Visibility of Cartesian products of Cantor sets
Tingyu Zhang, Kan Jiang, Wenxia Li

TL;DR
This paper characterizes the set of slopes for which lines are completely outside the Cartesian product of Cantor sets generated by a specific IFS, analyzing its Hausdorff dimension and topological properties.
Contribution
It provides a complete description of the visibility set for Cartesian products of Cantor sets, including its Hausdorff dimension and topological features.
Findings
Determined the Hausdorff dimension of the visibility set V.
Established topological properties of V.
Analyzed a related slicing problem.
Abstract
Let be the attractor of the following IFS \begin{equation*} \{f_1(x)=\lambda x, f_2(x)=\lambda x+1-\lambda\}, \;\;0<\lambda<1/2. \end{equation*} Given , we say the line is visible through if Let . In this paper, we give a completed description of , e.g., its Hausdoff dimension and its topological property. Moreover, we also discuss another type of visible problem which is related to the slicing problem.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
xxxx
Tingyu Zhang, Kan Jiang and Wenxia Li Tingyu Zhang is the corresponding author
Visibility of Cartesian products of Cantor sets
Tingyu Zhang, Kan Jiang and Wenxia Li Tingyu Zhang is the corresponding author
Abstract
Let be the attractor of the following IFS
[TABLE]
Given , we say the line is visible through if
[TABLE]
Let . In this paper, we give a completed description of , e.g., its Hausdoff dimension and its topological property. Moreover, we also discuss another type of visible problem which is related to the slicing problems.
1 Introduction
Projections, sections, geodesic curves and visiblity are the main problems in geometry measure theory. It is related to many aspects of fractal geometry, for instance, the arithmetic sum of two self-similar sets is indeed the projectional problem [9, 14]; sections of some fractal sets are connected to the multiple representations of real numbers [12]; geodesic curves on fractal sets are distinct from the classical differential manifolds [6]. For more results on these problems see [19, 20, 3, 22, 21, 23, 18] and references therein. In this paper, we shall consider the visiblity of the Cartesian products of some Cantor sets.
Given and some subset , we say the line is visible through if
[TABLE]
The concept of “visiblity” was investigated by many scholars. Nikodym [13] constructed a subset of such that every point of is visible from two diametrically opposite directions. In convex geometry, Krasnosel [7] offered a beautiful criterion which enables us to check whether the entire boundary of a compact set of is visible from an interior point. Falconer and Fraser [8] proved that for a class of plane self-similar sets when the attractor has Hausdorff dimension greater than then the Hausdorff dimension of the visible subset is . The readers can find more related results in [15, 10, 1, 5].
In this paper, we shall analyze the following self-similar set. Let be the attractor with the IFS
[TABLE]
i.e.,
[TABLE]
Let
[TABLE]
It is easy to verify that the line is visible through if and only if
[TABLE]
Thus,
[TABLE]
By we denote the set of interior points of , by we denote the Lebsgue measure of . In this paper, we obtain the following results.
Theorem 1.1**.**
Let be given by (1). Then
- (1)
When , ;
- (2)
When ,
[TABLE]
- (3)
When , . In particular, when , ; when , and .
There are mainly two types of visible problem [8]. Now, we shall consider another one. First, we introduce some definitions. Let denote the line going through the origin in direction that is,
[TABLE]
Given . The visible part of is defined as follows:
[TABLE]
Let . Define
[TABLE]
In other words, we project a point to the -axis in direction Moreover, we also define the following sets.
[TABLE]
[TABLE]
Generally, is not an interval. In what follows, we always assume that , which is a natural assumption [3]. Clearly, is the attractor of the following IFS,
[TABLE]
For the IFS of , i.e. , define for and . We denote the concatenation by . Let
[TABLE]
i.e. we define . The following two propositions are motivated by the results in open dynamical systems.
Proposition 1.1**.**
Suppose that . For any , if there are some such that
[TABLE]
then
[TABLE]
is a graph-directed self-similar sets with the strong separation condition, where denotes the cardinality.
Analogously, we have the following result.
Proposition 1.2**.**
. For any , if all the possible orbits of and hit finitely many points, then apart from a countable set
[TABLE]
is a graph-directed self-similar sets with the open set condition.
Propositions 1.1 and 1.2 give a sufficient conditition which allows us to calculate the dimension of the slicing set .
