Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated


Statement

Setting (p. 2). For a finite set DD of positive integers, H(D)\mathcal H(D) is the hypergraph on the vertex set DD whose hyperedges are the sets {d}∪E⊆D\{d\}\cup E\subseteq D with d∉Ed\notin E, E≠∅E\ne\varnothing and

1d=∑e∈E1e\frac1d=\sum_{e\in E}\frac1e

(the paper's equation (2)). A subset of DD is independent when it contains no hyperedge, and α(D)\alpha(D) is the largest size of an independent subset of DD. A forbidden solution is a choice of pairwise distinct a,b1,…,bka,b_1,\ldots,b_k with k≥1k\ge1 and 1/a=1/b1+⋯+1/bk1/a=1/b_1+\cdots+1/b_k (the paper's equation (1), p. 1), and [N]={1,…,N}[N]=\{1,\ldots,N\}.

Lemma 1 (Dilation lemma, p. 2). Let DD be a finite set of positive integers and mm a positive integer. If A⊆[N]A\subseteq[N] contains no solution of (1), then

∣A∩mD∣≤α(D)|A\cap mD|\le\alpha(D)

whenever mD⊆[N]mD\subseteq[N].

Source. Xinjun Wang, A 667/806 Upper Bound for Erdős Problem #301 on Unit-Fraction-Free Sets, unpublished manuscript dated May 27, 2026 on its title page, posted on ResearchGate (2026), identified on the source card: Lemma 1 and the definition of H(D)\mathcal H(D) on p. 2, in Section 2 (pp. 2--3). An unrefereed manuscript, which the site's Problem 301 discussion thread describes as AI-generated.

Read depth. Claims checked: the statement and the definitions it uses were read clause by clause on the page image. Nothing here is independently reviewed.

Proof pointer

P. 2. If A∩mDA\cap mD had more than α(D)\alpha(D) elements, the elements of DD it comes from would contain a hyperedge {d}∪E\{d\}\cup E; multiplying the reciprocal identity through by 1/m1/m turns it into a solution of (1) with pairwise distinct terms mdmd and meme (e∈Ee\in E), all in AA.

Dependencies

None.

Bears on

  • Problem 301: the lemma is the step that turns a finite certificate on DD into a count of elements a relation-free set must omit from each dilate; the manuscript's Theorem 1 applies it to the prefixes of the divisor set of Proposition 1. The lemma alone gives no density bound for f(N)f(N).