Homological algebra of Nakayama algebras and 321-avoiding permutations
Eirini Chavli, Rene Marczinzik
TL;DR
The paper studies linear Nakayama algebras in bijection with Dyck paths, which are themselves in bijection with $321$-avoiding permutations via the BJS map. It develops a homological interpretation of permutation data by translating Ext-type invariants into Dyck-path statistics, obtaining two main results: (i) for the Nakayama algebra $A_{\pi}$ associated to a $321$-avoiding permutation $\pi$, the number of indecomposable projective modules with injective dimension one equals the number of fixed points of $\pi$, and (ii) the self-extension space $\operatorname{Ext}^1_A(J,J)$ is isomorphic to $K^{\mathfrak{s}(\pi)}$, where $\mathfrak{s}(\pi)$ is the permutation’s support size. These findings yield enumerative corollaries: the number of linear Nakayama algebras with $n+2$ simples and a given Ext-dimension matches the number of standard tableaux of shape $[n,k]$, with maximal Ext-dimension attained in Catalan-number abundance; the work also provides explicit inverse correspondences between the combinatorial and algebraic data. The approach unifies homological algebra with classical combinatorics by leveraging the Dyck-path representation of Nakayama data and the BJS bijection, enabling precise combinatorial interpretations of homological invariants and guiding generalization to higher $k$ in future work.
Abstract
Linear Nakayama algebras over a field $K$ are in natural bijection to Dyck paths and Dyck paths are in natural bijection to 321-avoiding bijections via the Billey-Jockusch-Stanley bijection. Thus to every 321-avoiding permutation $π$ we can associate in a natural way a linear Nakayama algebra $A_π$. We give a homological interpretation of the fixed points statistic of 321-avoiding permutations using Nakayama algebras with a linear quiver. We furthermore show that the space of self-extension for the Jacobson radical of a linear Nakayama algebra $A_π$ is isomorphic to $K^{\mathfrak{s}(π)}$, where $\mathfrak{s}(π)$ is defined as the cardinality $k$ such that $π$ is the minimal product of transpositions of the form $s_i=(i,i+1)$ and $k$ is the number of distinct $s_i$ that appear.
