The h h -vectors of toric ideals of odd cycle compositions revisited

Kieran Bhaskara, Adam Van Tuyl, Sasha Zotine · 2025

Read The h h -vectors of toric ideals of odd cycle compositions revisited by Kieran Bhaskara, Adam Van Tuyl, Sasha Zotine (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-13087.txt.

Introduction

Kieran Bhaskara and co-authors revisit the h-vectors of toric ideals built from graphs made of odd cycles sharing one vertex. A recent paper computed the key polynomial from cycle counts and sizes. This note proves a stronger structural result instead. The ideals turn out to be geometrically vertex decomposable. That property yields the earlier formula as a simple corollary. The short paper serves readers in combinatorial algebra.

Catalog record

CREATORKieran Bhaskara, Adam Van Tuyl, Sasha Zotine
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE22 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-13087

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The h h -vectors of toric ideals of odd cycle compositions revisited Kieran Bhaskara, Adam Van Tuyl, Sasha Zotine Abstract . Let G G be a graph consisting of s s odd cycles that all share a common vertex. Bhaskara, Higashitani, and Shibu Deepthi recently computed the h h -polynomial for the quotient ring R / I G R/I_{G} , where I G I_{G} is the toric ideal of G G , in terms of the number and sizes of odd cycles in the graph. The purpose of this note is to prove the stronger result that these toric ideals are geometrically vertex decomposable, which allows us to deduce the result of Bhaskara, Higashitani, and Shibu Deepthi about the h h -polyhomial as a corollary. 2020 Mathematics Subject Classification 13D02, 13D40, 13P10, 13F65, 14M25, 05E40 1. Introduction Toric ideals of graphs provide a fruitful bridge between combinatorics and commutative algebra. Recall that if G=(V,E) is a finite simple graph on the vertex set V=\{x_{1},\ldots,x_{n}\} and edge set E=\{e_{1},\ldots,e_{m}\} , then the toric ideal of G is the kernel of the \mathbb{K} -algebra homomorphism \varphi:\mathbb{K}[e_{1},\ldots,e_{m}]\rightarrow\mathbb{K}[x_{1},\ldots,x_{n}] given by \varphi(e_{i})=x_{j}x_{k} if e_{i}=\{x_{j},x_{k}\} . We denote {\rm ker}(\varphi) by I_{G} and write \mathbb{K}[G] for \mathbb{K}[e_{1},\ldots,e_{m}]/I_{G} , which is sometimes referred to as the toric edge ring of G . Studying how the combinatorics of G are reflected in \mathbb{K}[G] , and vice-versa, is an active area of research; see [ 1 , 3 , 5 , 7 , 8 , 14 ] for a small sample of this work. A recent example of this interaction is work of Bhaskara-Higashitani-Shibu Deepthi [ 2 ] , generalizing earlier work of Higashitani-Shibu Deepthi [ 9 ] , on the h -polynomial of the toric edge rings of the family of graphs consisting of odd cycles that share a common vertex. More precisely, fix an integer s\geq 1 and let t_{1},\ldots,t_{s} be positive integers. Let G=G(t_{1},\ldots,t_{s}) be the graph consisting of s odd cycles of size 2t_{i}+1 for i=1,\ldots,s , where all the odd cycles share a common “center” vertex. An example of the graph G(2,1,2) is given in Figure 1 . Figure 1. The graph G(2,1,2) consisting of two 2\cdot 2+1=5 -cycles and one 2\cdot 1+1=3 -cycle meeting at a common vertex The main result of [ 2 ] is to compute the h -polynomial of \mathbb{K}[G(t_{1},\ldots,t_{s})] , that is, the numerator of the reduced Hilbert series of \mathbb{K}[G(t_{1},\ldots,t_{s})] (see the next section or [ 2 ] for any undefined terms). Note that when t_{1}=\cdots=t_{s}=1 , the graph G(1,\ldots,1) consists of s triangles that share a common vertex—this family is the one studied in [ 9 ] . The proof strategy in both [ 2 ] and [ 9 ] is to consider the initial ideal of the toric ideal of G(t_{1},\ldots,t_{s}) . For this family of graphs, the initial ideal is a squarefree monomial ideal and thus corresponds to a simplicial complex via the Stanley–Reisner correspondence. Through a careful study of the facets of this simplicial complex, the authors of [ 2 , 9 ] construct a recursive formula for the Hilbert series, allowing them to deduce their main result. The purpose of this article is to give an entirely different proof that avoids passing to the initial ideal of I_{G} . In fact, we prove that the ideals I_{G} satisfy a stronger algebraic property: Theorem 1.1 . The toric ideal I_{G} of G=G(t_{1},\ldots,t_{s}) is geometrically vertex decomposable. Geometrically vertex decomposable ideals