4
$\begingroup$

Let $S$ be a finite set, $I$ a finite index set, $\mathcal A=(A_i:i\in I)$ and $\mathcal B=(B_i:i\in I)$ families of subsets of $S$.

For $J\subseteq I$, let $A(J)$ denote $\bigcup_{j\in J} A_j$.

A partial transversal of $\mathcal A$ is the image of an injective map $f:J\to S$ from a subset $J$ of $I$ such that $f(i)\in A_i$ for all $i\in J$. We have a transversal if $J=I$. A common partial transversal of $\mathcal A$ and $\mathcal B$ is a subset that is a partial transversal of $\mathcal A$ and $\mathcal B$; it is a common transversal if the subset is a transversal of $\mathcal A$ and $\mathcal B$.

The following is the finite version of Theorem 7 in the article by J. de Sousa (also publishing as Joan Davies), "Disjoint Common Transversals," Combinatorial Mathematics and its Applications: Proceedings of a Conference held at the Mathematical Institute, Oxford, from 7-10 July, 1969 (D.J.A. Welsh, ed.).

If $m\ge1$, then $\mathcal A$ and $\mathcal B$ have $m$ pairwise disjoint common transversals if and only if, for all $J,K\subseteq I$, $|A(J)\cap B(K)|\ge m(|J|+|K|-|I|)$.

For $\\\$$123 (I am not Lavinia Clay! and please see the caveat below), can someone provide necessary and sufficient conditions for $\mathcal A$ and $\mathcal B$ to have $m\in\mathbb N$ pairwise disjoint common partial transversals of cardinalities $k_1$, $k_2$, $\dots$, $k_m\in\mathbb N$, under the assumption that the sets in $\mathcal A$ are pairwise disjoint, and likewise for $\mathcal B$?

The caveat is that this offer must be legal and allowed by the moderators; there must be some reasonable way for me to pay you (e.g., you are not in the Witness Protection Program or North Korea); I am the sole arbiter of what is an acceptable solution; there can be only one winner; I reserve the right to change my mind in case I missed some condition; and I place a time limit of 3 months from today.

See Theorem 2.2 of D. R. Fulkerson, "Disjoint Common Partial Transversals of Two Families of Sets," RAND Memorandum, Document RM-6102-PR (1969).

$\endgroup$
6
  • 1
    $\begingroup$ SE meta discussion on whether or not it is appropriate to offer cash for an answer: meta.stackexchange.com/q/25615/260547 (the downvotes suggest "no") $\endgroup$ Commented Jun 2 at 19:08
  • $\begingroup$ @CarloBeenakker Thanks. I wasn't asking whether it was appropriate (Lavinia Clay and Paul Erdos say it is), only whether the moderators of this site would allow it. $\endgroup$
    – Tri
    Commented Jun 3 at 12:10
  • $\begingroup$ @Stefan Kohl Well, you're a moderator, so I'll go with your change, but I've asked a variant of this question before without getting answers. The cash prize might induce people to look at it, whereas history shows that with no cash prize, probably no one will try to answer it. Please restore the cash offer to the title. $\endgroup$
    – Tri
    Commented Jun 3 at 12:12
  • $\begingroup$ on the issue whether the title should mention the cash offer, see web.archive.org/web/20150910105257/https://tea.mathoverflow.net/… [search for "title"] $\endgroup$ Commented Jun 3 at 12:45
  • $\begingroup$ @CarloBeenakker I knew him when he was in grad school. I wrote one of my papers because of a bet I made with him (or maybe it was a dare).... $\endgroup$
    – Tri
    Commented Jun 8 at 5:46

0

You must log in to answer this question.

Browse other questions tagged .