The paper is arranged as follows. In section 2, we give the proof of Theorem 1.1. In section 3, we give the proofs of Propositions 1.1 and 1.2. Finally, we give some remarks.
2 Proofs of Main results
Before we prove Theorem 1.1, we give some definitions, and prove a useful lemma. Let . For any , we call a basic interval of rank which has length . Denote by the collection of all these basic intervals of rank . Suppose and are the left and right endpoints of some basic intervals in for some , respectively. Denote by the union of all the basic intervals of rank which are contained in Let be a basic interval with rank . Define .
Lemma 2.1**.**
Let be a continuous function, where is a non-empty open set. Suppose and are the left and right endpoints of some basic intervals in for some respectively such that Then . Moreover, if for any and any two basic intervals , such that
[TABLE]
then
Proof.
By the construction of , i.e. for any , it follows that
[TABLE]
The continuity of yields that
[TABLE]
Without loss of generality, we may assume that
[TABLE]
where is a basic interval in . By the condition in lemma, i.e. for any and any two basic intervals , such that
[TABLE]
it follows that
[TABLE]
Therefore, ∎
Lemma 2.2**.**
Let , and be two basic intervals. If , and then
Proof.
Note that
[TABLE]
Therefore,
[TABLE]
where
[TABLE]
Note that . In the following, we verify that
Since and , we have
[TABLE]
Now it suffices to check that
[TABLE]
We have
[TABLE]
and
[TABLE]
Finally,
[TABLE]
If , then . Therefore, we have
[TABLE]
which leads to . However, if , then
[TABLE]
Thus, we finish checking that . ∎
Lemma 2.3**.**
We have
[TABLE]
Proof.
From Lemmas 2.1 and 2.2 it follows that if
[TABLE]
Each can be uniquely represented as
[TABLE]
Note that if and only if . Thus each is of form
[TABLE]
Thus for any two with one has
[TABLE]
Thus
[TABLE]
It is easy to check that when , and intervals are pairwise disjoint when . ∎
Pourbarat [16], making use of the thickness of the Cantor sets, proved the following result.
Theorem 2.5**.**
If , then contains an interior point.
Lemma 2.6**.**
If , then contains an interior point.
Proof.
If , then . Therefore, contains an interior point by Theorem 2.5. ∎
Lemma 2.7**.**
If , then has an interior point.
Proof.
Note that . Thus, by the argument in Lemma 2.3, we have
[TABLE]
Note that the intervals for are pairwise disjoint when . Therefore, has an interior point by (2). ∎
Lemma 2.8**.**
If , then has Lebesgue measure zero.
Proof.
If , then
[TABLE]
We note that for any , we have , where denotes the projection of on the axis along lines having angle with the axis. Therefore,
[TABLE]
where we use the fact that and . for a bounded set . Thus, has Lebesgue measure zero. ∎
We will use a result given by Simon and Solomyak [17].
Theorem 2.10**.**
Let be a self-similar -set in with the open set condition, which is not on a line. Then
[TABLE]
where
[TABLE]
Lemma 2.9**.**
* has Lebesgue measure zero.*
Proof.
Note that when , is a self-similar set with the following IFS
[TABLE]
[TABLE]
Clearly, the above IFS satisfies the open set condition. Therefore, the Hausdorff dimension of is 1, and . Let
[TABLE]
The Lebesgue measure of is [math] due to Theorem 2.10. Let
[TABLE]
Clearly, . The metric on , denoted by , is the arc metric. It is well known that on , the arc metric is equivalent to the Euclidean metric. Let
[TABLE]
The metric on the Euclidean metric (we denote it by ). We define the the map
[TABLE]
by
[TABLE]
The map is indeed mapping a point on into its associated polar angle in the polar coordinate system. Therefore, we may define in another way as follows: define
[TABLE]
Clearly, is well-defined, and it is a bijection. Moreover, we shall prove that is a Lipschitz map, i.e. there exists some constant such that
[TABLE]
Note that , and that
[TABLE]
Now, follows from , , and is Lipschitz. Theorefore, ∎
The following is from Bárány [4].
Theorem 2.11**.**
Let be an arbitrary self-similar set in not contain in any line. Suppose that is a map such that
[TABLE]
for any . Then
[TABLE]
Lemma 2.10**.**
When , .
Proof.
