Bounded ranges of cardinal functions
Jacek Marchwicki, Błażej Żmija · 2025
Read Bounded ranges of cardinal functions by Jacek Marchwicki, Błażej Żmija (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-2508-13016.txt.
Introduction
Marchwicki and Żmija study how many ways a number can be built from a fixed list of summands. Given a sequence of positive reals, its achievement set collects every sum its elements can form. The cardinal function then counts, for each such sum, in how many ways it can be assembled. The paper asks which sets of counts can actually arise as the range of such a function.
The authors give general constructions and several tests a set must pass to qualify. They focus on ranges whose largest value is six, and for interval-filling sequences they characterize those ranges completely. The work maps the border between possible and impossible multiplicity patterns for subset sums.
Catalog record
| CREATOR | Jacek Marchwicki, Błażej Żmija |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 89 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-13016
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Bounded ranges of cardinal functions Thanks: The second author was supported by Czech Science Foundation (GAČR) grant 21-00420M, Charles University Research Centre program UNCE/SCI/022, and PRIMUS/25/SCI/017. Jacek Marchwicki, Błażej Żmija Abstract . Let 𝐱 \mathbf{x} be a (non-empty) sequence of positive real numbers. Its achievement set 𝐱 \mathcal{\mathbf{x}} is the set of all the possible sums of the elements of 𝐱 \mathbf{x} . The cardinal function of 𝐱 \mathbf{x} is the function f : 𝒜 ( 𝐱 ) → ℕ ∪ { ω , 𝔠 } f:\mathcal{A}(\mathbf{x})\to\mathbb{N}\cup\{\omega,\mathfrak{c}\} that for every x ∈ 𝔸 ( 𝐱 ) x\in\mathbb{A}(\mathbf{x}) the value f ( x ) f(x) is equal to the number of ways x x is represented as a sum of elements of 𝐱 \mathbf{x} . In this paper we consider possible ranges of cardinal functions of sequences 𝐱 \mathbf{x} . We present some general constructions and several criteria that a set has to satisfy in order to be a range of a cardinal function. We put special attention to the case of sets with maximal element equal to 6 6 . In this case, in particular, we obtained a full characterisation of sets that are ranges of cardinal functions of interval-filling sequences. 2010 Mathematics Subject Classification Primary: 40A05; Secondary: 11K31 1. Introduction By the achievement set of a sequence \mathbf{x}=(x_{n}) we mean the set of all subsums of the series \sum_{n=1}^{\infty}x_{n} , that is, \displaystyle\mathcal{A}(\mathbf{x})=\left\{\ \sum_{n=1}^{\infty}\varepsilon_{n}x_{n}\ \Bigg|\ (\varepsilon_{n})\in\{0,1\}^{\mathbb{N}}\ \right\}=\left\{\ \sum_{n\in A}x_{n}\ \Bigg|\ A\subseteq\mathbb{N}\ \right\}. The cardinal function f for the sequence \mathbf{x} (or equivalently for the achievement set \mathcal{A}(\mathbf{x}) ) counts the number of representations as subsums for each element of the achievement set. That is, f:\mathcal{A}(\mathbf{x})\rightarrow\{1,2,\ldots,n,\ldots,\omega,\mathfrak{c}\} and for any x\in\mathcal{A}(\mathbf{x}) we define \displaystyle f(x):=\#\left\{\ (\varepsilon_{n})\in\{0,1\}^{\mathbb{N}}\ \Bigg|\ \sum_{x_{n}\neq 0}\varepsilon_{n}x_{n}=x\ \right\}, where the summation is over those x_{n} ’s which are non-zero. We will sometimes write f_{\textbf{x}} if it is not clear, which sequence we are referring to. Note that if \mathbf{x}=(x_{n}) is eventually zero, then \mathcal{A}(\mathbf{x}) is finite and f(x)<\omega for any x\in\mathcal{A}(\mathbf{x}) . By the boundedness of f we mean that there exists a natural number N such that f(x)\leq N for all x . There are two types of unboundedness: first, when f attains infinite value \omega or \mathfrak{c} and the second, when f achieves infinitely many distinct natural numbers. We will also use the notation f_{k} for the cardinal function for \mathcal{A}((x_{n})_{n\geq k}) . Cardinal functions were introduced in [ 8 ] . The first paper where achievement sets are considered is that of Kakeya [ 10 ] . The author claimed that if (x_{n})\in l_{1} then \mathcal{A}(x_{n}) is a finite set, a finite union of closed intervals or a Cantor set. However, due to [ 9 ] and [ 13 ] , we know that there is one more possibility: so called Cantovals. These are compact subsets of \mathbb{R} with dense interiors. Note that first known examples of Cantorvals were given in the papers [ 7 ] and [ 14 ] . The shape of a cardinal function may depend on one of the four classes to which the achievement set belongs. Therefore, throughout the paper we will need the following notation: \displaystyle\mathcal{I} \displaystyle:=\big\{\ \textrm{possible ranges of cardinal functions obtained for the achievements sets being unions of } \displaystyle\ \ \ \ \textrm{ closed intervals}\ \big\}, \displaystyle\mathcal{C}_{v} \displaystyle:=\big\{\ \textrm{possible ranges of cardinal functions obtained for the achievements sets being cantorvals}\ \big\}, \displaystyle\mathcal{C} \displaystyle:=\big\{\ \textrm{possible ranges of cardinal functions obtained for the achievements sets being Cantor sets}\ \big\}, \displaystyle\mathcal{F} \displaystyle:=\big\{\ \textrm{possible ranges of cardinal functions obtained for the achievements sets being finite sets}\ \big\}. The subset of \mathcal{I} corresponding to those f whose domain (an achievement set) is a single close interval will be denoted by \mathcal{I}_{1} . Let us recall some known facts about cardinal functions. Most of the results come from the paper [ 8 ] . For example, it was proved that \mathcal{F}\subseteq\mathcal{C} , namely for any finite achievement set its counterpart Cantor set with the same range of cardinal function was constructed. Moreover if M\in\mathcal{F} , then M\cup\{\mathfrak{c}\}\in\mathcal{I} . Let us denote \displaystyle\mathcal{R}:=\mathcal{I}\cup\mathcal{C}_{v}\cup\mathcal{C}\cup\mathcal{F}, that is, the set of all the possible ranges of cardinal functions of summable sequences. It is easy to see that \mathcal{R}\subsetneq\mathcal{P}(\{1,2,\ldots,n,\ldots,\omega,\mathfrak{c}\})\setminus\{\emptyset\} , where \mathcal{P}(X) is the power set of X . Indeed, for any summable sequences (x_{n}) the maximal and minimal elements of \mathcal{A}(x_{n}) are obtained uniquely. Thus for each M\in\mathcal{R} we have 1\in M . The more complicated question is whether the equality \displaystyle\mathcal{R}=\big\{\ \{1\}\cup M\ \big|\ M\in\mathcal{P}(\{2,\ldots,n,\ldots,\omega,\mathfrak{c}\})\ \big\} holds, since we do not know any nontrivial method of showing that a given set is not in \mathcal{R} . It is also known that \{1\}\in\mathcal{C}\setminus(\mathcal{I}\cup\mathcal{C}_{v}) , that is, the case when each point in achievement set is obtained by a unique representation is reserved for Cantor sets (and for \mathcal{F} in the more trivial sense of finite subsums). Indeed, note that the mapping \{0,1\}^{\mathbb{N}}\rightarrow\mathcal{A}((x_{n})) defined as (\varepsilon_{n})\mapsto\sum_{n=1}^{\infty}\varepsilon_{n}x_{n} is continuous on the compact domain. If {\rm rng}\hskip 0.85355pt(f) denotes the range of f , then the equality {\rm rng}\hskip 0.85355pt(f)=\{1\} means that f is a bijection and hence a homeomorphism. Thus the image \mathcal{A}((x_{n})) is homeomorphic to the Cantor space \{0,1\}^{\mathbb{N}} . There are still some unsolved problems in the area. For instance, the question: does \omega\in M imply that \mathfrak{c}\in M ? remains unanswered. The method of proving that some set belongs to \mathcal{R} are not well developed. From the few of them the most important is that if M\in\mathcal{R} and L\in\mathcal{F} , then M\cdot L\in\mathcal{R} . In particular, this implies that M\cup 2M=\{1,2\}\cdot M\in\mathcal{R} , and similarly M\cup 3M\in\mathcal{R} , M\cup 4M\cup 6M\in\mathcal{R} or in general \bigcup_{j=0}^{p}{p\choose j}M\in\mathcal{R} for each p\in\mathbb{N} . The problem whether the conditions M\in\mathcal{R} and L\in\mathcal{R} imply that M\cdot L\in\mathcal{R} remains open. Probably the simplest set about which we still do not know if it belongs to \mathcal{R} is \{1,4\} . Note that the sets \{1,2\} , \{1,3\} , \{1,2,3\} , \{1,2,3,4\} , \{1,2,4\} , \{1,3,4\} are all known to be the elements of \mathcal{R} . Another known fact is that if M\in\mathcal{R} , then M\cup\{2\}\in\mathcal{R} . The paper is organised as follows. In Section 2 we present some basic properties. Section 3 is dedicated to answering the questions does \{1,4\}\in\mathcal{I}_{1} ? and more general does \{1,m\}\in\mathcal{I}_{1} for m\in\mathbb{N} ? by using the concept of lockers [ 5 ] . In Section 4 we discuss how the upper bound of cardinal function is connected with Kakeya conditions [ 11 ] . Then in Section 5 we introduce a useful tool called Cardinality Invariant Sets which help us to recognize unbounded ranges. Section 6 is dedicated to distinguish \mathcal{I} from \mathcal{I}_{1} by showing that \mathcal{I}\setminus\mathcal{I}_{1} is nonempty. In Section 7 we obtain more results about \mathcal{I}_{1} , which allow us to complete the characterization of ranges bounded by 6 . Sections 8 and 9 are dedicated to studying \mathcal{F} , \mathcal{C} and \mathcal{C}_{v} respectively. In Section 10 we briefly summarize the results by presenting a table showing which sets M with \max M\leq 6 belong to sets \mathcal{I} , \mathcal{C}_{v} , \mathcal{\mathbb{C}} , \mathcal{F} and \mathcal{R} . Filling this table is the reason, why we write down many corollaries specifically for sets