Random runners are very lonely

TL;DR

This paper proves that for random sets of runners on a circular track, with high probability, each runner will be at least 1/2 minus epsilon away from all others, using Fourier analysis.

Contribution

It establishes a probabilistic version of the Lonely Runner Conjecture with a stronger separation bound for random configurations.

Findings

With high probability, runners are at least 1/2 - epsilon apart.

Fourier analysis is used to prove the probabilistic bound.

Implications for coloring random integer distance graphs.

Abstract

Suppose that $k$ runners having different constant speeds run laps on a circular track of unit length. The Lonely Runner Conjecture states that, sooner or later, any given runner will be at distance at least $1/k$ from all the other runners. We prove that, with probability tending to one, a much stronger statement holds for random sets in which the bound $1/k$ is replaced by \thinspace $1/2-\varepsilon $. The proof uses Fourier analytic methods. We also point out some consequences of our result for colouring of random integer distance graphs.