The Möbius function is strongly orthogonal to nilsequences (Q410107): Difference between revisions

From MaRDI portal
Import240304020342 (talk | contribs)
Set profile property.
Set OpenAlex properties.
Property / OpenAlex ID
 
Property / OpenAlex ID: W2165859273 / rank
 
Normal rank

Revision as of 20:40, 19 March 2024

scientific article
Language Label Description Also known as
English
The Möbius function is strongly orthogonal to nilsequences
scientific article

    Statements

    The Möbius function is strongly orthogonal to nilsequences (English)
    0 references
    0 references
    0 references
    17 April 2012
    0 references
    This paper is a part of the Green-Tao program to establish asymptotic results for the number of solutions of systems of linear equations under the assumption that all variables are assigned prime values. A landmark paper in this direction was [Ann. Math. (2) 167, No. 2, 481--547 (2008; Zbl 1191.11025)] where the authors showed that the sequence of primes contains arbitrarily long arithmetic progressions. In [the authors, Ann. Math. (2) 171, No. 3, 1753--1850 (2010; Zbl 1242.11071)] they proved a general theorem including asymptotic formulas for the number of solutions, conditionally on the truth of two conjectures MN(s) and GI(s). Conjecture GI(s) was proved by the first author et al. [Ann. Math. (2) 176, No. 2, 1231--1372 (2012; Zbl 1282.11007)], while the present paper contains (as one application) the proof of conjecture MN(s) for \(s \geq 3\). Conjecture MN(2) had already been proved by the authors [Ann. Inst. Fourier 58, No. 6, 1863--1935 (2008; Zbl 1160.11017)]. As a further remark, the present paper has a companion paper [Zbl 1251.37012] which is longer and more technical, and which establishes several deep results that are used here (mainly quantitative equidistribution results for polynomial orbits on nilmanifolds). The main result of the paper is the following. Let \(G\) be a simply connected nilpotent Lie group which has a discrete and cocompact subgroup \(\Gamma\), meaning that \(G / \Gamma\) is a nilmanifold. Let \(g: \mathbb{Z} \mapsto G\) be a polynomial sequence, and let \(F:~G/\Gamma \mapsto \mathbb{R}\) be a Lipschitz function. Then the authors prove that \[ \left| N^{-1} \sum_{n=1}^N \mu(n) F(g(n) \Gamma) \right| \ll_{F,G,\Gamma,A} \log^{-A} N \] for all positive \(A\). This quantitative result goes beyond the previously known qualitative results from [\textit{J. Bourgain} et al., Dev. Math. 28, 67--83 (2013; Zbl 1336.37030)]. The proof uses a classical decomposition into ``Type I'' and ``Type II'' sums, a method going back to Vinogradov. Further applications of the methods developed in this paper are: a) an orthogonality result between the Möbius function and bracket polynomials, which is a somewhat natural supplement to the main theorem due to the close relation between bracket polynomials and nilmanifolds shown by \textit{V. Bergelson} and \textit{A. Leibman} [Acta Math. 198, No. 2, 155--230 (2007; Zbl 1137.37005)]. b) corresponding results when the Möbius function \(\mu\) is replaced by the Liouville function \(\lambda\). c) a recurrence/convergence theorem along the primes.
    0 references
    Möbius function
    0 references
    nilsequences
    0 references
    progressions in primes
    0 references
    nilmanifolds
    0 references

    Identifiers