2
$\begingroup$

Let $P$ be a poset, or partially ordered set. Let $\le$ denote the reflexive order on $P$, and $<$ the corresponding irreflexive order. Let the phrase "a maximal pair" in $P$ refer to an ordered pair $(u,v)$ of elements in $P$ such that

  • $u>v$,
  • $u$ is a maximal element of $P$, and
  • $v$ is a maximal element of the subposet $P\setminus\{u\}$.

We say a subposet $P'$ of $P$ is obtained by elementary collapse from $P$, if $P' = P \setminus \{u,v\}$ for some maximal pair $(u,v)$ of $P$. We say $P$ is collapsible if there is a sequence $$ P = P_0 \supset P_1 \supset \cdots \supset P_{k-1} \supset P_k = \left\{\ast\right\} $$ of subposets of $P$, in which $P_i$ is obtained by elementary collapse from $P_{i-1}$ for $i=1,2,\ldots,k$.

My questions are:

  • I needed to prepare by myself the terminology for maximal pairs, elementary collapse, and collapsiblity; is there any better, more common way to refer to them?
  • I am working on proving collapsibility for a class of posets, conjecturally being the face posets of some convex polytopes. Are there any general ways to prove collapsibility of a poset, assuming or not assuming realizations as convex polytopes?
$\endgroup$
5
  • $\begingroup$ So you have a finite partial order, and you just remove maximal elements one by one (each time maximal in the "collapsed" order) until you've cleared up the entire thing. What am I missing? Also what does $v$ matter in the choice of collapsing the order? You just remove $u$ which is a maximal element, and that's it. $\endgroup$
    – Asaf Karagila
    Commented May 7, 2016 at 9:36
  • $\begingroup$ Note that: $P'=P\setminus\left\{u,v\right\}$. And $v$ must be maximal in $P\setminus \left\{ u \right\} $. So $\left\{x > y < z\right\}$ (in which x and z are incomparable) is not collapsible. $\endgroup$
    – gksato
    Commented May 7, 2016 at 9:42
  • $\begingroup$ Is the poset $P=\{0,1\}$ ordered by $0<1$ collapsible? As I understand $(1,0)$ is a maximal pair, but $P \setminus\{1,0\} = \emptyset \neq \{*\}$ as you demand. - If I understand correctly, one simple criterion for $P$ to be collapsible is that it has an odd number of points in the base set. $\endgroup$ Commented May 10, 2016 at 12:45
  • $\begingroup$ No. It is NOT collapsible. and yes,that's very important necessary condition. $\endgroup$
    – gksato
    Commented May 10, 2016 at 14:55
  • $\begingroup$ In fact, in case of ranked posets (=all maximal chains has the same finite length), $P$ must have Euler characteristic 1. $\endgroup$
    – gksato
    Commented May 10, 2016 at 22:06

0

You must log in to answer this question.

Browse other questions tagged .