On the distribution of lengths of short vectors in a random lattice (Q269911)

The author is interested in the investigation of the statistics of short vectors of a random lattice. In [\textit{A. Södergren}, Math. Z. 269, No. 3--4, 945--954 (2011; Zbl 1257.60002)], the following assertion was proved: for a lattice \(L\) in \(\mathbb R^n\) and \(t\geq 0\), let \(\widetilde N_t^n(L)\) be the number of nonzero vectors (up to sign) of \(L\) in a ball of volume \(t\). Taking \(L\) to be a random lattice, one may view \({\widetilde N}_t^n\) as a stochastic process on the positive real line \(\{t\in\mathbb R:t\geq 0\}\). As \(n\to \infty\), \({\widetilde N}_t^n\) weakly converges to a Poisson process on the positive real line with intensity \(1/2\). The joint distribution of the angles and the lengths of the first \(N\) shortest vectors of a random lattice was investigated in [\textit{A. Södergren}, J. Lond. Math. Soc., II. Ser. 84, No. 3, 749--764 (2011; Zbl 1268.60068)]. Using the main result of that paper (Theorem 5), the author is able to obtain the following estimate: Theorem. Let \(S\) be a Borel measurable set in \(\mathbb R^n\), which is symmetric at the origin with Euclidean measure \(V\). Fix \(\varepsilon>0\). Suppose that \(0\leq k\leq (n/2)^{1/2-\varepsilon}\) is a non-negative integer, possibly depending on \(n\), and suppose also that \(8V\leq\sqrt{n/2}-k\). Let \(P(S,k)\) be the probability that an \(n\)-dimensional random lattice has at most \(k\) nonzero vectors (up to sign) in \(S\). Then, \(P(S,k)\) is close to \(P_{V/2}(k)\) for \(n\) sufficiently large, where \(P_{V/2}\) is the (left) distribution function of the Poisson distribution with mean \(V/2\).