Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Statement
Theorem 3 (p. 2, quoted). "For all and there is a triangle-free -degenerate -uniform hypergraph with chromatic number ."
Here a hypergraph is -degenerate when every subhypergraph has a vertex lying in at most of its edges, and a triangle in an -uniform hypergraph is three edges whose union is a set of vertices (p. 1, abstract and footnote 1; p. 2); the full conventions are on the Lemma 4 page. The paper presents the theorem as showing that the greedy bound for -degenerate hypergraphs is tight for every , ruling out for the and bounds that its background Theorems 1 and 2 might suggest (p. 2).
Source. David R. Wood, Hypergraph Colouring and Degeneracy, arXiv:1310.2972v3, as identified on the source card: Theorem 3 on p. 2.
Read depth. Claims checked: the statement was read clause by clause on the print. Nothing here is independently reviewed.
Proof pointer
p. 2: the paper states Theorem 3 as a corollary of Lemma 4, whose hypergraph has the three properties; the lemma's extra conclusion on the sizes of colour classes is not needed.
Consequence for Problem 1022
This application is the corpus's; the paper does not mention the problem. Fix and take , . The theorem gives a finite -uniform hypergraph with , so does not have property B.
Let be any nonempty set. The edges of inside are the edges of the induced subhypergraph , . If is empty there are none. Otherwise is a subhypergraph, and so is every induced subhypergraph of it, so its vertices can be deleted one at a time, each lying in at most two remaining edges when deleted. Charge each edge to its first deleted vertex: each vertex is charged at most two edges, and the last vertex none, since every edge has vertices. Hence the number of edges inside is at most .
So meets the hypothesis of the corrected statement (every nonempty ) with every yet has no property B. Any constant for which that implication holds at a given therefore satisfies , and no sequence of such constants tends to infinity.
Bears on
- Problem 1022: by the consequence above, every constant for which the implication of the corrected statement (nonempty ) holds is below at every , so the corrected statement has the answer no.