Key Words and Phrases

EDGE IDEALS: ALGEBRAIC AND COMBINATORIAL PROPERTIES
SUSAN MOREY AND RAFAEL H. VILLARREAL

arXiv:1012.5329v3 [math.AC] 29 Mar 2011
Abstract. Let C be a clutter and let I(C) ⊂ R be its edge ideal. This is a survey paper on
the algebraic and combinatorial properties of R/I(C) and C, respectively. We give a criterion to
estimate the regularity of R/I(C) and apply this criterion to give new proofs of some formulas for
the regularity. If R/I(C) is sequentially Cohen-Macaulay, we present a formula for the regularity
of the ideal of vertex covers of C and give a formula for the projective dimension of R/I(C). We
also examine the associated primes of powers of edge ideals, and show that for a graph with a
leaf, these sets form an ascending chain.
1. Introduction
A clutter C is a finite ground set X together with a family E of subsets of X such that if
f1 , f2 ∈ E, then f1 6⊂ f2 . The ground set X is called the vertex set of C and E is called the
edge set of C, denoted by V (C) and E(C) respectively. Clutters are simple hypergraphs and
are sometimes called Sperner families in the literature. We can also think of a clutter as the
maximal faces of a simplicial complex over a ground set. One example of a clutter is a graph
with the vertices and edges defined in the usual way.
Let C be a clutter with vertex set X = {x1 , . . . , xn } and with edge set E(C). Permitting
an abuse of notation, we will also denote by xi the ith variable in the polynomial ring R =
QK. The edge ideal of C, denoted by I(C), is the ideal of R generated
K[x1 , . . . , xn ] over a field
by all monomials xe = xi ∈e xi such that e ∈ E(C). Edge ideals of graphs and clutters were
introduced in [109] and [39, 47, 53], respectively. The assignment C 7→ I(C) establishes a natural
one-to-one correspondence between the family of clutters and the family of square-free monomial
ideals. Edge ideals of clutters are also called facet ideals [39].
This is a survey paper on edge ideals, which includes some new proofs of known results and
some new results. The study of algebraic and combinatorial properties of edge ideals and clut-
ters (e.g., Cohen-Macaulayness, unmixedness, normality, normally torsion-freeness, shellability,
vertex decomposability, stability of associated primes) is of current interest, see [22, 24, 25, 32,
39, 40, 41, 45, 51, 61, 62, 89, 116] and the references there. In this paper we will focus on
the following algebraic properties: the sequentially Cohen-Macaulay property, the stability of
associated primes, and the connection between torsion-freeness and combinatorial problems.
The numerical invariants of edge ideals have attracted a great deal of interest [1, 46, 53, 77,
87, 90, 91, 106, 110, 117, 118]. In this paper we focus on the following invariants: projective
dimension, regularity, depth and Krull dimension.
We present a few new results on edge ideals. We give a criterion to estimate the regularity
of edge ideals (see Theorem 3.14). We apply this criterion to give new proofs of some formulas
for the regularity of edge ideals (see Corollary 3.15). If C is a clutter and R/I(C) is sequentially
2000 Mathematics Subject Classification. Primary 13-02, 13F55; Secondary 05C10, 05C25, 05C65, 05E40.
Key words and phrases. Edge ideal, regularity, associated prime, sequentially Cohen-Macaulay.
The second author was partially supported by CONACyT grant 49251-F and SNI.
1
2 SUSAN MOREY AND RAFAEL H. VILLARREAL
Cohen-Macaulay, we present a formula for the regularity of the ideal of vertex covers of C (see
Theorem 3.31) and give a formula for the projective dimension of R/I(C) (see Corollary 3.33).
We also give a new class of monomial ideals for which the sets of associate primes of powers are
known to form ascending chains (Proposition 4.23).
For undefined terminology on commutative algebra, edge ideals, graph theory, and the theory
of clutters and hypergraphs we refer to [10, 28, 102], [38, 110], [7, 54], [17, 97], respectively.
2. Algebraic and combinatorial properties of edge ideals
Let C be a clutter with vertex set X = {x1 , . . . , xn } and let I = I(C) ⊂ R be its edge ideal.
A subset F of X is called independent or stable if e 6⊂ F for any e ∈ E(C). The dual concept
of a stable vertex set is a vertex cover , i.e., a subset C of X is a vertex cover of C if and only
if X \ C is a stable vertex set. A first hint of the rich interaction between the combinatorics of
C and the algebra of I(C) is that the number of vertices in a minimum vertex cover of C (the
covering number α0 (C) of C) coincides with the height ht I(C) of the ideal I(C). The number of
vertices in a maximum stable set (the stability number of C) is denoted by β0 (C). Notice that
n = α0 (C) + β0 (C).
A less immediate interaction between the two fields comes from passing to a simplicial com-
plex and relating combinatorial properties of the complex to algebraic properties of the ideal.
