Wiki
Wiki

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

Updated


Source. Hughes, arXiv:2609.10902v1, Lemma 4 (greedy step), p. 2, with its three-line proof; read on the page image.

Statement

Suppose any two consecutive divisors of the integer NN differ by a factor of at most 22, and take a remainder RR with 1≤R≤N1\le R\le N. When R∣NR\mid N (for example R=NR=N), the greedy expansion stops at this step. When R∤NR\nmid N, write d<R<bd<R<b for the two consecutive divisors of NN on either side of RR; then

R−d<d,R−d≤2Rlog⁡bd.R-d<d,\qquad R-d\le2R\log\frac bd.

So each divisor the greedy expansion picks (the largest divisor not exceeding the current remainder) is smaller than the one picked before it, and the picked divisors are distinct.

Proof

The ratio hypothesis gives b≤2db\le2d, and R<bR<b, so R−d<b−d≤dR-d<b-d\le d: the new remainder is below dd, hence so is the next divisor chosen. For the second bound, b−d=d(b/d−1)b-d=d(b/d-1) and d<Rd<R give R−d<R(b/d−1)R-d<R(b/d-1), and b/d−1≤2log⁡(b/d)b/d-1\le2\log(b/d) because b/db/d lies in [1,2][1,2], where x−1≤2log⁡xx-1\le2\log x (p. 2).

Reconstruction

An author-recorded reconstruction, not an independent review, is filed as the Lemma 4 reconstruction; it also supplies a proof of the ratio hypothesis for N=n!N=n!.

Dependencies

None beyond the hypothesis; for N=n!N=n! the ratio hypothesis is the standard fact, recalled by the paper from Tenenbaum–Yokota's Lemma 4 and Yokota's Lemma 2, that consecutive divisors of n!n! have ratio at most 22.

Bears on

  • Problem 18: the step of the greedy construction counted in Theorem 1.