arXiv++ Combinatorics

Dyer (1991) proved that, for each fixed length $k$, only finitely many isomorphism types of Bruhat intervals occur in finite Coxeter groups. In this short paper, we prove that this result holds also for affine Coxeter groups, and moreover that this characterizes the affine groups among all infinite Coxeter groups. We also show that the coefficients of $q$ in the Kazhdan-Lusztig polynomials of a Coxeter group are bounded if and only if the group is finite or affine.

We prove the combinatorial invariance of the coefficient of $q$ in Kazhdan-Lusztig polynomials for arbitrary Coxeter groups. As a result, we obtain the Combinatorial Invariance Conjecture for Bruhat intervals of length at most $6$. We also prove the Gabber-Joseph conjecture for the second-highest $\mathrm{Ext}$ group of a pair of Verma modules, as well as the combinatorial invariance of the dimension of this group.

The extended weak order on a Coxeter group $W$ is the poset of biclosed sets in its root system. In (Barkley-Speyer 2024), it was shown that when $W=\widetilde{S}_n$ is the affine symmetric group, then the extended weak order is a quotient of the lattice $L_n$ of translation-invariant total orderings of the integers. In this article, we give a combinatorial introduction to $L_n$ and the extended weak order on $\widetilde{S}_n$. We show that $L_n$ is an algebraic completely semidistributive lattice. We describe its canonical join representations using a cyclic version of Reading's non-crossing arc diagrams. We also show analogous statements for the lattice of all total orders of the integers, which is the extended weak order on the symmetric group $S_\infty$. A key property of both of these lattices is that they are profinite; we also prove that a profinite lattice is join semidistributive if and only if its compact elements have canonical join representations. We conjecture that the extended weak order of any Coxeter group is a profinite semidistributive lattice.

Blundell, Buesing, Davies, Veličković, and Williamson (BBDVW) introduced the notion of a hypercube decomposition of an interval in Bruhat order. They conjectured a recursive formula in terms of this structure which, if shown for all intervals, would imply the Combinatorial Invariance Conjecture of Lusztig and Dyer, for Kazhdan-Lusztig polynomials of the symmetric group. In this article, we prove implications between the BBDVW Conjecture and several other recurrences for hypercube decompositions, under varying hypotheses, which have appeared in the recent literature. As an application, we prove the BBDVW Conjecture for lower intervals $[e,v]$, the first non-trivial class of intervals for which it has been established.

We introduce the concepts of an amazing hypercube decomposition and a double shortcut for it, and use these new ideas to formulate a conjecture implying the Combinatorial Invariance Conjecture of the Kazhdan--Lusztig polynomials for the symmetric group. This conjecture has the advantage of being combinatorial in nature. The appendix by Grant T. Barkley and Christian Gaetz discusses the related notion of double hypercubes and proves an analogous conjecture for these in the case of co-elementary intervals.

We show that the Combinatorial Invariance Conjecture for Kazhdan-Lusztig polynomials due to Lusztig and to Dyer, its parabolic analog due to Marietti, and a refined parabolic version that we introduce, are equivalent. We use this to give a new proof of Marietti's conjecture in the case of lower Bruhat intervals and to prove several new cases of the parabolic conjectures.

Coxeter groups are equipped with a partial order known as the weak order, such that $u \leq v$ if the inversions of $u$ are a subset of the inversions of $v$. In finite Coxeter groups, weak order is a complete lattice, but in infinite Coxeter groups it is only a meet semi-lattice. Motivated by questions in Kazhdan-Lusztig theory, Matthew Dyer introduced a larger poset, now known as extended weak order, which contains the weak order as an order ideal and coincides with it for finite Coxeter groups. The extended weak order is the containment order on certain sets of positive roots: those which satisfy a geometric condition making them "biclosed". The finite biclosed sets are precisely the inversion sets of Coxeter group elements. Generalizing the result for finite Coxeter groups, Dyer conjectured that the extended weak order is always a complete lattice, even for infinite Coxeter groups. In this paper, we prove Dyer's conjecture for Coxeter groups of affine type. To do so, we introduce the notion of a clean arrangement, which is a hyperplane arrangement where the regions are in bijection with biclosed sets. We show that root poset order ideals in a finite or rank 3 untwisted affine root system are clean. We set up a general framework for reducing Dyer's conjecture to checking cleanliness of certain subarrangements. We conjecture this framework can be used to prove Dyer's conjecture for all Coxeter groups.

We adapt the hypercube decompositions introduced by Blundell-Buesing-Davies-Veličković-Williamson to prove the Combinatorial Invariance Conjecture for Kazhdan-Lusztig $R$-polynomials in the case of elementary intervals in $S_n$. This significantly generalizes the main previously-known case of the conjecture, that of lower intervals.

Let $W$ be a Coxeter group and let $Φ^+$ be its positive roots. A subset $B$ of $Φ^+$ is called biclosed if, whenever we have roots $α$, $β$ and $γ$ with $γ\in \mathbb{R}_{>0} α+ \mathbb{R}_{>0} β$, if $α$ and $β\in B$ then $γ\in B$ and, if $α$ and $β\not\in B$, then $γ\not\in B$. The finite biclosed sets are the inversion sets of the elements of $W$, and the containment between finite inversion sets is the weak order on $W$. Matthew Dyer suggested studying the poset of all biclosed subsets of $Φ^+$, ordered by containment, and conjectured that it is a complete lattice. As progress towards Dyer's conjecture, we classify all biclosed sets in the affine root systems. We provide both a type uniform description, and concrete models in the classical types $\widetilde{A}$, $\widetilde{B}$, $\widetilde{C}$, $\widetilde{D}$. We use our models to prove that biclosed sets form a complete lattice in types $\widetilde{A}$ and $\widetilde{C}$.

Let $m_G$ denote the number of perfect matchings of the graph $G$. We introduce a number of combinatorial tools for determining the parity of $m_G$ and giving a lower bound on the power of 2 dividing $m_G$. In particular, we introduce certain vertex sets called channels, which correspond to elements in the kernel of the adjacency matrix of $G$ modulo $2$. A result of Lovász states that the existence of a nontrivial channel is equivalent to $m_G$ being even. We give a new combinatorial proof of this result and strengthen it by showing that the number of channels gives a lower bound on the power of $2$ dividing $m_G$ when $G$ is planar. We describe a number of local graph operations which preserve the number of channels. We also establish a surprising connection between 2-divisibility of $m_G$ and dynamical systems by showing an equivalency between channels and billiard paths. We exploit this relationship to show that $2^{\frac{\gcd(m+1,n+1)-1}{2}}$ divides the number of domino tilings of the $m\times n$ rectangle. We also use billiard paths to give a fast algorithm for counting channels (and hence determining the parity of the number of domino tilings) in simply connected regions of the square grid.