An extremal problem for polynomials (Q5920280): Difference between revisions

From MaRDI portal
Set OpenAlex properties.
ReferenceBot (talk | contribs)
Changed an Item
 
Property / cites work
 
Property / cites work: Computational excursions in analysis and number theory / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q4459960 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Some unsolved problems / rank
 
Normal rank
Property / cites work
 
Property / cites work: An inequality for the maximum of trigonometric polynomials / rank
 
Normal rank
Property / cites work
 
Property / cites work: Sur Les Polynomes a Coefficients Unimodulaires / rank
 
Normal rank
Property / cites work
 
Property / cites work: Generalized bounds on the crest-factor distribution of OFDM signals with applications to code design / rank
 
Normal rank
Property / cites work
 
Property / cites work: On Polynomials ∑±nzm,∑eαminzm,z=e0i / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q5577283 / rank
 
Normal rank

Latest revision as of 10:54, 7 July 2024

scientific article; zbMATH DE number 6265630
Language Label Description Also known as
English
An extremal problem for polynomials
scientific article; zbMATH DE number 6265630

    Statements

    An extremal problem for polynomials (English)
    0 references
    0 references
    5 March 2014
    0 references
    The author considers the following special extremal problem for polynomials: how small can the sup-norm of a polynomial over the unit circle be if each of the coefficients is \(\pm 1\) times a fixed complex number with absolute value \(1\)? More precisely, let \[ B_n(\phi_0,\ldots,\phi_n):=\max\{\| \sum_{k=0}^{n-1} \epsilon_k e^{2\pi i \phi_k}z^k \|: \epsilon_0,\ldots,\epsilon_{n-1}\in \{-1,1\}\} \] where \(\phi_0,\ldots,\phi_{n-1}\in [0,1]\) are given and \(\|.\|\) denotes the sup norm over the unit circle. Furthermore, let \[ b(n):=\min\{B_n(\phi_0,\ldots,\phi_n): \phi_0,\ldots,\phi_{n-1}\in [0,1]\}. \] The author shows that \(b(n)/n\rightarrow 2/\pi\) in Theorem 1 by taking real parts and using cosine polynomials and the equidistribution of \(\{\alpha k^2\}_k\) where \(\{.\}\) denotes the fractional part and \(k\) runs over the positive integers. This result gives an answer in a special case to a general problem-class raised by Littlewood, Borwein, Erdélyi and others.
    0 references
    0 references
    polynomials
    0 references
    extremal problems
    0 references
    0 references