The Stanley-Reisner complex of I(C), denoted by ∆C , is the simplicial complex whose faces are
the independent vertex sets of C. The complex ∆C is also called the independence complex of
C. Recall that ∆C is called pure if all maximal independent vertex sets of C, with respect to
inclusion, have the same number of elements. If ∆C is pure (resp. Cohen-Macaulay, shellable,
vertex decomposable), we say that C is unmixed (resp. Cohen-Macaulay, shellable, vertex de-
composable). Since minor variations of the definition of shellability exist in the literature, we
state here the definition used throughout this article.
Definition 2.1. A simplicial complex ∆ is shellable if the facets (maximal faces) of ∆ can be
ordered F1 , . . . , Fs such that for all 1 ≤ i < j ≤ s, there exists some v ∈ Fj \ Fi and some
ℓ ∈ {1, . . . , j − 1} with Fj \ Fℓ = {v}.
We are interested in determining which families of clutters have the property that ∆C is pure,
Cohen-Macaulay, or shellable. These properties have been extensively studied, see [10, 89, 92,
93, 94, 100, 102, 110, 111] and the references there.
The above definition of shellable is due to Björner and Wachs [6] and is usually referred to as
nonpure shellable, although here we will drop the adjective “nonpure”. Originally, the definition
of shellable also required that the simplicial complex be pure, that is, all facets have the same
dimension. We will say ∆ is pure shellable if it also satisfies this hypothesis. These properties
are related to other important properties [10, 102, 110]:
pure shellable ⇒ constructible ⇒ Cohen-Macaulay ⇐ Gorenstein.
If a shellable complex is not pure, an implication similar to that above holds when Cohen-
Macaulay is replaced by sequentially Cohen-Macaulay.
Definition 2.2. Let R = K[x1 , . . . , xn ]. A graded R-module M is called sequentially Cohen-
Macaulay (over K) if there exists a finite filtration of graded R-modules
(0) = M0 ⊂ M1 ⊂ · · · ⊂ Mr = M
3
such that each Mi /Mi−1 is Cohen-Macaulay, and the Krull dimensions of the quotients are
increasing:
dim(M1 /M0 ) < dim(M2 /M1 ) < · · · < dim(Mr /Mr−1 ).
We call a clutter C sequentially Cohen-Macaulay if R/I(C) is sequentially Cohen-Macaulay.

As first shown by Stanley [102], shellable implies sequentially Cohen-Macaulay.
A related notion for a simplicial complex is that of vertex decomposability [5]. If ∆ is a
simplicial complex and v is a vertex of ∆, then the subcomplex formed by deleting v is the
simplicial complex consisting of the faces of ∆ that do not contain v, and the link of v is
lk(v) = {F ∈ ∆|v 6∈ F and F ∪ {v} ∈ ∆}.
Suppose ∆ is a (not necessarily pure) simplicial complex. We say that ∆ is vertex-decomposable
if either ∆ is a simplex, or ∆ contains a vertex v such that both the link of v and the subcomplex
formed by deleting v are vertex-decomposable, and such that every facet of the deletion is a
facet of ∆. If C is vertex decomposable, i.e., ∆C is vertex decomposable, then C is shellable and
sequentially Cohen-Macaulay [5, 116]. Thus, we have:
vertex decomposable ⇒ shellable ⇒ sequentially Cohen-Macaulay.
Two additional properties related to the properties above are also of interest in this area.
One is the unmixed property, which is implied by the Cohen-Macaulay property. The other is
balanced. To define balanced, it is useful to have a matrix that encodes the edges of a graph or
clutter.
Definition 2.3. Let f1 , . . . , fq be the edges of a clutter C. The incidence matrix or clutter
matrix of C is the n × q matrix A = (aij ) given by aij = 1 if xi ∈ fj and aij = 0 otherwise. We
say that C is a totally balanced clutter (resp. balanced clutter) if A has no square submatrix of
order at least 3 (resp. of odd order) with exactly two 1’s in each row and column.
If G is a graph, then G is balanced if and only if G is bipartite and G is totally balanced if

and only if G is a forest [96, 97].
While the implications between the properties mentioned above are interesting in their own
right, it is useful to identify classes of ideals that satisfy the various properties. We begin with
the Cohen-Macaulay and unmixed properties. There are classifications of the following families
in terms of combinatorial properties of the graph or clutter:
(c1 ) [94, 111] unmixed bipartite graphs,
(c2 ) [36, 60] Cohen-Macaulay bipartite graphs,
(c3 ) [109] Cohen-Macaulay trees,
(c4 ) [38] totally balanced unmixed clutters,
(c5 ) [89] unmixed clutters with the König property without cycles of length 3 or 4,
(c6 ) [89] unmixed balanced clutters.
We now focus on the sequentially Cohen-Macaulay property.
