Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Problem 339
claims/: The 1 claim page of Problem 339, one per claimant's result; the problem's standing derives from them.
Statement. Let be a basis of order . Must the set of integers representable as the sum of exactly distinct elements from have positive lower density?
Status. The site labels the problem proved (page last edited 2025-10-14): Hegyvári, Hennecart and Plagne [HHP03] answered it, and the companion upper-density question of Erdős and Graham, in the affirmative. The accepted claim is Hegyvári, Hennecart and Plagne 2003.
Source. erdosproblems.com/339, accessed 2026-09-04. Cite as: T. F. Bloom, Erdős Problem #339, https://www.erdosproblems.com/339.
References.
- [HHP03] Hegyvári, N. and Hennecart, F. and Plagne, A., A proof of two Erdős' conjectures on restricted addition and further results. J. Reine Angew. Math. (2003), 199-220.
Formalization. None recorded on the site, and the community database
records the problem unformalized. Boris Alexeev's lean-proofs repository
holds a Lean 4 file that declares itself a formalization of Hegyvári,
Hennecart and Plagne's solution, with the AI systems Codex and GPT-5.6 Sol as
its formal authors, linked at its pinned commit from the claim page; this
corpus has not built it.
Current assessment
Proved by a refereed paper of 2003. The site's formulation above (export of 2026-09-04, page read) asks whether, for a basis of order , the integers that are sums of exactly distinct elements of have positive lower density. The answer is yes, proved by Hegyvári, Hennecart and Plagne in J. Reine Angew. Math. 560 (2003), 199--220, which also proves the companion statement that positive upper density of the -fold sums implies positive upper density of the restricted sums; the claim page records the acceptance. The paper is not held here, its theorem numbering is not recorded, and its proof is not compiled or reviewed here.
Scope. Read: the site's problem page, the Crossref record of the paper, and the header and theorem statement of the Lean file. Not done: reading the paper, building the Lean file, and any literature search beyond these sources.