Using block designs in crossing number bounds

From MaRDI portal
Publication:5204969

DOI10.1002/JCD.21665zbMATH Open1427.05060arXiv1807.03430OpenAlexW2963894154MaRDI QIDQ5204969FDOQ5204969


Authors: John Asplund, Gregory J. Clark, Garner Cochran, Éva Czabarka, Arran Hamm, Gwen Spencer, László A. Székely, Libby Taylor, Zhiyu Wang Edit this on Wikidata


Publication date: 10 December 2019

Published in: Journal of Combinatorial Designs (Search for Journal in Brave)

Abstract: The crossing number mboxcr(G) of a graph G=(V,E) is the smallest number of edge crossings over all drawings of G in the plane. For any kge1, the k-planar crossing number of G, mboxcrk(G), is defined as the minimum of mboxcr(G1)+mboxcr(G2)+ldots+mboxcr(Gk) over all graphs G1,G2,ldots,Gk with cupi=1kGi=G. Pach et al. [emph{Computational Geometry: Theory and Applications} {�f 68} 2--6, (2018)] showed that for every kge1, we have mboxcrk(G)leleft(frac2k2frac1k3ight)mboxcr(G) and that this bound does not remain true if we replace the constant frac2k2frac1k3 by any number smaller than frac1k2. We improve the upper bound to frac1k2(1+o(1)) as kightarrowinfty. For the class of bipartite graphs, we show that the best constant is exactly frac1k2 for every k. The results extend to the rectilinear variant of the k-planar crossing number.


Full work available at URL: https://arxiv.org/abs/1807.03430




Recommendations





Cited In (4)





This page was built for publication: Using block designs in crossing number bounds

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5204969)