Wiki
Wiki

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

Updated

Ascending waves

../

theorem_1_1: Alon and Spencer's theorem that the least f(k) for which every 2-coloring of 1 to f(k) has a monochromatic ascending wave of length k satisfies c_1 k^3 <= f(k) <= c_2 k^3 for all k >= 1, so the lower bound k^2 - k + 1 of Brown, Erdős and Freedman is not the exact value.

theorem_2_1: Alon and Spencer's theorem that the largest g(n) such that every subset of 1 to n with at least n/2 elements contains an ascending wave of length g(n) satisfies Omega(log^2 n / log log n) <= g(n) <= O(log^2 n).


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 file prints "Copyright © 1989 by Academic Press, Inc. All rights of reproduction in any form reserved." in the footer of its first page (printed p. 275), every other right reserved.

Source: https://web.math.princeton.edu/~nalon/PDFS/publications.html.

An ascending wave of length kk is a sequence of integers x1<⋯<xkx_1<\cdots<x_k whose consecutive differences never decrease (p. 275). The paper has two main results, both stated on p. 276. Theorem 1.1 gives c1k3≤f(k)≤c2k3c_1k^3\le f(k)\le c_2k^3 for all k≥1k\ge1, where f(k)f(k) is the least integer such that every 2-coloring of {1,…,f(k)}\{1,\ldots,f(k)\} has a monochromatic ascending wave of length kk; the lower bound, proved by a random block coloring in Section 1 (pp. 276--282), shows that the Brown--Erdős--Freedman lower bound k2−k+1k^2-k+1 is not the exact value. Theorem 2.1 gives Ω(log⁡2n/log⁡log⁡n)≤g(n)≤O(log⁡2n)\Omega(\log^2n/\log\log n)\le g(n)\le O(\log^2n) for the largest length g(n)g(n) of an ascending wave guaranteed in every subset of {1,…,n}\{1,\ldots,n\} with at least n/2n/2 elements, proved in Section 2 (pp. 282--286). Section 3 (pp. 286--287) states a real-interval form of the Theorem 1.1 construction, a sharpness remark for sets of nαn^\alpha elements, and the conjecture g(n)=Θ(log⁡2n)g(n)=\Theta(\log^2n) (p. 287).

Read status: claims checked for Theorems 1.1 and 2.1, their definitions and the remarks of Section 3, read clause by clause on the page images; the proofs read for structure only. Nothing here is independently reviewed. Result pages: theorem_1_1 and theorem_2_1.

Bears on. #781: Theorem 1.1 (p. 276) gives constants c1,c2>0c_1,c_2>0 with c1k3≤f(k)≤c2k3c_1k^3\le f(k)\le c_2k^3 for all k≥1k\ge1, and the paper presents it as showing false the question of Brown, Erdős and Freedman whether f(k)=k2−k+1f(k)=k^2-k+1 for all kk, which is the problem's particular question. The paper's waves have non-decreasing differences and the problem's descending waves non-increasing ones; reversing {1,…,n}\{1,\ldots,n\} by x↦n+1−xx\mapsto n+1-x exchanges the two, a step the paper does not write out.

Results.

  • Theorem 1.1 (p. 276): Ω(k3)≤f(k)≤O(k3)\Omega(k^3)\le f(k)\le O(k^3) for the two-color ascending-wave number.
  • Theorem 2.1 (p. 276): Ω(log⁡2n/log⁡log⁡n)≤g(n)≤O(log⁡2n)\Omega(\log^2n/\log\log n)\le g(n)\le O(\log^2n) for ascending waves in subsets of {1,…,n}\{1,\ldots,n\} of at least n/2n/2 elements.

No file of this source is held: no license on record permits its redistribution, and the card cites the edition it names above.