Crossing number inequalities for curves on surfaces
Alfredo Hubard, Hugo Parlier
TL;DR
The paper addresses crossing numbers for drawings of non-homotopic curves on orientable surfaces, proving a lower bound of cr$(\Gamma) \gtrsim (m \log m)^2$ for a family of $m$ distinct curves and extending the result to general surfaces with a bound depending on $|\chi|$. The authors leverage the Koebe–Andreev–Thurston uniformization to work in a hyperbolic metric, apply circle packing to realize a length/area framework, and derive length-based lower bounds via averaging to obtain the main crossing inequality. They then translate these curve results into corollaries for arcs and graphs (via arc-to-curve translations) and provide explicit upper bounds on the size of arc/curve families with bounded pairwise intersections. Finally, they demonstrate near-optimality of the bounds through combinatorial pants models and explicit pants-based constructions, yielding tight growth rates up to constants on both arcs and curves across surfaces. These results settle questions of Pach–Tardos–Toth about crossing numbers for non-homotopic edges and establish optimal growth rates on orientable surfaces, enriching the understanding of topological graph drawings.
Abstract
We prove that, as $m$ grows, any family of $m$ homotopically distinct closed curves on a surface induces a number of crossings that grows at least like $(m \log m)^2$. We use this to answer two questions of Pach, Tardos and Toth related to crossing numbers of drawings of multigraphs where edges are required to be non-homotopic. Furthermore, we generalize these results, obtaining effective bounds with optimal growth rates on every orientable surface.
