Numbers omitting digits in certain base expansions
Alexia Yavicoli, Han Yu · 2025
Read Numbers omitting digits in certain base expansions by Alexia Yavicoli, Han Yu (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-2503-09528.txt.
Introduction
Alexia Yavicoli and Han Yu study integers whose expansions in several bases all skip the digit zero. An earlier result showed that for large enough bases with a multiplicative independence condition, infinitely many such integers exist. This paper drops that condition. It also makes the result quantitative, giving an explicit value for how large the bases must be.
The authors extend the finding beyond zero: the missing digits need not be zero. They also generalize the method to missing-digit sets in an algebraic setting. The work sharpens a known theorem in number theory by removing a restrictive hypothesis and replacing an implicit bound with a concrete one. The authors present their technique as reusable for related digit problems.
Catalog record
| CREATOR | Alexia Yavicoli, Han Yu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-09528
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Numbers omitting digits in certain base expansions Alexia Yavicoli, Han Yu Abstract . In [ 6 ] the author proved that for each integer k k there is an implicit number M > 0 M>0 such that if b 1 , ⋯ , b k b_{1},\cdots,b_{k} are multiplicatively independent integers greater than M M , there are infinitely many integers whose base b 1 , b 2 , ⋯ , b k b_{1},b_{2},\cdots,b_{k} expansions all do not have zero digits. In this paper we don’t require the multiplicative independence condition and make the result quantitative, getting an explicit value for M M . We also obtain a result for the case when the missing digit(s) may not be zero. Finally, we extend our method to study various missing-digit sets in an algebraic setting. 1991 Mathematics Subject Classification MSC 28A78, MSC 28A80, MSC28A12 and MSC 11B25 1. Introduction and main results In this paper, we utilise the notion of thickness for sets in \mathbb{R}^{d},d\geq 1 to prove several results concerning numbers with missing digits in various different possibly algebraic bases. To start, we first consider the easy setting with ‘usual’ missing digit sets with rational integer bases. Definition 1 . We say that a natural number N can be written in basis b without the digit 0 , if there is a j\in{\mathbb{N}}_{0} so that N=\sum_{0\leq i\leq j}c_{i}b^{i}\text{ where }c_{i}\in\{1,\cdots,b-1\}. We say that a positive real number R can be written in basis b without the digit 0 , if there is a j\in{\mathbb{Z}} so that R=\sum_{-\infty<i\leq j}c_{i}b^{i}\text{ where }c_{i}\in\{1,\cdots,b-1\}. In [ 6 ] the author proved that for each integer k there is an implicit number M>0 such that if b_{1},\cdots,b_{k} are multiplicatively independent integers greater than M , there are infinitely many integers whose base b_{1},b_{2},\cdots,b_{k} expansions all do not have zero digits. In this paper, we make the result quantitative, getting an explicit value for M , where the multiplicative independence hypothesis is not required. Theorem 2 . Let k\geq 2 be an integer. Let (b_{1},\cdots,b_{k}) be a k -tuple of natural numbers. Let M=M(k)\in{\mathbb{N}}_{\geq 5} be the smallest number satisfying that g_{k}(M-2):=\frac{M-2}{\log(M-2)}-k\frac{e4(432)^{2}}{\log(4)}>0 . If b_{i}\geq M=M(k) for every 1\leq i\leq k , then there are infinitely many integers whose base b_{1},\cdots,b_{k} expansions all omit the digit 0 . Remark 3 . For a concrete example, we can take k=3. In this case, we need M such that \frac{M-2}{\log(M-2)}>\frac{2239488e}{\log(4)}>4391245. Thus M=79904626 would be enough. It is convenient to find some easier “gauge” functions to replace g_{k} . For example, for any \tilde{\varepsilon}\in(0,1) , if k is large enough in terms of \tilde{\varepsilon} then M\approx k\log^{1+\tilde{\varepsilon}}(k) works. We also prove a result for when the missing digit(s) may not be zero. Theorem 4 . Let k\geq 2 be an integer, (b_{1},\cdots,b_{k}) be a k -tuple of natural numbers and r\in[\frac{1}{\min_{1\leq i\leq k}b_{i}},1) . Let M_{1},\cdots,M_{k} be sets (of missing digits) satisfying: • 0\in M_{i}\subseteq\{0,\cdots,b_{i}-2\} , • M_{i} does not have consecutive numbers, • \{0,\cdots,b_{i}-1\}\setminus M_{i} is formed by rows of at least b_{i}r consecutive