Sharp Bounds for Sets with Distinct Subset Products
Abstract
Let be such that for any pair of distinct subsets , the products and are distinct. We prove that , where is the prime counting function, answering a question of Erdős.
1 Introduction
For , let denote the size of the largest subset of such that for any pair of distinct subsets , . We say that a set satisfying these conditions has distinct subset products. Erdős [3] initiated the study of the quantity , proving111See Definition 1.9 for the asymptotic notation used in this paper.
where is the prime counting function.
He also produced the following example, establishing that the above bound is sharp up to the implicit constant:
Example 1.1.
Let Then has distinct subset products, so .
He then asked [3], [4] whether this estimate is optimal up to lower order terms. This problem is also listed at https://www.erdosproblems.com/795.
Question 1.2 (Erdős #795).
Is
We answer this question affirmatively:
Theorem 1.3.
.
In view of the above estimate, it is natural to ask for more precise information about the lower-order term. Erdős also considered this in [3], and speculated that a refinement of Example 1.1 may be optimal. To be more specific, for , let be the smallest element of for which there is an of size such that for any distinct , . In other words, let be the smallest possible maximal element in a set of size with distinct subset sums. Then, by setting
we see that the subset products of are distinct, and thus
Since , , , and (with ), Erdős established , and speculated that the above infinite sum may be best possible. However, we produce an example that improves upon this.
Theorem 1.4.
.
It is also natural to consider the additive variant of Question 1.2, or in other words, to determine the asymptotic behavior of the function . This is another problem of Erdős, which can be seen at https://www.erdosproblems.com/1. Although our methods do not address this problem, the interested reader may consult [2], [5] for the best-known lower bounds on and some history.
Since squares feature so prominently in Example 1.1, one may ask about what kinds of estimates can be obtained in the absence of squares. Our method answers this question as well, with a somewhat smaller second-order term.
Theorem 1.5.
Let be the maximal size of a subset of consisting of squarefree integers with distinct subset products. Then .
This estimate is also sharp up to the error term.
Theorem 1.6.
.
1.1 Notation and strategy of proof
Definition 1.7 (Subset Product Set).
Given a finite set , we define the subset product set .
Throughout the proof, we will use the fact that an element of can be divisible by at most one prime in , and at most two primes (with multiplicity) in . We thus define
Definition 1.8 (Small, Medium, and Large Primes, Valuations).
Let denote the set of primes in the set of primes in , and the set of primes in . We will also use the terms “small primes”, “medium primes”, and “large primes” to refer to elements of , , and , respectively.
For a fixed prime , and , let denote the valuation of at , i.e., the largest nonnegative integer such that divides . Then define functions
We will also define as .
We will also use some standard asymptotic notation:
Definition 1.9 (Big-O and Little-o Notation).
Given functions , , and from to , we say
-
•
if there is a constant such that for all ,
-
•
if for all , there is an such that for all , ,
-
•
if .
To prove Theorem 1.3, we will use a graph-theoretic approach. Our graphs will be simple, i.e., containing no loops or multiple edges. We will need to pay attention to certain subgraphs in our analysis:
Definition 1.10 (Paths, Cycles, Circuits).
Let be a graph.
-
•
A path of length is a collection of edges of the form
, where are distinct vertices in .
-
•
A cycle of length is a collection of edges of the form
, where are distinct vertices in .
-
•
A circuit of length is a collection of edges of the form
, where are (not necessarily distinct) vertices in .
In [3], to prove , Erdős counted the possible number of prime factorizations in elements of , where has distinct subset products. His proof is a counting argument based on the following three observations:
-
•
An element of is divisible by at most one large prime, and at most two large or medium primes (with multiplicity).
-
•
The range of on is small (see Proposition 2.1).
-
•
There are at most elements of such that there is a prime for which divides , but no other element of is a multiple of .
Optimizing his argument, one can obtain the bound . We will utilize each of these observations in our approach, but we take into account some more refined information as well. For example, if are large primes and are medium primes, the argument from [3] does not use the fact that cannot contain the elements , , , and (although it does show that cannot contain this configuration for many primes ).
