2
$\begingroup$

Let $\mathcal A$ be a central hyperplane arrangement in $\mathbb R^d$ and let's assume that it is essential, meaning the hyperplanes in $\mathcal A$ intersect in the origin. The intersection lattice $L(\mathcal A)$ consists of those subspaces of $\mathbb R^d$ that are intersections of hyperplanes from $\mathcal A$ (including the empty intersection which is interpreted as $\mathbb R^d$) ordered by reverse inclusion. So notice that $\mathbb R^d$ is the bottom of $L(\mathcal A)$ and $\{0\}$ is the top. Let $\mu$ be the Mobius function of $L(\mathcal A)$.

Consider the quantity $$Q_{\mathcal A}=\sum_{W\in L(\mathcal A)}|\mu (W,\{0\})|.$$

Does $Q_{\mathcal A}$ have a natural interpretation in terms of the combinatorics of the arrangement? Note that Zaslavsky's theorem gives $$\sum_{W\in L(\mathcal A)}|\mu(\mathbb R^d,W)|$$ is the number of chambers of the arrangement, but this is, in some sense that is imprecise, a dual quantity.

I know that when $\mathcal A$ is the braid arrangement in $\mathbb R^n$, $Q_{\mathcal A}$ counts the number of necklaces of set partitions of $n$ (i.e., you look at ordered set partitions of $\{1,\ldots,n\}$ up to cyclic permutations).

$\endgroup$
2
  • $\begingroup$ You mean "central," not "essential." $\endgroup$ May 14 '21 at 20:57
  • $\begingroup$ @RichardStanley I meant both but only so I could call the intersection of the hyperplanes 0. Maybe my terminology is slightly off $\endgroup$ May 14 '21 at 21:16
7
$\begingroup$

Let $\Delta$ be a matroid complex, i.e., an abstract simplicial complex whose faces are the independent sets of a matroid $M$. Let $K[\Delta]$ denote the face ring (aka "Stanley-Reisner ring") of $\Delta$ over a field $K$. Let $\beta_i(K[\Delta])$ denote the Betti numbers of a minimal free resolution of $K[\Delta]$, regarded as a module (in fact, quotient ring) over $K[x_1,\dots,x_n]$, where $x_1,\dots,x_n$ are the vertices of $\Delta$. Then $$ \beta_i(K[\Delta]) = \sum |\mu(W,\{0\})|, $$ where $W$ ranges over all flats of $M$ of corank $i$. In particular, the sum of all the Betti numbers is the sum requested. See Theorem 9 of http://math.mit.edu/~rstan/pubs/pubfiles/34.pdf.

$\endgroup$
4
  • $\begingroup$ Thanks. Let me assimilate this. $\endgroup$ May 14 '21 at 21:23
  • $\begingroup$ Can these betti numbers be computed from the simplicial complex? $\endgroup$ May 14 '21 at 21:48
  • $\begingroup$ See the above reference. For general simplicial complexes, the Betti numbers are sums of dimensions of homology groups of restrictions $\Delta_W$ of $\Delta$ to some subset $W$ of the vertices (Theorem 4). For matroid complexes, the homology of each $\Delta_W$ vanishes except in the top dimension (this property is in fact equivalent to being a matroid complex), so the top dimensional homology is just the reduced Euler characteristic of $\Delta_W$, up to sign. $\endgroup$ May 14 '21 at 23:27
  • $\begingroup$ OK. Thanks. I remember reading about this once long ago in connection with Leray numbers. Ok. I'll try to read the paper. Thanks for your answer. $\endgroup$ May 15 '21 at 0:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.