Earth-Moon Problem
Status partial high confidence
The Earth–Moon problem—finding the maximum chromatic number of a biplanar graph (union of two planar graphs)—remains open with bounds $9 \leq \chi \leq 12$ unchanged since the problem was posted. In 2023, Eppstein disproved Gethner's conjecture that 2-blowups of planar graphs are always biplanar, using iterated Kleetopes as counterexamples; and Kirchweger, Scheucher, and Szeider used SAT-based methods to show that at least one of Gethner's proposed 10-chromatic candidate graphs is not biplanar.
Cited literature (2)
-
Disproves Gethner's conjecture that 2-blowups of planar graphs are always biplanar by constructing iterated Kleetopes whose 2-blowups are not biplanar, while also giving positive biplanarity results for certain graph classes.
-
Applies SAT encodings with dynamic symmetry breaking to test biplanarity, and shows that a graph derived from $C_5 \boxtimes K_4$ with one vertex removed (one of Gethner's candidate 10-chromatic biplanar graphs) is not biplanar.
Reviewer notes. Gethner's 2018 survey 'To the Moon and Beyond' (Springer book chapter in 'Graph Theory: Favorite Conjectures and Open Problems – 2', edited by Gera, Haynes, Hedetniemi) introduced the conjecture that the answer is 11 and proposed $C_7 \boxtimes K_4$ and a vertex deletion from $C_5 \boxtimes K_4$ as candidate 10-chromatic biplanar graphs; the Springer page was inaccessible due to authentication redirect so this paper is not in since_posted. The fundamental bounds ($9 \leq \chi \leq 12$) are unchanged. Eppstein's paper won Best Paper at GD 2023 and was subsequently published in the Journal of Graph Algorithms and Applications (2024). The DROPS page for Kirchweger et al. confirmed the venue and DOI but the full text (PDF) was unreadable by WebFetch; the specific Earth-Moon result attributed to that paper (eliminating one of Gethner's candidates) is reported in Wikipedia's Earth–Moon problem article.
Discussion
In term of graphs, it can be rephrased as follows. What is the maximum chromatic number of a graph $ G $ which is the union of two planar graphs (on the same vertex set)? If a graph $ G $ on $ n $ vertices is the union of two planar graphs, then it has at most $ 2(3n-6) $ edges, and so it is has a vertex of degree at most $ 11 $ . An easy induction shows that $ G $ is 12-colourable, as observed by Heawood [He]. Gardner [G] reported an example requiring 9 colours (the join of $ C_5 $ and $ K_6 $ ). It is not known if configurations exist requiring 10, 11, or 12 colours. More generally, one may ask for the maximum chromatic number of the union of $ k $ planar graphs. Problem What is the maximum chromatic number of a graph $ G $ which is the union of $ k $ planar graphs? The same reasoning as above shows that $ 6k $ colours are always sufficient. The minimum number of planar graphs to decompose a complete graph [BW] gives a lower bound of $ 6k-2 $ for $ k \ne 2 $ .
Bibliography
-
[BW]L. W. Beineke and A. T. White, Topological Graph Theory, Selected Topics in Graph Theory, L. W. Beineke and R. J. Wilson, eds., Academic Press, 15-50, 1978. -
[G]M. Gardner, Mathematical Recreations: The Coloring of Unusual Maps Leads Into Uncharted Territory. Sci. Amer. 242, 14-22, 1980. -
[He]P.J. Heawood, Map Colour Theorems. Quart. J. Pure Appl. Math. 24, 332-338, 1890. -
★
[R]G. Ringel, Färbungsprobleme auf Flachen und Graphen. Berlin: Deutsche Verlag der Wissenschaften, 1959.