← All problems

Erdős Problem #318 — Signed reciprocal subsums on arithmetic progressions

46 years old · traced to

Let N\mathbb{N} denote the positive integers. For every infinite arithmetic progression A⊆NA\subseteq\mathbb{N} and every non-constant function f:A→{−1,1}f:A\to\{-1,1\}, must there exist a finite, non-empty set S⊆AS\subseteq A such that

∑n∈Sf(n)n=0?\sum_{n\in S}\frac{f(n)}{n}=0?

The related questions are whether the same assertion holds when AA is an arbitrary set of positive density, and whether it holds when A={k2:k∈N, k≥2}A=\{k^2:k\in\mathbb{N},\ k\ge 2\}, the squares excluding 11. Erdős Problems #318

Background

The problem concerns exact cancellation among finitely many signed reciprocal terms, with the signs prescribed by a non-constant function on the chosen set. The full set of positive integers and the odd integers were early cases: Erdős and Straus established the assertion for A=NA=\mathbb{N}, and Sattler established it for the odd numbers. The exclusion of 11 in the square case is essential: since ∑k≥21/k2<1\sum_{k\ge2}1/k^2<1, assigning one sign to the term at 11 and the opposite sign to every other square prevents any finite signed sum from cancelling it. Erdős Problems #318 Erdős and Straus (1975)

References

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.