Erdős Problem #281 — Suppose n1<n2<…n_1 < n_2 < \dots are such that for every choice of aia_i the set of integers not satisfying any of the congruences ai(modni)a_i \pmod{n_i} has density 0.

About 46 years old · traced to

Suppose n1<n2<…n_1 < n_2 < \dots are such that for every choice of aia_i the set of integers not satisfying any of the congruences ai(modni)a_i \pmod{n_i} has density 0. In this case we must have ∑i1ni=∞\sum_i \frac{1}{n_i} = \infty and, if the nin_i are pairwise relatively prime, then this suffices. This property clearly holds if for every ε\varepsilon there is a kk so that the density of integers not satisfying ai(modni)a_i \pmod{n_i}, 1⩽i⩽k1 \leqslant i \leqslant k, is less than ε\varepsilon. Is this in fact necessary?

References

Additional references

Erdős, P. and Graham, R., Old and new problems and results in combinatorial number theory. Monographies de L'Enseignement Mathématique 28 (1980).

Progress summary

Refreshed
Claimed solved

The question has been settled affirmatively: classical results show that sufficiently many initial congruence classes leave arbitrarily little uncovered.

The problem asks whether pointwise density-zero coverage by an infinite sequence of moduli can be made uniform over all choices after a finite initial segment. The affirmative conclusion is attributed to Somani, with an argument also presented through classical theorems of Davenport–Erdős and Rogers.

Known results

  • Davenport–Erdős: the relevant lower density equals the limit of the finite-prefix densities.
  • Rogers: uncovered density is maximized when every selected residue class is zero.

Affirmative resolution

Somani’s proof discussion derives the required uniform finite-prefix conclusion from those two theorems. A Lean formalization records the argument and labels the result “research solved,” providing corroborating formal evidence.

Current status (as of March 2026): The stated problem is resolved affirmatively; only questions about possible intended variants remain separate.

Sources

Solutions 0

No solutions have been posted yet.