Long shortest vectors in low dimensional lattices

and relate such sets to lattices of a particular form. As a main result, we present an infinite family of such lattices with the property that the normalised norm of the shortest vector of each lattice converges to the square root of the Hermite constant

g a m m a_{3}

. We obtain a similar result for the generalisation of our construction to

4

and

5

dimensions.

Review text

The author considers points sets of the form \[ \Pi_{N,v} = \{ na\pmod N,\ nb \pmod{N},\ nc \pmod{N} \mid 0 \leq n < N \} \] where \(\mathbf{v} = (a,b,c)\) with \(0 < a < b < c <N\) and \(\gcd(N,a)=1\). This definition ensures that there are \(N\) different points in the unit cube \([0,N-1]^3.\) The shortest distance between points in \(\Pi_{N,v}\) is the minimum of all distinct \(\mathbf{x}, \mathbf{y} \in \Pi_{N,v}\) of \(||\mathbf{x}-\mathbf{y}||\). The author studies which parameters generate point sets with long shortest distances, relates them to certain lattices, and presents an infinite family of such lattices such that the normalized norm of the shortest vector of each lattices converges to the square root of the Hermite constant \(\gamma_3\). The construction is then generalized to four and five dimensions.

Recommendations

Cites work

Cited in

(7)

Describes a project that uses

Uses Software

This page was built for publication: Long shortest vectors in low dimensional lattices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2005735)