Distinct Directions and Distinct Distances in $\mathbb{R}^d$
Noga Alon, Rom Pinchasi
TL;DR
This work addresses how many distinct directions and distances a $d$-dimensional point set in $\R^d$ can determine. It combines a high-dimensional extension of Ungar's theorem with a bichromatic, generalized-segment framework and central projections, leveraging a DS&Wigderson bound to prove a linear-in-$n$ bound: there exists $b \,\ge\, \frac{1}{48}$ such that a $d$-dimensional set not contained in a hyperplane determines at least $b d n$ lines with pairwise distinct directions. Furthermore, for typical $d$-norms, the paper shows the existence of $(b d - o(1)) n$ distinct distances for large $n$, aligning with a broader conjecture on dimensional scaling. The results illuminate a linear-growth regime in high dimensions and contrast with the Euclidean case, suggesting Theta$(n d)$ behavior for typical norms and highlighting connections to Dirac-type problems in discrete geometry.
Abstract
We show that there exists an absolute positive constant $b (\geq \frac{1}{48})$ so that any set of $n$ points in $\mathbb{R}^d$ that is $d$-dimensional determines at least $bdn$ lines with pairwise distinct directions. As a consequence we prove that there are $d$-dimensional real norms $\|\cdot\|$ so that every set of $n>n_0(d)$ points that is $d$-dimensional determines at least $(bd-o(1))n$ distinct distances with respect to $\|\cdot \|$.