The strategy of our proof is as follows. Given a set with distinct subset products, we can produce a graph whose vertices correspond to large and medium primes, and whose edges correspond to (most) elements of . We call this graph the prime factorization graph of . The condition that has distinct subset products restricts the number of short cycles that can appear in this graph. We can exploit this fact to prove an upper bound on the number of edges in this graph, and thus an upper bound on . In Section 2, we show how to construct this prime factorization graph. In Section 3, we use the condition that has distinct subset products to remove circuits and cycles from the prime factorization graph without removing too many edges. In Section 4, we estimate the number of edges in a graph without any short cycles, and finally establish Theorems 1.3 and 1.5.
Acknowledgements
The author would like to thank Terence Tao for many helpful conversations and for introducing the author to this problem.
2 Constructing the Prime Factorization Graph
Given a set with distinct subset products, we will construct a graph that encodes the large and medium prime factors of (most) elements of . For this to be effective, we first need to control how much information is lost by ignoring small primes.
Proposition 2.1.
Let denote the subset product set of . Then
Proof.
For a fixed and , . Since an element of is a product of at most elements of , for a fixed , . Thus
In view of this proposition, if has distinct subset products, there cannot be many elements of with the same valuations at medium and large primes. In fact, the same is true for .
Proposition 2.2.
Given , let . Let . Then
Proof.
For each , let and each have size . Then
each have the same valuations at all large and medium primes. There are thus at most elements of of the form . There are at least
elements of this form, where we used the inequality for all . Thus
and taking logarithms gives the desired inequality. ∎
Corollary 2.3.
Given with distinct subset products, one can remove elements of to get a set such that is injective when restricted to . By removing at most one more element, we may assume for all
Because of this corollary, we will assume here and in the next section that is injective and nonzero on . Observe also that for , is divisible by at most one large prime, and at most two large or medium primes (with multiplicity). Thus the tuple either consists of:
-
•
a 1 at one index, with 0 at all other indices,
-
•
a 1 at each of two indices, at most one of them from , and a 0 at all other indices, or
-
•
a 2 at an index from and a 0 at all other indices.
Under these conditions, we may define the prime factorization graph of a set .
Definition 2.4 (Prime Factorization Graph).
The prime factorization graph associated to has vertex set . We connect to a vertex by an edge if there is an such that and for all other primes . We connect by an edge if there is an element of divisible by both and .
By the aforementioned discussion, there is a bijection between the edges of and the elements of which are not divisible by the square of any medium prime. We will still need to consider those elements, so we define the following.
Definition 2.5 (, ).
We define and .
Then is the sum of the number of edges in and .
Example 2.6.
Suppose , so , , and is the remaining set of primes in . On the set , is injective and nonzero. Its prime factorization graph is (omitting the isolated vertices corresponding to the primes in ):
Each element of corresponds to an edge in the graph except for , since is a multiple of and .
Remark. Suppose is divisible by for some . In our definition of , we do not include an edge corresponding to . We could alternatively define the prime factorization graph so that would correspond to a loop from to itself. One can still make sense of the arguments in the following sections that way, but we found it clearer to express our graph-theoretic arguments using only simple graphs.
Having defined the necessary sets of primes, we will state our main estimate.
Theorem 2.7.
Let have distinct subset products and let . Then
3 Cycle Removal
Throughout this section, will have distinct subset products and be such that is injective and nonzero on . Unless otherwise specified, the graph will denote the prime factorization graph of .
Lemma 3.1.
Let be the sets of edges in a maximal collection of edge-disjoint even circuits in , each with length at most . Then , so .
Proof.
For each , let be the set of vertices in the cycle indexed so that . Let be the subset of corresponding to the edges , and let be the subset of corresponding to all the other edges in . Then
Thus, for any choice of ,
has the same valuation at all large and medium primes. By Proposition 2.1, we thus have , as desired. ∎
Example 3.2.
