Rogers' mean value theorem for \(S\)-arithmetic Siegel transforms and applications to the geometry of numbers (Q2161327)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Rogers' mean value theorem for \(S\)-arithmetic Siegel transforms and applications to the geometry of numbers
scientific article

    Statements

    Rogers' mean value theorem for \(S\)-arithmetic Siegel transforms and applications to the geometry of numbers (English)
    0 references
    0 references
    4 August 2022
    0 references
    The Siegel transform of a bounded and compactly supported function \(f : \mathbb R^d \to \mathbb R_{\geq0}\) is defined as \[ \tilde{f}(\Lambda) = \sum_{\boldsymbol v \in \Lambda \setminus \{\boldsymbol 0\}} f(\boldsymbol v),\ \text{ for all }\ \Lambda \in \operatorname{SL}_d(\mathbb R)/\operatorname{SL}_d(\mathbb Z). \] The classical Siegel's integral formula reads \[\int_{\operatorname{SL}_d(\mathbb R)/\operatorname{SL}_d(\mathbb Z)} \tilde{f}(g\mathbb Z^d) dg = \int_{\mathbb R^d} f(\boldsymbol v) d\boldsymbol v, \] where \(dg\) is the normalized \( \operatorname{SL}_d(\mathbb R)\)-invariant measure on the space \(\operatorname{SL}_d(\mathbb R)/\operatorname{SL}_d(\mathbb Z)\) of unimodular lattices in \(\mathbb R^d\) and \(d\boldsymbol v\) is the usual Lebesgue measure on \(\mathbb R^d\). For each \(1 \leq k \leq d-1\), there are also \(k\)-moment formulas for the Siegel transform produced by \textit{C. A. Rogers} in [Acta Math. 94, 249--287 (1955; Zbl 0065.28201)]. While there are known gaps in Rogers' original proof, they have since been fixed by \textit{W. M. Schmidt} [Monatsh. Math. 61, 269--276 (1957; Zbl 0078.03903)]. The paper under review is devoted to the higher moment formulas for Siegel transforms over the space of unimodular \(S\)-lattices in \(\mathbb Q_S^d\) for \(d \geq 3\). Specifically, let \(S\) be the union of the archimedean place \(\infty\) and a finite set of primes. The space \(\mathbb Q e_S\) is defined as the direct product product of the corresponding completions \(\mathbb Q_S = \prod_{v \in S} \mathbb Q_v\), so the corresponding ring of \(S\)-integers is \[ \mathbb Z_S = \left\{ (z,\dots,z) \in \mathbb Q_S : z \in \mathbb Q,\ |z|_v \leq 1\ \text{ for all }\ v \in S \right\}. \] Then an \(S\)-lattice is a free \(\mathbb Z_S\)-module \(\Lambda\) of rank \(d\) in \(\mathbb Q_S^d\) so that \(\mathbb Q_S^d/\Lambda\) is compact and thus has a finite \(\mathbb Q_S^d\)-invariant measure. There is then a natural (albeit a bit technical) way to describe the space of unimodular \(S\)-lattices in \(\mathbb Q_S^d\), analogous to the space \(\operatorname{SL}_d(\mathbb R)/\operatorname{SL}_d(\mathbb Z)\) of unimodular lattices in \(\mathbb R^d\). Now, for a bounded and compactly supported function \(f\) on \(\mathbb Q_S^d\) the Siegel transform on this space of unimodular \(S\)-lattices is given by \[ \tilde{f}(\Lambda) = \sum_{\boldsymbol v \in \Lambda \setminus \{\boldsymbol 0\}} f(\boldsymbol v). \] An analogue of Siegel's formula also holds in this setting, and the author's main results provide \(k\)-moment formulas for this generalized Siegel's transforms. As applications, the author obtains certain results related to random Gauss circle problem and effective Oppenheim conjecture for \(S\)-arithmetic quadratic forms.
    0 references
    0 references
    \(S\)-arithmetic geometry of numbers
    0 references
    Rogers' higher moment formulas
    0 references
    lattice counting problem
    0 references
    random Gauss circle problem
    0 references
    effective Oppenheim conjecture
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references