On sparse universally rigid graphs in $\mathbb{R}$

Abstract

A graph is said to be 1-universally rigid if, for any generic realization of its vertices on the real line, the realization cannot be deformed out of the line in any higher-dimensional Euclidean space while preserving all edge lengths. In this paper, we investigate the sparsity of 1-universally rigid graphs in different manners. First, we show that an Erdős-Rényi random graph $G(n,p)$ is 1-universally rigid with high probability whenever $p \geq C\log n \cdot \log\log n/n$ for some sufficiently large constant $C>0$. These techniques also imply the existence of 1-universally rigid graphs of arbitrarily large girth, which resolves a problem suggested by Dantas e Moura, Jordán, and Silverman in a non-constructive manner. We also study how sparse a 1-universally rigid graph can be, proving that there exists an absolute constant $\varepsilon>0$ such that every sufficiently large 1-universally rigid graph on $n$ vertices has at least $(3/2+\varepsilon)n$ edges. This breaks the natural $3/2$ barrier arising from degree-2 extensions and makes progress towards a question of Jordán and Nguyen.