Word posets, complexity, and Coxeter groups

TL;DR

This paper introduces word posets to analyze the structure of commutation classes in monoids and Coxeter groups, demonstrating that counting words in these classes is #P-complete and providing recursive formulas for enumeration.

Contribution

It defines word posets as a new tool to study commutation classes, establishes their correspondence with these classes, and applies this framework to Coxeter groups, including complexity results and enumeration methods.

Findings

Counting words in commutation classes is #P-complete.

Enumerating reduced words in Coxeter groups is #P-complete.

A recursive formula for the number of commutation classes of reduced words is derived.

Abstract

A monoid $M$ generated by a set $S$ of symbols can be described as the set of equivalence classes of finite words in $S$ under some relations that specify when some contiguous sequence of symbols can be replaced by another. If $a,b\in S$, a relation of the form $ab=ba$ is said to be a \emph{commutation relation}. Words that are equivalent using only a sequence of commutation relations are said to be in the same \emph{commutation class}. We introduce certain partially ordered sets that we call \emph{word posets} that capture the structure of commutation classes in monoids. The isomorphism classes of word posets are seen to be in bijective correspondence with the commutation classes, and we show that the linear extensions of the word poset correspond bijectively to the words in the commutation class, using which we demonstrate that enumerating the words in commutation classes of monoids is #P-complete. We then apply the word posets to Coxeter groups. We show that the problem of enumerating the reduced words of elements of Coxeter groups is #P-complete. We also demonstrate a method for finding the word posets for commutation classes of reduced words of an element, then use this to find a recursive formula for the number of commutation classes of reduced words.