Proposition 2.4. [43] The only sequentially Cohen-Macaulay cycles are C3 and C5
The next theorem generalizes a result of [36] (see (c2 ) above) which shows that a bipartite
graph G is Cohen-Macaulay if and only if ∆G has a pure shelling.
Theorem 2.5. [107] Let G be a bipartite graph. Then G is shellable if and only if G is sequen-
tially Cohen-Macaulay.
Recently Van Tuyl [106] has shown that Theorem 2.5 remains valid if we replace shellable by
vertex decomposable.
Additional examples of sequentially Cohen-Macaulay ideals depend on the chordal structure
of the graph. A graph G is said to be chordal if every cycle of G of length at least 4 has a chord.
A chord of a cycle is an edge joining two non-adjacent vertices of the cycle. Chordal graphs
have been extensively studied, and they can be constructed according to a result of G. A. Dirac,
see [21, 63, 104]. A chordal graph is called strongly chordal if every cycle C of even length at
least six has a chord that divides C into two odd length paths. A clique of a graph is a set
of mutually adjacent vertices. Totally balanced clutters are precisely the clutters of maximal
cliques of strongly chordal graphs by a result of Farber [37]. Faridi [39] introduced the notion
of a simplicial forest. In [62, Theorem 3.2] it is shown that C is the clutter of the facets of a
simplicial forest if and only if C is a totally balanced clutter. Additionally, a clutter C is called
d-uniform if all its edges have size d.
Theorem 2.6. Any of the following clutters is sequentially Cohen-Macaulay:
(a) [116] graphs with no chordless cycles of length other than 3 or 5,
(b) [43] chordal graphs,
(c) [63] clutters whose ideal of covers has linear quotients (see Definitions 2.7 and 3.1),
(d) [55] clutters of paths of length t of directed rooted trees,
(e) [39] simplicial forests, i.e., totally balanced clutters,
(f) [52] uniform admissible clutters whose covering number is 3.
The clutters of parts (a)–(f) are in fact shellable, and the clutters of parts (a)–(b) are in
fact vertex decomposable, see [22, 63, 106, 107, 115, 116]. The family of graphs in part (b) is
contained in the family of graphs of part (a) because the only induced cycles of a chordal graph
are 3-cycles.
A useful tool in examining invariants related to resolutions comes from a carefully chosen
ordering of the generators.
Definition 2.7. A monomial ideal I has linear quotients if the monomials that generate I can
be ordered g1 , . . . , gq such that for all 1 ≤ i ≤ q − 1, ((g1 , . . . , gi ) : gi+1 ) is generated by linear
forms.
If an edge ideal I is generated in a single degree and I has linear quotients, then I has a linear
resolution (cf. [39, Lemma 5.2]). If I is the edge ideal of a graph, then I has linear quotients if
and only if I has a linear resolution if and only if each power of I has a linear resolution [64].
Let G be a graph. Given a subset A ⊂ V (G), by G \ A, we mean the graph formed from G by
deleting all the vertices in A, and all edges incident to a vertex in A. A graph G is called vertex-
critical if α0 (G \ {v}) < α0 (G) for all v ∈ V (G). An edge critical graph is defined similarly. The
final property introduced in this section is a combinatorial decomposition of the vertex set of a
graph.
Definition 2.8. [2] A graph G without isolated verticesSis called a B-graph if there is a family
G consisting of independent sets of G such that V (G) = C∈G C and |C| = β0 (G) for all C ∈ G.
The notion of a B-graph is at the center of several interesting families of graphs. One has the
following implications for any graph G without isolated vertices [2, 110]:
edge-critical =⇒
B-graph =⇒ vertex-critical
Cohen-Macaulay =⇒ unmixed =⇒
5
In [2] the integer α0 (G) is called the transversal number of G.

Theorem 2.9. [34, 46] If G is a B-graph, then β0 (G) ≤ α0 (G).
3. Invariants of edge ideals: regularity, projective dimension, depth
Let C be a clutter and let I = I(C) be its edge ideal. In this section we study the regularity,
depth, projective dimension, and Krull dimension of R/I(C). There are several well-known
results relating these invariants that will prove useful. We collect some of them here for ease of
reference.
The first result is a basic relation between the dimension and the depth (see for example [28,
Proposition 18.2]):
(3.1) depth R/I(C) ≤ dim R/I(C).
The deviation from equality in the above relationship can be quantified using the projective
dimension, as is seen in a formula discovered by Auslander and Buchsbaum (see [28, Theo-
rem 19.9]):
(3.2) pdR (R/I(C)) + depth R/I(C) = depth(R).
Notice that since in the setting of this article R is a polynomial ring in n variables, depth(R) = n.
Another invariant of interest also follows from a closer inspection of a minimal projective
resolution of R/I. Consider the minimal graded free resolution of M = R/I as an R-module:
M M
F⋆ : 0 → R(−j)bgj → · · · → R(−j)b1j → R → R/I → 0.
j j
The Castelnuovo-Mumford regularity or simply the regularity of M is defined as
reg(M ) = max{j − i| bij 6= 0}.
