Wiki
Wiki

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

Updated


Statement

Setting (pp. 1--5). Un\mathbb U_n, M(F)M(F), m(F)m(F), W(F)W(F) and MnM_n, mnm_n, WnW_n are as on the page for the conjecture on p. 4. For even nn, the skew-symmetric polynomials are the F∈UnF\in\mathbb U_n with F(z)=±znF(−1/z)F(z)=\pm z^nF(-1/z), where the sign has to be (−1)n/2(-1)^{n/2} (p. 4); they have n/2+1n/2+1 free coefficients (p. 5). Mn∗M_n^*, mn∗m_n^* and Wn∗W_n^* are defined as MnM_n, mnm_n and WnW_n but over skew-symmetric polynomials of degree nn only (p. 5).

Conjecture (p. 5). As n→∞n\to\infty through even values, Mn∗→MM_n^*\to M, mn∗→mm_n^*\to m and Wn∗→WW_n^*\to W, where MM, mm, WW are the limits of MnM_n, mnm_n, WnW_n conjectured on p. 4.

The paper presents this as suggested by Fig. 1 (p. 3), where for even 10≤n≤5010\le n\le50 the skew-symmetric optimum usually coincides with the unrestricted one and otherwise usually differs only slightly; the largest exception it reports is W24=0.8344W_{24}=0.8344 against 0.95280.9528 for the best skew-symmetric polynomial (pp. 4--5).

Scope

A conjecture, not a theorem. Since Mn≤Mn∗M_n\le M_n^* and mn≥mn∗m_n\ge m_n^* by definition, the skew-symmetric values bound the unrestricted extremes from one side only.

Read depth

Claims checked: the definitions and the conjecture were read on the page images of the print. Nothing here is independently reviewed.

Dependencies

Conjecture, p. 4, which supplies the limits named here.

Source. Andrew Odlyzko, "Search for Ultraflat Polynomials with Plus and Minus One Coefficients," in Connections in Discrete Mathematics, pp. 39--55, Cambridge University Press, 2018, doi:10.1017/9781316650295.004; the version read, the author's revised version of 18 May 2017, and its page numbering are named on the source card.

Bears on

  • Problem 1150: the skew-symmetric minima Mn∗M_n^* are upper bounds for MnM_n, so they cannot give the lower bound the problem asks for; this conjecture bears on the problem only together with the conjecture on p. 4.