Suppose , , and }. Below is the prime factorization graph of , containing two even circuits. The product of the elements on the blue edges and the product of the elements on the red edges have the same valuations at all large and medium primes. By swapping the red-blue labeling in either circuit, we can produce four subsets whose products have the same valuations at all large and medium primes.
Corollary 3.3.
By removing edges from , we may assume that has no even circuits of length at most .
Proof.
Let be the sets of edges in a maximal collection of edge-disjoint even circuits in , each with length at most . We then have , so we may remove all the edges from from . After doing so, if there remains an even circuit of length at most in , this would contradict the maximality of . ∎
From here forward, we will assume that contains no even circuits of length at most . Having removed short even circuits from , we now turn our focus to cycles with odd length. We first have the following proposition for a general graph .
Proposition 3.4.
Let be a graph with no even circuits of length at most . Then any two odd cycles in of length at most are vertex-disjoint.
Proof.
We will prove the contrapositive. Suppose contains two odd cycles with vertex sets and , respectively, such that . Supposing also that , we will show that has an even circuit of length at most . Let and be their respective edge sets. We consider two cases.
First, assume Then form a circuit of length , which is even and at most .
On the other hand, suppose . Since , we may assume without loss of generality that , , but . Let be the maximal index such that for all , . Let and be the edge sets from the two paths in from to . Since is odd, we can combine one of or with the edges from to form an even circuit of length at most . ∎
Lemma 3.5.
Let be the sets of edges in a maximal collection of edge-disjoint odd circuits in , each with length at most and each having a vertex from . Then , so
Proof.
For each , let be the set of vertices in the cycle indexed so that and . For each , let be the prime corresponding to , and let be the element of such that divides . Note that by Proposition 3.4, the elements are distinct for distinct . Let be the set of elements of corresponding to the edges , and let be the set of elements of corresponding to the other edges in , union .
Then
Thus, for any choice of ,
has the same valuation at all large and medium primes. By Proposition 2.1, we thus have , as desired. ∎
By a similar argument to Corollary 3.3, by removing edges, we may assume without loss of generality that has no odd cycles of length at most with a vertex from .
Example 3.6.
Suppose , , , and . Below is the prime factorization graph of . The product of the red elements (including ) and the product of the blue elements have the same valuations at all large and medium primes.
Lemma 3.7.
Let denote the edge set of . There is a set of edges of size at most such that the graph with edge set contains no cycles of length at most .
Proof.
Let be the sets of edges of all cycles of length at most in . By Proposition 3.4, we may assume that these sets are disjoint and that they share no vertices. None of these cycles have a vertex from , and no vertices in are adjacent, so each contains an edge between two vertices from . Since the vertex sets are disjoint, we thus have . We may take to consist of one edge from each cycle . ∎
4 Proof of Theorems 1.3 and 1.5
Having removed all cycles of length at most from the prime factorization graph, we are ready to estimate the number of edges in it. We first record the following general estimate. This can be deduced from a result of Alon, Hoory, and Linial [1], but it is easier in our case, so we present a self-contained proof. The proof uses a standard breadth-first search argument; a similar argument can be found in [6], Theorem 1.6.5, for example.
Lemma 4.1.
Let be a graph with vertices and at least edges. Then has a cycle of length at most .
Proof.
Without loss of generality, we may assume that is connected, since we may pass to a component satisfying . We can remove degree vertices and their incident edges while preserving the inequality , so we may assume without loss of generality that each vertex has degree at least 2. If contains a path of length greater than , then we can remove all internal edges and vertices from that path while preserving the inequality , so we may assume that has no such paths.
With all these assumptions set, we fix a vertex and consider a breadth-first search starting at . For each , let . Fix with and let be the least integer greater than .
If there is a such that there are two distinct paths of length from to , then we are done, since must thus contain a cycle of length at most . Otherwise, for each , let be the vertex from on the shortest path from to . For each , we must have , since otherwise the path from to the only element in would be a path of length greater than consisting only of degree two vertices. We thus have . Letting be the least integer greater than , we must thus have , so there must be a with at least two distinct paths from to . There is thus a cycle of length at most in , as desired. ∎
We will apply Lemma 4.1 with The following fact is required to ensure that the error term from Lemma 4.1 is multiplied by instead of in our final estimate.