An excellent reference for the regularity is the book of Eisenbud [29]. There are methods to
compute the regularity of R/I avoiding the construction of a minimal graded free resolution,
see [3] and [50, p. 614]. These methods work for any homogeneous ideal over an arbitrary field.
We are interested in finding good bounds for the regularity. Of particular interest is to be
able to express reg(R/I(C)) in terms of the combinatorics of C, at least for some special families
of clutters. Several authors have studied the regularity of edge ideals of graphs and clutters
[18, 20, 53, 74, 75, 77, 81, 90, 103, 106, 114]. The main results are general bounds for the
regularity and combinatorial formulas for the regularity of special families of clutters. The
estimates for the regularity are in terms of matching numbers and the number of cliques needed
to cover the vertex set. Covers will play a particularly important role since they form the basis
for a duality.
Definition 3.1. The ideal of covers of I(C), denoted by Ic (C), is the ideal of R generated by all
the monomials xi1 · · · xik such that {xi1 , . . . , xik } is a vertex cover of C. The ideal Ic (C) is also
called the Alexander dual of I(C) and is also denoted by I(C)∨ . The clutter of minimal vertex
covers of C, denoted by C ∨ , is called the Alexander dual clutter or blocker of C.
To better understand the Alexander dual, let e ∈ E(C) and consider the monomial prime
ideal (e) = ({xi |xi ∈ e}). Then the duality is given by:
I(C) = (xe1 , xe2 , . . . , xeq ) = p1 ∩ p2 ∩ · · · ∩ ps
(3.3) l l
Ic (C) = (e1 ) ∩ (e2 ) ∩ · · · ∩ (eq ) = (xp1 , xp2 , . . . , xps ),
Q
where p1 , . . . , ps are the associated primes of I(C) and xpk = xi ∈pk xi for 1 ≤ k ≤ s. Notice the
equality Ic (C) = I(C ∨ ). Since (C ∨ )∨ = C, we have Ic (C ∨ ) = I(C). In many cases I(C) reflects
properties of Ic (C) and viceversa [27, 58, 86]. The following result illustrates this interaction.
Theorem 3.2. [103] Let C be a clutter. If ht(I(C)) ≥ 2, then
reg I(C) = 1 + reg R/I(C) = pd R/Ic (C),
where Ic (C) is the ideal of minimal vertex covers of C.
If |e| ≥ 2 for all e ∈ E(C), then this formula says that the regularity of R/I(C) equals 1 if
and only if Ic (C) is a Cohen-Macaulay ideal of height 2. This formula will be used to show
that regularity behaves well when working with edge ideals with disjoint sets of variables (see
Proposition 3.4). This formula also holds for edge ideals of height one [61, Proposition 8.1.10].
Corollary 3.3. If ht(I(C)) = 1, then reg R/I(C) = pd R/Ic (C) − 1.
Proof. We set I = I(C). The formula clearly holds if I = (x1 · · · xr ) is a principal ideal. Assume
that I is not principal. Consider the primary decomposition of I
I = (x1 ) ∩ · · · ∩ (xr ) ∩ p1 ∩ · · · ∩ pm ,
where L = p1 ∩ · · · ∩ pm is an edge ideal of height at least 2. Notice that I = f L, where
f = x1 · · · xr . Then the Alexander dual of I is
I ∨ = (x1 , . . . , xr , xp1 , xp2 , . . . , xpm ) = (x1 , . . . , xr ) + L∨ .
f
The multiplication map L[−r] → f L induces an isomorphism of graded R-modules. Thus
reg(L[−r]) = r + reg(L) = reg(I). By the Auslander-Buchsbaum formula, one has the equality
pd(R/I ∨ ) = r + pd(R/L∨ ). Therefore, using Theorem 3.2, we get
3.2
reg(R/I) = reg(R/L) + r = (pd(R/L∨ ) − 1) + r = pd(R/I ∨ ) − 1.
Thus reg(R/I) = pd(R/I ∨ ) − 1, as required.
Next we show some basic properties of regularity. The first such property is that regularity
behaves well when working with the edge ideal of a graph with multiple disjoint components or
with isolated vertices, as can be seen by the following proposition.
Proposition 3.4. [117, Lemma 7] Let R1 = K[x] and R2 = K[y] be two polynomial rings over
a field K and let R = K[x, y]. If I1 and I2 are edge ideals of R1 and R2 respectively, then
reg R/(I1 R + I2 R) = reg(R1 /I1 ) + reg(R2 /I2 ).
Proof. By abuse of notation, we will write Ii in place of Ii R for i = 1, 2 when it is clear from
context that we are using the generators of Ii but extending to an ideal of the larger ring.
