3
$\begingroup$

An $n$-dimensional Euclidean lattice $L$ is called of $\textbf{Voronoi's first kind}$ if it satisfies the $\textbf{obtuse superbasis}$ condition: There exist $x_0, \dots, x_n \in L$ such that

$1) x_1, \dots, x_n$ is a basis of $L$

$2) x_i \cdot x_j \leq 0$, unless $i=j$

$3) \sum_{i=0}^n x_i = 0$

Now I am reading this paper: The spine which was no spine.

The main result is the following: It uses the identification of the space $S_n ‎ =  SO_n(\mathbb{R})$\ $SL_n(\mathbb{R})$ of (equivalence classes of) full-rank $n$-dimensional volume $1$ lattices up to rotation with the Teichmüller space $T_n$ of flat metrics of unit volume on the torus $\mathbb{T}^n = \mathbb{R}^n/\mathbb{Z}^n$.

Moreover, it identifies:

a)The $\textbf{systole} \operatorname{syst}(\rho)$ of a point $\rho \in T_n$, which is the length of the shortest homotopically essential geodesic in the flat torus $(\mathbb{T}^n, \rho)$, with the function $f:S_n \to (0, \infty), f(A) = \min_{v \in \mathbb{Z}^n, v\neq 0}|Av|$ which gives the length of the shortest vector in $A$.

b)The set $S(\rho)$ of homotopy classes of geodesics in $(\mathbb{T}^n, \rho)$ with length $\operatorname{syst}(\rho)$ with the set of shortest vectors of an element $A \in S_n$.

c)The set $X \subseteq T_n$ consisting of those points $\rho$ with the property that $S(\rho)$ generates a finite index subgroup of $\pi_1(\mathbb{T}^n)$ with the set of well-rounded lattices in $S_n$, that is, those lattices whose set of shortest vectors is a basis for $\mathbb{R}^n$.

to prove that $X$ is a minimal contractible deformation retraction of $S_n$ (that is, there does not exist a strict subset $Y \subset X$ with the above properties).

Now, I am a bit new in lattice theory in general, and especially in identifying lattice concepts with their respective diff. geometric counterparts, so is there a way to translate lattices of Voronoi's first kind into the diff. geometric context described above?

$\endgroup$
2
  • 1
    $\begingroup$ While I don't understand diff geom either (hence why this is a comment), it is worth mentioning that a standard consequence of being VFK is that any relevant vector may be written as $\sum_{i\in I}x_i$, where $\emptyset\subsetneq I\subsetneq\{1,\dots,n+1\}$. See Theorem 3 of Conway and Sloane's Low dimensional lattices VI. This then implies that the definition of $\mathsf{syst}$ can minimize over a finite set (rather than infinite), among other things. It is also worth mentioning though that being VFK is a basis-dependent property, and determining $\endgroup$ Commented Mar 20 at 3:27
  • 1
    $\begingroup$ if a lattice is VFK is thought to be computationally hard. See for example section 7 of Finding a Closest Point in a Lattice of Voronoi's First Kind. As the diff geom component of your question seems basis independent, this could be an indication that VFK lattices don't fit into the picture that well (but I don't know). $\endgroup$ Commented Mar 20 at 3:29

0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.