Lemma 4.2.
Let be the prime factorization graph of , containing no even circuits of length at most . The set of vertices from with degree at least has size at most .
Proof.
Let be the set of vertices of degree at least in . Recall that no pair of vertices in is connected by an edge, so if , then there are distinct which are each connected to by an edge.
Consider the graph with vertex set , where is adjacent to if there is a path of length from to in with the middle vertex in . By construction, there is a surjection from the set of edges in this graph to . Moreover, this graph is simple since contains no -cycles.
Any cycle of length in this graph corresponds to a circuit of length in . This graph thus does not contain any cycles of length at most . Applying Lemma 4.1 with , we thus find that the number of edges is at most , so . ∎
Proof of Theorem 2.7.
Let have distinct subset products. By Corollary 2.3, we may remove elements from to ensure that is injective and nonzero when restricted to . We may now let be the prime factorization graph of .
By Corollary 3.3 and Lemma 3.5, we may remove elements from to ensure that has no even circuits of length at most , and to ensure that has no odd cycles of length at most with a vertex from . Then, we may apply Lemma 3.7 and remove edges to remove all odd cycles of length at most from .
Let be the set of all vertices in with degree at least two. We remove all vertices from and their incident edges. Let and be the remaining sets of vertices and edges after all these removals. We have
and
The graph has no cycles of length at most . Applying Lemma 4.1 with , we find
This results in the inequality
where we used Lemma 4.2 so that .
Finally, using the equation , we find
since and . ∎
5 Examples
In this section, we will discuss some examples of large sets with distinct subset products.
Our graph-theoretic perspective enables us to produce some new examples of sets with distinct subset products and .
Example 5.1.
Let be a tree with vertex set . Let . Then and has distinct subset products.
Next, we will produce an example of a set with distinct subset products and , establishing Theorem 1.4.
Proof of Theorem 1.4.
Construct a set as follows. For any prime , let . For any prime , let . Let be distinct primes in . The set is a subset of with distinct subset products. We may thus divide the primes from into disjoint subsets of size (perhaps leaving one or two primes out) and for each triple , add the elements to . The number of elements of is thus at least
Finally, we may use our graph-theoretic perspective to produce a set of squarefree integers with distinct subset products and nearly elements, establishing Theorem 1.6.
Proof of Theorem 1.6.
Let , and let be sufficiently large (depending on ). By the prime number theorem, for every , the number of primes satisfying the inequality
is less than twice the number of primes satisfying
We may thus list all but at most of the primes in as so that there are distinct primes in with , for all .
Let be the set of large primes not in . Let
Then and has distinct subset products.
The prime factorization graph of is as follows (omitting the vertices from ); it is constructed so that the argument from Lemma 3.7 is sharp.
We have established that for all , there is an such that for all , . Thus, . ∎
References
- [1] Noga Alon, Shlomo Hoory, and Nathan Linial. The Moore bound for irregular graphs. Graphs and Combinatorics, 18:53–57, 2002.
- [2] Quentin Dubroff, Jacob Fox, and Max Wenqiang Xu. A note on the Erdős distinct subset sums problem. SIAM Journal on Discrete Mathematics, 35(1):322–324, 2021.
- [3] Paul Erdős. Extremal problems in number theory II. Mat. Lapok, 17:135–155, 1966.
- [4] Paul Erdős. Some applications of graph theory to number theory. In The Many Facets of Graph Theory: Proceedings of the Conference held at Western Michigan University, Kalamazoo/MI., October 31–November 2, 1968, pages 77–82. Springer, 1969.
- [5] Stefan Steinerberger. Some remarks on the Erdős distinct subset sums problem. International Journal of Number Theory, 19(08):1783–1800, 2023.
- [6] Yufei Zhao. Graph Theory and Additive Combinatorics: Exploring Structure and Randomness. Cambridge University Press, 2023.