Let x = {x1 , . . . , xn } and y = {y1 , . . . , ym } be two disjoint sets of variables. Notice that
(I1 + I2 )∨ = I1∨ I2∨ = I1∨ ∩ I2∨ where Ii∨ is the Alexander dual of Ii (see Definition 3.1). Hence by
Theorem 3.2 and using the Auslander-Buchsbaum formula, we get
reg(R/(I1 + I2 )) = n + m − depth(R/(I1∨ ∩ I2∨ )) − 1,
reg(R1 /I1 ) + reg(R2 /I2 ) = n − depth(R1 /I1∨ ) − 1 + m − depth(R2 /I2∨ ) − 1
Therefore we need only show the equality
depth(R/(I1∨ ∩ I2∨ )) = depth(R1 /I1∨ ) + depth(R2 /I2∨ ) + 1.
7
Since depth(R/(I1∨ + I2∨ )) = depth(R1 /I1∨ ) + depth(R2 /I2∨ ), the proof reduces to showing the
equality
(3.4) depth(R/(I1∨ ∩ I2∨ )) = depth(R/(I1∨ + I2∨ )) + 1.
We may assume that depth(R/I1∨ ) ≥ depth(R/I2∨ ). There is an exact sequence of graded R-
modules:
ϕ φ
(3.5) 0 −→ R/(I1∨ ∩ I2∨ ) −→ R/I1∨ ⊕ R/I2∨ −→ R/(I1∨ + I2∨ ) −→ 0,
where ϕ(r) = (r, −r) and φ(r 1 , r 2 ) = r1 + r2 . From the inequality
depth(R/I1∨ ⊕ R/I2∨ ) = max{depth(R/Ii∨ )}2i=1 = depth(R/I1∨ ) = depth(R1 /I1∨ ) + m
> depth(R1 /I1∨ ) + depth(R2 /I2∨ ) = depth(R/(I1∨ + I2∨ ))
and applying the depth lemma (see [10, Proposition 1.2.9] for example) to Eq. (3.5), we obtain
Eq. (3.4).
Another useful property of regularity is that one can delete isolated vertices of a graph without
changing the regularity of the edge ideal. The following lemma shows that this can be done
without significant changes to the projective dimension as well.
Lemma 3.5. Let R = K[x1 , . . . , xn ] and I be an ideal of R. If I ⊂ (x1 , . . . , xn−1 ), and R′ =
R/(xn ) ∼
= K[x1 , . . . , xn−1 ], then reg(R/I) = reg(R′ /I) and pdR (R/I) = pdR′ (R′ /I). Similarly,
if xn ∈ I and I ′ = I/(xn ), then reg(R/I) = reg(R′ /I ′ ) and pdR (R/I) = pdR′ (R′ /I ′ ) + 1.
Proof. The first result for projective dimension follows from the Auslander-Buchsbaum formula
since depth(R/I) = depth(R′ /I) + 1 and depth(R) = depth(R′ ) + 1. Since depth(R/I) =
depth(R′ /I ′ ) the second result for projective dimension holds as well. The results for regularity
follow from Proposition 3.4 by noting that the regularity of a polynomial ring K[x] is 0, as is
the regularity of the field K = K[x]/(x).
While adding variables to the ring will preserve the regularity, other changes to the base
ring, such as changing the characteristic, will affect this invariant. The following example shows
that, even for graphs, a purely combinatorial description of the regularity might not be possible.
Results regarding the role the characteristic of the field plays in the resolution of the ideal appear
in [19, 74].
Example 3.6. Consider the edge ideal I ⊂ R = K[x1 , . . . , x10 ] generated by the monomials
x1 x3 , x1 x4 , x1 x7 , x1 x10 , x1 x11 , x2 x4 , x2 x5 , x2 x8 , x2 x10 ,
x2 x11 , x3 x5 , x3 x6 , x3 x8 , x3 x11 , x4 x6 , x4 x9 , x4 x11 , x5 x7 ,
x5 x9 , x5 x11 , x6 x8 , x6 x9 , x7 x9 , x7 x10 , x8 x10 .
Using Macaulay2 [49] we get that reg(R/I) = 3 if char(K) = 2, and reg(R/I) = 2 if char(K) = 3.
As mentioned in Theorem 2.6(b), chordal graphs provide a key example of a class of clutters
whose edge ideals are sequentially Cohen-Macaulay. Much work has been done toward finding
hypergraph generalizations of chordal graphs, typically by looking at cycles of edges or at tree
hypergraphs [32, 53, 113]. The papers [53, 113, 115] contribute to the algebraic approach that
is largely motivated by finding hypergraph generalizations that have edge ideals with linear
resolutions.
It is useful to consider the homogeneous components of the ideals when using linear resolutions.
Let (Id ) denote the ideal generated by all degree d elements of a homogeneous ideal I. Then I
is called componentwise linear if (Id ) has a linear resolution for all d. If I is the edge ideal of a
clutter, we write I[d] for the ideal generated by all the squarefree monomials of degree d in I.
