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. 276). Ascending waves are as on the Theorem 1.1 page: integers x1<⋯<xkx_1<\cdots<x_k with non-decreasing consecutive differences. g=g(n)g=g(n) is the largest positive integer such that every set A⊆{1,2,…,n}A\subseteq\{1,2,\ldots,n\} with ∣A∣≥12n|A|\ge\frac12n contains an ascending wave of length gg.

Theorem 2.1 (p. 276, quoted). "Ω(log⁡2n/log⁡log⁡n)≤g(n)≤O(log⁡2n)\Omega(\log^2 n/\log\log n)\leq g(n)\leq O(\log^2 n)."

The abstract (p. 275) states it with constants: there are positive constants c3,c4c_3,c_4 with c3(log⁡n)2/log⁡log⁡n≤g(n)≤c4(log⁡n)2c_3(\log n)^2/\log\log n\le g(n)\le c_4(\log n)^2 for all n≥1n\ge1.

Context (p. 276). The paper records the earlier bounds Ω(log⁡n)≤g(n)≤O(n)\Omega(\log n)\le g(n)\le O(\sqrt n) of Brown, Erdős and Freedman.

Proof pointer

Section 2 (pp. 282--286). The upper bound (p. 282) removes from {1,…,n}\{1,\ldots,n\} a union of short intervals placed like a discrete Cantor set, keeping a set SS with ∣S∣≥n/2|S|\ge n/2 in which every ascending wave has length less than 8log⁡2n+4log⁡n=(8+o(1))log⁡2n8\log^2n+4\log n=(8+o(1))\log^2n. The lower bound (pp. 282--286) works with the gaps of a set SS with ∣S∣=n/2|S|=n/2, sorted by dyadic length, and builds a wave greedily by appending short waves of 10−5log⁡n/log⁡log⁡n10^{-5}\log n/\log\log n terms each.

  • Section 3 (p. 286) says the upper-bound construction adapts to show that the bound clog⁡nc\log n implied by Brown, Erdős and Freedman for sets of at least nαn^\alpha elements, 0<α<10<\alpha<1, is sharp: some A⊂{1,…,n}A\subset\{1,\ldots,n\} with ∣A∣≥nα|A|\ge n^\alpha has no ascending wave of length greater than d(α)log⁡nd(\alpha)\log n.
  • After imprecise remarks on removing the log⁡log⁡n\log\log n factor, the paper closes (p. 287) with the conjecture g(n)=Θ(log⁡2n)g(n)=\Theta(\log^2n).

Read depth

Claims checked: the definition of gg, Theorem 2.1, the abstract's form of it and the remarks of Section 3 were read clause by clause on the page images of the print; the proofs of Section 2 were read for structure only. Nothing here is independently reviewed.

Dependencies

None in the corpus.

Source. N. Alon and J. Spencer, Ascending waves, J. Combin. Theory Ser. A 52 (1989), no. 2, 275--287, doi:10.1016/0097-3165(89)90033-2; the edition read is named on the source card.

Bears on

No Erdős problem of the corpus states this density question.