By the argument in Lemma 2.3, we have
[TABLE]
Thus
[TABLE]
Clearly, is a two-dimensional self-similar set which is not contained in any line. Let , then
[TABLE]
for any . Therefore, in terms of Theorem 2.11,
[TABLE]
Hence, if , then
[TABLE]
∎
Proof of Theorem 1.1.
Theorem 1.1 (1) and (2) follows from Lemmas 2.3. Theorem 1.1 (3) follows from 2.6, 2.7, 2.8, 2.9 and 2.10. ∎
3 Visible sets, slicing sets and open dynamical systems
In this section, we give the proofs of Propositions 1.1 and 1.2. Define
[TABLE]
It is easy to see that the length of is In what follows, we always assume that . Clearly, in terms of , is the attractor of the following IFS,
[TABLE]
In other words,
[TABLE]
For any , there exists a sequence such that
[TABLE]
We call such a sequence a coding of . Usually, the coding of is not unique. If has a unique coding, then we call a univoque point. Write for all the univoque points of with respect to the IFS
The following result is proved in [2, Lemma 2.1] which states that any self-similar set can be regarded as a topological dynamical system. For the IFS of , i.e. , define for and . We denote the concatenation by .
Lemma 3.1**.**
Let Then is a coding for if and only if for all
Motivated by Lemma 3.1, we may define the orbits of the points of .
Definition 3.2**.**
Let with a coding , we call the set
[TABLE]
an orbit set of , where
It is easy to see that for different codings, the orbits of may be distinct.
The set can be viewed as a slicing set, i.e.
[TABLE]
where denotes the cardinality. The following lemma is trivial.
Lemma 3.2**.**
Suppose that . Then
[TABLE]
is exactly the univoque set of under the IFS i.e.
[TABLE]
Proof.
The proof is trivial. We leave it to readers. ∎
If , let , i.e.
[TABLE]
Let . Similarly, if , then we can also define the Now the following result is a corollary of the main result of [2].
Proposition 3.1**.**
Suppose that . For any , if there are some such that
[TABLE]
then
[TABLE]
is a graph-directed self-similar sets with the strong separation condition.
The following is a corollary of the main result of [11].
Proposition 3.2**.**
. For any , if all the possible orbits of and hit finitely many points, then apart from a countable set
[TABLE]
is a graph-directed self-similar sets with the open set condition.
4 Some remarks
The main idea of this paper is to establish a connection between the visible problem and arithmetic on the fractal sets. Our idea can be implemented for other overlapping self-similar sets. Similar results can be obtained if we replace the line in the definition of , by some parabolic curves or hyperbolic curves. Nevertheless, for these cases, the analysis can be difficult. We shall discuss these problems in another paper.
Acknowledgements
The work is supported by National Natural Science Foundation of China (Nos.11701302,
11671147). The work is also supported by K.C. Wong Magna Fund in Ningbo University.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] Ida Arhosalo, Esa Järvenpää, Maarit Järvenpää, Michał Rams, and Pablo Shmerkin. Visible parts of fractal percolation. Proc. Edinb. Math. Soc. (2) , 55(2):311–331, 2012.
- 2[2] Simon Baker, Karma Dajani, and Kan Jiang. On univoque points for self-similar sets. Fund. Math. , 228(3):265–282, 2015.
- 3[3] B.Bárány and M.Rams. Dimension of slices of Sierpinski-like carpets. J. Fractal Geom. , 1(3): 273-294, 2014.
- 4[4] Balázs Bárány. On some non-linear projections of self-similar sets in ℝ 3 superscript ℝ 3 \mathbb{R}^{3} . Fund. Math. , 237(1):83–100, 2017.
- 5[5] M. Bond, I.Łaba, and J. Zahl. Quantitative visibility estimates for unrectifiable sets in the plane. Trans. Amer. Math. Soc. , 368(8):5475–5513, 2016.
- 6[6] Yiming Li and Lifeng Xi. Manhattan property of geodesic paths on self-affine carpets. Arch. Math. (Basel) , 111(3):279–285, 2018.
- 7[7] Ludwig Danzer, Branko Grünbaum, and Victor Klee. Helly’s theorem and its relatives. In Proc. Sympos. Pure Math., Vol. VII , pages 101–180. Amer. Math. Soc., Providence, R.I., 1963.
- 8[8] Kenneth J. Falconer and Jonathan M. Fraser. The visible part of plane self-similar sets. Proc. Amer. Math. Soc. , 141(1):269–278, 2013.