Theorem 3.7. Let K be a field and C be a clutter. Then
(a) [27] R/I(C) is Cohen-Macaulay if and only if Ic (C) has a linear resolution.
(b) [58] R/I(C) is sequentially Cohen-Macaulay if and only if Ic (C) is componentwise linear.
(c) [58] I(C) is componentwise linear if and only if I(C)[d] has a linear resolution for d ≥ 0.
(d) [44] If G is a graph, then I(G) has a linear resolution if and only if Gc is chordal.
(e) [103] If G is a graph, then reg(R/I(G)) = 1 if and only if Ic (G) is Cohen-Macaulay.
A graph whose complement is chordal is called co-chordal. A consequence of this result and
Theorem 3.2 is that an edge ideal I(C) has regularity 2 if and only if ∆C is the independence
complex of a co-chordal graph. In this case the complex ∆C turns out to be a quasi-forest in
the sense of Zheng [118]. In [61, Theorem 9.2.12] it is shown that a complex ∆ is a quasi-forest
if and only if ∆ is the clique complex of a chordal graph.
Information about the regularity of a clutter can also be found by examining smaller, closely
related clutters. Let S be a set of vertices of a clutter C. The induced subclutter on S, denoted
by C[S], is the maximal subclutter of C with vertex set S. Thus the vertex set of C[S] is S and
the edges of C[S] are exactly the edges of C contained in S. Notice that C[S] may have isolated
vertices, i.e., vertices that do not belong to any edge of C[S]. If C is a discrete clutter, i.e., all
the vertices of C are isolated, we set I(C) = 0 and α0 (C) = 0. A clutter of the form C[S] for
some S ⊂ V (C) is called an induced subclutter of C.
Proposition 3.8. If D is an induced subclutter of C, then reg(R/I(D)) ≤ reg(R/I(C)).
Proof. There is S ⊂ V (C) such that D = C[S]. Let p be the prime ideal of R generated by the
variables in S. By duality, we have
\ \ \
Ic (C) = (e) =⇒ Ic (C)p = (e)p = (e)p = Ic (D)p .
e∈E(C) e∈E(C) e∈E(D)
Therefore, using Theorem 3.2 and Lemma 3.5, we get

reg(R/I(C)) = pd(R/Ic (C)) − 1
≥ pd(Rp /Ic (C)p ) − 1 = pd(Rp /Ic (D)p ) − 1
= pd(R′ /Ic (D)) − 1 = pd(R/Ic (D)) − 1 = reg(R/I(D)),
where R′ is the polynomial ring K[S]. Thus, reg(R/I(C)) ≥ reg(R/I(D)).
Several combinatorially defined invariants that bound the regularity or other invariants of a
clutter are given in terms of subsets of the edge set of the clutter. An induced matching in a
clutter C is a set of pairwise disjoint edges f1 , . . . , fr such that the only edges of C contained in
∪ri=1 fi are f1 , . . . , fr . We let im(C) be the number of edges in the largest induced matching.
The next result was shown in [53, Theorem 6.5] for the family of uniform properly-connected
hypergraphs.
Corollary 3.9. Let C be a clutter and let f1 , . . . , fr be an induced matching of C with di = |fi |
for i = 1, . . . , r. Then
(a) ( ri=1 di ) − r ≤ reg(R/I(C)).
P
(b) [74, Lemma 2.2] im(G) ≤ reg(R/I(G)) for any graph G.
9
Proof. Let D = C[∪ri=1 fi ]. Notice that I(D) = (xf1 , . . . , xfr ). Thus I(D) is a complete inter-
section and the regularity of R/I(D) is the degree of its h-polynomial. The Hilbert series of
R/I(D) is given by
Qr
(1 + t + · · · + tdi −1 )
HSD (t) = i=1 .
(1 − t)n−r
Thus, the degree of the h-polynomial equals ( ri=1 di ) − r. Therefore, part (a) follows from
P
Proposition 3.8. Part (b) follows from part (a).
Corollary 3.10. If C is a clutter and R/Ic (C) is Cohen-Macaulay, then im(C) = 1.
Proof. Let r be the induced matching number of C and let d be the cardinality of any edge of C.
Using Theorem 3.2 and Corollary 3.9, we obtain d − 1 ≥ r(d − 1). Thus r = 1, as required.
The following example shows that the inequality obtained in Corollary 3.9(b) can be strict.
Example 3.11. Let G be the complement of a cycle C6 = {x1 , . . . , x6 } of length six. The edge
ideal of G is
I(G) = (x1 x3 , x1 x5 , x1 x4 , x2 x6 , x2 x4 , x2 x5 , x3 x5 , x3 x6 , x4 x6 ).
Using Macaulay2 [49], we get reg(R/I(G)) = 2 and im(G) = 1.
Lemma 3.12. [28, Corollary 20.19] If 0 → N → M → L → 0 is a short exact sequence of
graded finitely generated R-modules, then
(a) reg(N ) ≤ max(reg(M ), reg(L) + 1).