numbers. Let M=M(k)\in{\mathbb{N}}_{\geq 4} be the smallest number satisfying that g_{k}(r(M-2)):=\frac{r(M-2)}{\log(r(M-2))}-k\frac{e4(432)^{2}}{\log(4)}>0. If b_{i}\geq M=M(k) for every 1\leq i\leq k , then there are infinitely many integers whose base b_{1},\cdots,b_{k} expansions omit the digits in the sets M_{1},\cdots,M_{k} respectively. Finally, we extend our method to study various missing-digit sets in an algebraic setting. The sets in this scenario are self-similar, possibly with rotations. We defer the detailed description of all the terminologies and notations in Section 4 , where we prove the following result. Theorem 5 . Let \mathbb{K} be a number field so that O_{\mathbb{K}} is an Euclidean domain. For each k\geq 2, there is a number M>0 such that for each k different integers in O_{\mathbb{K}} , say, b_{1},\dots,b_{k} with equal norm Galois embeddings, and with norm at least M , there are infinitely many integers z\in O_{\mathbb{K}} with the property that z misses one arbitrarily fixed digit in base b_{j} for all j . Remark 6 . Here O_{\mathbb{K}} is the ring of integers in \mathbb{K} . We require that it is an Euclidean domain. This is more for convenience than necessity. Our method would work in general number fields with careful modifications. In this paper, we do not pursue this level of generality. Another condition is that b_{1},\dots,b_{k} are integers with equal norm Galois embeddings, e.g. integers in complex quadratic fields. This condition is crucial for the application of the thickness argument. Should this condition fail, the geometry of the missing-digit set would be not self-similar but self-affine with small thickness. Remark 7 . This result can be made quantitative with the explicit constants provided in Theorem 11 as well as the detailed knowledge of the number field \mathbb{K} . We do not pursue this refinement in this paper. We need the notion of thickness on the real line. Recall that every compact set C on the real line can be constructed by starting with a closed interval I (its convex hull), and successively removing disjoint open complementary intervals (they are the path-connected components of the complement of C ). Clearly there are finitely or countably many disjoint open complementary intervals (G_{n})_{n} , which we may assume are ordered so that their length |G_{n}| is decreasing (if some intervals have the same length, we order them arbitrarily). We emphasize that the two unbounded connected components of \mathbb{R}\setminus C are not considered as we start with the convex hull I . Note that G_{n+1} is a subset of some connected component I_{n+1} (a closed interval) of I\setminus(G_{1}\cup\cdots\cup G_{n}) . We say that G_{n+1} is removed from I_{n+1} . Definition 8 . Let C\subset\mathbb{R} a compact set with convex hull I , and let (G_{n}) be the open intervals making up I\setminus C , ordered in decreasing length. Each G_{n} is removed from a closed interval I_{n} , leaving behind two closed intervals L_{n} and R_{n} ; the left and right pieces of I_{n}\setminus G_{n} . We define the thickness of C as \tau(C):=\inf_{n\in{\mathbb{N}}}\frac{\min\{|L_{n}|,|R_{n}|\}}{|G_{n}|}. Note that the sequence of complementary intervals (G_{n})_{n} may be finite, and in this case the infimum is taken over the finite set of indices. We define the thickness of a singleton as 0 , and the thickness of a (non-degenerate) interval as +\infty . It can be checked that if there are some intervals of equal length, then the way in which we order them does not affect the value of \tau(C) . See [ 4 , 3 , 1 ] for more information on Newhouse thickness and alternative definitions. Example 9 . Let M_{\varepsilon} be the middle- \varepsilon Cantor set obtained by starting with the interval [0,1] and repeatedly deleting from each interval appearing in the construction the middle open interval of relative length \varepsilon . Then \tau(M_{\varepsilon})=\frac{1-\varepsilon}{2\varepsilon} . For the algebraic setting, we need the following result from [ 2 , Theorem 6] to guarantee that intersections of thick sets in \mathbb{R}^{d},d\geq 1 are non-empty. The definition of thickness can be generalised to \mathbb{R}^{d} . See [ 2 ] . Given a compact subset C of \mathbb{R}^{d} , we define (G_{n})_{n=1}^{\infty} to be the (at most) countably many open bounded path-connected components of the complement C and