On the minimum of a positive definite quadratic form over non-zero lattice points. Theory and applications (Q1782001)

Let \(\Sigma^{++}_{d}\) (\(d\geqslant 2\)) be the set of positive definite matrices with determinant \(1\). Denote by \(\mathcal{SC}\) the class of probability measures arising either in Cholesky or spectral decomposition of an element of \(\Sigma^{++}_{d}\). Let \(\Sigma\in\Sigma^{++}_{d}\). Denote by \(M_d(\Sigma)=\min\{^{t}a\cdot \Sigma\cdot a\mid a\in\mathbb{Z}^{d}\setminus\{0\}\} \). In this paper, the following problem is studied: for a given probability measure \(\mu\) of the class \(\mathcal{SC}\), estimate the probability \(\mu(M_d(\Sigma))\leqslant\delta\) as a function of \(\delta >0\). The authors prove sharp upper and lower bounds of the probability \(\mu(M_d(\Sigma)) \leqslant \delta\). Recall that Kleinbock and Margulis obtained a similar result for the locally symmetric space \(\mathrm{SL}_{d}(\mathbb{Z})\setminus \mathrm{SL}_{d}(\mathbb{R})/\mathrm{SO}_{d}(\mathbb{R})\) equipped with the probability measure induced by the Haar measure on \(\mathrm{SL}_{d}(\mathbb{R})\) (see [\textit{D. Y. Kleinbock} and \textit{G. A. Margulis}, Invent. Math. 138, No. 3, 451--494 (1999; Zbl 0934.22016), erratum ibid. 211, No. 2, 855--862 (2018; Zbl 1404.22021)]). The main results are applied to information theory.