(b) reg(M ) ≤ max(reg(N ), reg(L)).
(c) reg(L) ≤ max(reg(N ) − 1, reg(M )).
Definition 3.13. If x is a vertex of a graph G, then its neighbor set, denoted by NG (x), is the
set of vertices of G adjacent to x.
The following theorem gives a precise sense in which passing to induced subgraphs can be
used to bound the regularity. Recall that a discrete graph is one in which all the vertices are
isolated.
Theorem 3.14. Let F be a family of graphs containing any discrete graph and let β : F → N
be a function satisfying that β(G) = 0 for any discrete graph G, and such that given G ∈ F,
with E(G) 6= ∅, there is x ∈ V (G) such that the following two conditions hold:
(i) G \ {x} and G \ ({x} ∪ NG (x)) are in F.
(ii) β(G \ ({x} ∪ NG (x))) < β(G) and β(G \ {x}) ≤ β(G).
Then reg(R/I(G)) ≤ β(G) for any G ∈ F.
Proof. The proof is by induction on the number of vertices. Let G be a graph in F. If G

is a discrete graph, then I(G) = (0) and reg(R) = β(G) = 0. Assume that G has at least
one edge. There is a vertex x ∈ V (G) such that the induced subgraphs G1 = G \ {x} and
G2 = G \ ({x} ∪ NG (x)) satisfy (i) and (ii). There is an exact sequence of graded R-modules
x
0 −→ R/(I(G) : x)[−1] −→ R/I(G) −→ R/(x, I(G)) −→ 0.
Notice that (I(G) : x) = (NG (x), I(G2 )) and (x, I(G)) = (x, I(G1 )). The graphs G1 and G2 have
fewer vertices than G. It follows directly from the definition of regularity that reg(M [−1]) =
1 + reg(M ) for any graded R-module M . Therefore applying the induction hypothesis to G1
and G2 , and using conditions (i) and (ii) and Lemma 3.5, we get
reg(R/(I(G) : x)[−1]) = reg(R/(I(G) : x)) + 1 = reg(R′ /I(G2 )) + 1 ≤ β(G2 ) + 1 ≤ β(G),
reg(R/(x, I(G))) ≤ β(G1 ) ≤ β(G)
where R′ is the ring in the variables V (G2 ). Therefore from Lemma 3.12, we get that the
regularity of R/I(G) is bounded by the maximum of the regularities of R/(I(G) : x)[−1] and
R/(x, I(G)). Thus reg(R/I(G)) ≤ β(G), as required.
As an example of how Theorem 3.14 can be applied to obtain combinatorial bounds for the
regularity, we provide new proofs for several previously known results. Let G be a graph. We let
β ′ (G) be the cardinality of any smallest maximal matching of G. Hà and Van Tuyl proved that
the regularity of R/I(G) is bounded from above by the matching number of G and Woodroofe
improved this result showing that β ′ (G) is an upper bound for the regularity.
Corollary 3.15. Let G be a graph and let R = K[V (G)]. Then
(a) [53, Corollary 6.9] reg(R/I(G)) = im(G) for any chordal graph G.
(b) ([53, Theorem 6.7], [117]) reg(R/I(G)) ≤ β ′ (G).
(c) [106, Theorem 3.3] reg(R/I(G)) = im(G) if G is bipartite and R/I(G) is sequentially
Cohen-Macaulay.
Proof. (a) Let F be the family of chordal graphs and let G be a chordal graph with E(G) 6= ∅.
By Corollary 3.9 and Theorem 3.14 it suffices to prove that there is x ∈ V (G) such that
im(G1 ) ≤ im(G) and im(G2 ) < im(G), where G1 and G2 are the subgraphs G \ {x} and
G \ ({x} ∪ NG (x)), respectively. The inequality im(G1 ) ≤ im(G) is clear because any induced
matching of G1 is an induced matching of G. We now show the other inequality. By [104,
Theorem 8.3], there is y ∈ V (G) such that G[NG (y) ∪ {y}] is a complete subgraph. Pick
x ∈ NG (y) and set f0 = {x, y}. Consider an induced matching f1 , . . . , fr of G2 with r = im(G2 ).
We claim that f0 , f1 , . . . , fr is an induced matching of G. Let e be an edge of G contained in
∪ri=0 fi . We may assume that e ∩ f0 6= ∅ and e ∩ fi 6= ∅ for some i ≥ 1, otherwise e = f0 or e = fi
for some i ≥ 1. Then e = {y, z} or e = {x, z} for some z ∈ fi . If e = {y, z}, then z ∈ NG (y)
and x ∈ NG (y). Hence {z, x} ∈ E(G) and z ∈ NG (x), a contradiction because the vertex set of
G2 is disjoint from NG (x) ∪ {x}. If e = {x, z}, then z ∈ NG (x), a contradiction. This completes
the proof of the claim. Hence im(G2 ) < im(G).