were introduced by Klein and Rajchgot [ 10 ] , extending work of Knutson–Miller–Yong [ 11 ] on vertex decomposability. Very roughly speaking, geometrically vertex decomposable ideals generalize the properties of Stanley–Reisner ideals of vertex decomposable simplicial complexes to all ideals. As shown in [ 5 , 6 , 12 ] , geometrically vertex decomposable ideals are a useful addition to the commutative algebra toolkit. Theorem 1.1 also demonstrates the advantage of identifying geometrically vertex decomposable ideals. Indeed, using the properties of the Hilbert series of quotients by geometrically vertex decomposable ideals developed by Nguyễn–Rajchgot–Van Tuyl [ 12 ] , the main results of [ 2 , 9 ] are now corollaries of Theorem 1.1 . Corollary 1.2 ( [ 2 , Theorem 1.1] ) . The h-polynomial of {\mathbb{K}}[G(t_{1},\ldots,t_{s})] is (1.1) h\bigl({\mathbb{K}}[G(t_{1},\ldots,t_{s})];z\bigr)=\prod_{i=1}^{s}(1+z+z^{2}+\cdots+z^{t_{i}})-z\prod_{i=1}^{s}(1+z+z^{2}+\cdots+z^{t_{i}-1}). As a second application, we compute the Castelnuovo-Mumford regularity of \mathbb{K}[G(t_{1},\ldots,t_{s})] . Corollary 1.3 . The Castelnuovo-Mumford regularity of {\mathbb{K}}[G(t_{1},\ldots,t_{s})] is {\rm reg}({\mathbb{K}}[G(t_{1},\ldots,t_{s})])=\begin{cases}t_{1}+\cdots+t_{s}&~~\mbox{if $s\geq 2$}\\ 0&~~\mbox{if $s=1$.}\end{cases} In the next section we recall the relevant background and facts on geometrically vertex decomposable ideals and h -polynomials. In Section 3 , we prove Theorem 1.1 and its consequences, Corollary 1.2 and Corollary 1.3 . 2. Geometrically vertex decomposable ideals and Hilbert series Throughout this paper \mathbb{K} will denote a field. Let R=\mathbb{K}[x_{1},\ldots,x_{n}] be a polynomial ring and fix a variable y\in R . For any f\in R , there is some integer d\geq 0 such that we can write f as f=\sum_{i=0}^{d}\alpha_{i}y^{i} where \alpha_{i} is a polynomial in the variables \{x_{1},\ldots,x_{n}\}\setminus\{y\} . The initial y -form of f is {\rm in}_{y}(f)=\alpha_{d}y^{d} . For any ideal I of R , the ideal {\rm in}_{y}(I)=\langle{\rm in}_{y}(f)~|~f\in I\rangle is the ideal generated by all the initial y -forms of f in the ideal I . A monomial order < is y -compatible if {\rm in}_{<}(f)={\rm in}_{<}({\rm in}_{y}(f)) for all f\in R . Fixing a y -compatible monomial order < , suppose that \mathcal{G}(I)=\{g_{1},\ldots,g_{t}\} is a Gröbner basis of I with resect to < . For each g_{i} in the Gröbner basis, we write it as g_{i}=q_{i}y^{d_{i}}+r_{i} where {\rm in}_{y}(g_{i})=q_{i}y^{d_{i}} . Note that this means y does not divide any term of q_{i} and y^{d_{i}} does not divide any term of r_{i} . With this notation, we define two new ideals: \mathcal{C}_{y,I}=\langle q_{1},\ldots,q_{t}\rangle~~\mbox{and}~~\mathcal{N}_{y,I}=\langle q_{i}~|~d_{i}=0\rangle. We now recall the definition of a geometrically vertex decomposable ideal. Definition 2.1 ( [ 10 , Definition 2.7] ) . An ideal I of R=\mathbb{K}[x_{1},\ldots,x_{n}] is geometrically vertex decomposable (GVD) if I is unmixed (that is, all of the associated primes of I have the same height) and (1) I=\langle 1\rangle , or I is generated by a (possible empty) subset of variables, or (2) there exists a variable y and y -compatible monomial order such that {\rm in}_{y}(I)=\mathcal{C}_{y,I}\cap(\mathcal{N}_{y,I}+\langle y\rangle) and the contractions of \mathcal{C}_{y,I} and \mathcal{N}_{y,I} to \mathbb{K}[x_{1},\ldots,\hat{y},\ldots,x_{n}] are geometrically vertex decomposable. The following lemma allows us to conclude the equality {\rm in}_{y}(I)=\mathcal{C}_{y,I}\cap\bigl(\mathcal{N}_{y,I}+\langle y\rangle\bigr) automatically holds as long as we preserve a suitably nice generating set. Lemma 2.2 . [ 11 , Theorem 2.1] Let I be an ideal of R , set < to be a y -compatible monomial order, and assume \{g_{1},\ldots,g_{t}\} is a Gröbner basis of I with respect to < . Suppose that {\rm in}_{y}(g_{i})=q_{i}y^{d_{i}} with d_{i}=0 or 1 for all i. Then (1) \{q_{1},\ldots,q_{t}\} is a Gröbner basis of \mathcal{C}_{y,I} , \{q_{i}~|~d_{i}=0\} is a Gröbner basis of \mathcal{N}_{y,I} , and (2) {\rm in}_{y}(I)=\mathcal{C}_{y,I}\cap\bigl(\mathcal{N}_{y,I}+\langle y\rangle\bigr). For our proof of Theorem 1.1 , we need that certain complete intersections are GVD. Lemma 2.3 ( [ 5 , Corollary 2.12] ) . If I is a squarefree monomial ideal that is also a complete intersection, then I is geometrically vertex decomposable. We now turn our attention to the

Read the complete text