Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Statement
Setting. Cross-IU is defined on the Theorem 9 page: every member of one family meets every member of the other, and no such pair has union .
Theorem 10 (p. 3). If are pairwise cross-IU, then
The paper adds (p. 3, quoted) that "the equality holds only if and [sic] for some or and ", the index standing where is evidently meant. It compares the result with Hilton's theorem for pairwise intersecting families of -sets.
Theorem 12 (p. 7). If are cross-IU and , then
Corollary 1 (p. 6). If are pairwise cross-IU, and , then
with strict inequality unless .
Lemma 1 (p. 6). If and , then (12).
The paper says Theorem 10 follows at once from Corollary 1 and Theorem 12. For this is Corollary 1, after ordering the families by size. For , ordering the families by size, Theorem 12 applied to the two largest gives (the step spelled out on this page; is trivial).
Proof pointer
Corollary 1, pp. 6--7: with , (7) of Theorem 9 gives for , and Lemma 1 gives when ; adding single sets to the families handles the equality case. Theorem 12, p. 7: with , , (7) gives , which settles ; for with , Harris–Kleitman applied to the generated up-sets and down-sets gives .
Read depth
Claims checked: Theorem 10 with its equality remark, Lemma 1, Corollary 1 and Theorem 12 were read clause by clause on the print, and their proofs on pp. 6--7 were followed for their structure and not checked line by line. Nothing here is independently reviewed.
Dependencies
Theorem 9; external input named by the paper: the Harris–Kleitman inequality (its Theorem 2, p. 2).
Source. P. Frankl and A. Kupavskii, Perfect matchings in down-sets, Discrete Math. 346 (2023), Paper No. 113323, DOI 10.1016/j.disc.2023.113323; read in arXiv:2201.03865v1, Theorem 10 on p. 3, Lemma 1 and Corollary 1 on p. 6, Theorem 12 on p. 7, with proofs on pp. 6--7. The edition is identified on the source card.
Bears on
No Erdős problem in the corpus is linked to these bounds.