the horizontal translation flow is equivalent to the eventual appearance of all horizontal saddle connections as sides of the polygonal represenation of a surface. The main goal of this note is to study the relations between horizontal saddle connections and the combinatorics of the inverse RauzyVeech algorithm for translation surfaces as well as dynamics of the horizontal translation flows. In [2] (Proposition 4.3) Marmi, Ulcigrai and Yoccoz prove that if a translation surface does not have horizontal saddle connections, then its backward Rauzy-Veech induction orbit is indefinitely well-defined and ∞-complete, that is every symbol is a backward winner infinitely many times. In the same article the authors pose a question, whether these two conditions are equivalent. We answer affirmatively to this question in Theorem 11. The proof utilizes only combinatorics and geometry of translation surfaces. However, before proving Theorem 11, we prove Proposition 7 which states that, typically, possessing horizontal saddle connections does not rule out that the backward orbit with respect to the inverse RauzyVeech algorithm is well defined. Moreover, in Theorem 12 we prove that appearance of horizontal connections as sides of polygonal representations of translations surfaces is closely tied to the minimality of the horizontal translation flow. More precisely, we show that the horizontal translation flow is minimal if and only if all (if any) horizontal saddle connections appear as sides of a polygonal representation of a surface after applying a finite number of backward Rauzy-Veech induction steps. MSC CLASSIFICATION: 37E05, 37E35 1 ar X iv :2 10 9. 13 69 1v 2 [ m at h. D S] 2 3 Fe b 20 22 P. BERK HORIZONTAL SADDLE CONNECTIONS Acknowledgments: The author would like to thank Corinna Ulcigrai for pointing out the problem and her continuous support and Frank Trujillo for many useful remarks. The research that lead to this result was supported by Swiss National Science Foundation Grant 200021 188617/1 and Narodowe Centrum Nauki Grant OPUS 142017/27/B/ST1/00078. 1. Interval exchange transformations and translations surfaces We recall first basic notions and properties related to IETs and translation surfaces. Let A be an alphabet of #A ≥ 2 elements. For more information and basic properties, including the ergodic properties of interval exchange transformations, translation surfaces and Rauzy-Veech algorithm we refer the reader e.g. to [4] and [5]. Let SA 0 :={π = (π0, π1) : A → {1, . . . ,#A} × {1, . . . ,#A}; π1 ◦ π−1 0 {1, . . . , k} = {1, . . . , k} ⇒ k = #A} be the set of irreducible permutations, where π0 and π1 are bijections. Let us also denote by R>0 the set of all d-dimensional positive real vectors and for every λ ∈ R>0 let |λ| := ∑ α∈A λα. An interval exchange transformation on [0, |λ|) (IET) T = (π, λ) ∈ SA 0 × R>0 is a bijective piecewise translation, where the intervals Iα :=  ∑ β∈A;π0(β)<π0(α) λβ, ∑ β∈A;π0(β)≤π0(α) λβ  for α ∈ A are rearranged inside [0, |λ|) with respect to the permutation π. More precisely, for every α ∈ A, we have T (x) = x+ δα if x ∈ Iα, where δα = ∑ β∈A; π1(β)<π1(α) λα − ∑ β∈A; π0(β)<π0(α) λα. Note that T preserves Lebesgue measure. We denote by Ω = [ωαβ]α,β the associated translation matrix, with coefficients given by ωαβ :=  +1 if π0(α) < π0(β) and π1(α) > π1(β); −1 if π0(α) > π0(β) and π1(α) < π1(β); 0 otherwise. Then, if δ := [δα]α∈A, we get δ = Ωπ · λ. 2 P. BERK HORIZONTAL SADDLE CONNECTIONS On the space SA 0 × R>0 we consider an operator R called RauzyVeech induction, defined as R(π, λ) = (π, λ), where (π, λ) is the first return map of (π, λ) to the interval [0, |λ| −min{λπ−1 0 (d), λπ−1 1 (d)}). If λπ−1 0 (d) > λπ −1 1 (d) we say that R is of ”top” type and we say that it is of ”bottom” type if λπ−1 0 (d) < λπ −1 1 (d) . We denote the symbol corresponding to the longer interval as w (the winner) and to the shorter one as l (the loser). The map R(π, λ) is properly defined as an interval exchange transformation of d intervals if and only if λπ−1 0 (d) 6= λπ−1 1 (d). Keane [1] gave an equivalent condition on (π, λ), for the iterations of Rauzy-Veech induction to be defined indefinitely. More precisely, we say that IET T satisfies Keane’s condition if for every two discontinuities a and b of T equality T (a) = b for some n ∈ N implies n = 1, a = T−1(0) and b = 0. In particular, if the vector λ is rationally independent, that is for every choice of cα ∈ Z, α ∈ A we have ∑ α∈A cαλα = 0 ⇒ cα = 0 for every α ∈ A, then (π, λ) satisfies Keane’s condition. When it is well defined, we denote R(π, λ) = (π, λ) for every n ∈ N. We say that the orbit of (π, λ) via Rauzy-Veech induction is ∞-complete if every symbol in A appears infinitely many times in the sequence of winners {w}. Note that λ = A(π, λ)λ, where a matrix A(π, λ) is defined in the following way Aαβ =  1 if α = β; −1 if α = w and β = l; 0 otherwise. Inductively, for every n ∈ N we define A(π, λ) = A1(πn−1, λn−1)An−1(π, λ). Then λ = A(π, λ)λ. We will refer to A(π, λ) as Rauzy-Veech matrices. Note that for every n ∈ N, the matrix (A(π, λ))−1 is non-negative. For every π ∈ SA 0 let ΘA = ΘA(π) = { τ ∈ RA; ∑ α∈A;π0(α)≤k τα > 0 and ∑ α∈A;π1(α)≤k τα < 0 for every k ∈ {1, . . . , d− 1} } . Then every (π, λ, τ) ∈ SA 0 × ΛA × ΘA1 may be see as a translation surface as follows. More precisely, first we consider two broken line 1Note that this space is not really a product space since ΘA depends on π and thus SA 0 ×ΛA×ΘA = ⋃ π∈SA 0 {π}×ΛA×ΘA(π). However, we shall use this notation for simplicity. 3 P. BERK HORIZONTAL SADDLE CONNECTIONS Figure 1. A translation surface and one step of backward Rauzy-Veech induction. The parallel segments are identified via translation. The winning segment is the one which is first crossed by a rightward separatrix starting from (0, 0).
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
Continued fractions of cubic Laurent series and their effective irrationality exponents Joint normality of representations of numbers: an ergodic approach Rigidity of non-compact static domains in hyperbolic space via positive mass theorems A characterization of chains in dimension three Divides and hyperbolic volumes
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1