Generalised Erd\H{o}s distance theory on graphs

TL;DR

This paper introduces a unified framework for generalized distance problems on graphs using $g$-rigidity, providing sharp bounds for the number of distinct realizations across various metrics and applications.

Contribution

It develops a general theory connecting graph rigidity with diverse distance problems, offering new bounds and insights for multiple metric spaces and tensor completion.

Findings

Sharp lower bounds for the number of $g$-distinct realizations.

Results applicable to pseudo-Euclidean, $ ext{l}_p$, and matrix completion problems.

A proof linking the unit distance conjecture to the pinned distance conjecture.

Abstract

The famous Erd\H{o}s distinct distances problem asks the following: how many distinct distances must exist between a set of $n$ points in the plane? There are many generalisations of this question that ask one to consider different spaces and metrics, or larger structures of points. We bring these problems into a common framework using the concept of $g$-rigidity. Specifically, if $G=(V,E)$ is a (hyper)graph, $g$ is a map assigning polynomial measurements to the edges of $G$ and $f_{g,G}(P^V)$ gives the set of $g$-distinct realisations of the $g$-rigid graph $G$, where vertices must lie in a point set $P$, our main results describe sharp lower bounds for the size of $\big|f_{g,G}(P^V)\big|$. This allows us to obtain results for pseudo-Euclidean metrics, $\ell_p$ metrics, dot-product problems, matrix completion problems, and symmetric tensor completion problems. In addition, we use the recent work of Alon, Buci\'c and Sauermann along with a simple colouring argument to prove that the number of $\| \cdot\|$-distinct realisations of a graph $G=(V,E)$ within a $d$-dimensional point set $P$ is at least $\Omega\left(\frac{|P|^{|V|-1}}{(\log |P|)^2} \right)$ for almost all $d$-norms. Our methods here also provide a short proof that the unit distance conjecture implies the pinned distance conjecture.