(b) Let F be the family of all graphs and let G be a graph with E(G) 6= ∅. By Theorem 3.14
it suffices to prove that there is x ∈ V (G) such that β ′ (G1 ) ≤ β ′ (G) and β ′ (G2 ) < β ′ (G), where
G1 and G2 are the subgraphs G \ {x} and G \ ({x} ∪ NG (x)), respectively.
Let f1 , . . . , fr be a maximal matching of G with r = β ′ (G) and let x, y be the vertices of
f1 . Clearly f2 , . . . , fr is a matching of G1 . Thus we can extend it to a maximal matching
f2 , . . . , fr , h1 , . . . , hs of G1 . Notice that s ≤ 1. Indeed if s ≥ 2, then hi ∩ f1 = ∅ for some
i ∈ {1, 2} (otherwise y ∈ h1 ∩ h2 , which is impossible). Hence f1 , . . . , fr , hi is a matching of G,
a contradiction because f1 , . . . , fr is maximal. Therefore β ′ (G1 ) ≤ r − 1 + s ≤ β ′ (G).
The set f2 , . . . , fr contains a matching of G2 , namely those edges fi that do not degenerate.
Reorder the edges so that f2 , . . . , fm are the edges that do not degenerate. Then this set can
be extended to a maximal matching f2 , . . . , fm , fm+1 ′ , . . . , fk′ of G2 . Now consider fm+1
′ . Since
′
f1 , . . . , fr is a maximal matching of G, fm+1 has a nontrivial intersection with fi for some i.
Note that i 6= 1 since fm+1 ′ ′
is an edge of G2 , and i ≥ m + 1 since f2 , . . . , fm and fm+1 are all
part of a matching of G2 . Reorder so that i = m + 1. Repeat the process with fm+2 . As before, ′
11
′
fm+2 ′
has a nontrivial intersection with fi for some i ≥ m + 1. If i = m + 1, then since fm+1 and
′
fm+2 are disjoint, each must share a different vertex with fm+1 . But fm+1 degenerated when
passing to G2 , and G2 is induced on the remaining vertices, so this is a contradiction. Thus we
may reorder so that fm+2′ nontrivially intersects fm+2 . Repeating this process, we see that fj′
nontrivially intersects fj for all m+1 ≤ j ≤ k. Thus k ≤ r. Now β ′ (G2 ) ≤ k −1 ≤ r −1 < β ′ (G).
(c) Let F be the family of all bipartite graphs G such that R/I(G) is sequentially Cohen-
Macaulay, and let β : F → N be the function β(G) = im(G). Let G be a graph in F with
E(G) 6= ∅. By Corollary 3.9 and Theorem 3.14 it suffices to observe that, according to [107,
Corollary 2.10], there are adjacent vertices x and y with deg(y) = 1 such that the bipartite graphs
G \ ({x} ∪ NG (x)) and G \ ({y} ∪ NG (y)) are sequentially Cohen-Macaulay. Thus conditions (i)
and (ii) of Theorem 3.14 are satisfied.
Corollary 3.15 shows that the regularity of R/I(G) equals im(G) for any forest G, which was
first proved by Zheng [118]. If G is an unmixed graph, Kummini [77] showed that reg(R/I(G))
equals the induced matching number of G. If G is claw-free and its complement has no induced
4-cycles, then reg(R/I(G)) ≤ 2 with equality if its complement is not chordal [90] (note that in
this case reg(R/I(G)) = im(G) + 1). Formulas for the regularity of ideals of mixed products are
given in [71]. The regularity and depth of lex segment edge ideals are computed in [33]. The
regularity and other algebraic properties of the edge ideal I(G) associated to a Ferrers graph
G are studied in detail in [18]. If R/I(G) is Cohen-Macaulay, then a formula for reg(R/I(G))
follows from [95, Corollary 4.2].
The following result about regularity was shown by Kalai and Meshulam for square-free
monomial ideals and by Herzog for arbitrary monomial ideals. Similar inequalities hold for the
projective dimension.
Proposition 3.16. [73, 57] Let I1 and I2 be monomial ideals of R. Then
(a) reg R/(I1 + I2 ) ≤ reg(R/I1 ) + reg(R/I2 ),
(b) reg R/(I1 ∩ I2 ) ≤ reg(R/I1 ) + reg(R/I2 ) + 1.
Corollary 3.17. If C1 , . . . , Cs are clutters on the vertex set X, then
reg(R/I(∪si=1 Ci )) ≤ reg(R/I(C1 )) + · · · + reg(R/I(Cs )).
C = ∪si=1 Ci equals ∪si=1 E(Ci ). By Proposition 3.16, it suffices to notice

Proof. The set of edges ofP
the equality I(∪i=1 Ci ) = si=1 I(Ci ).
s
A clutter C is called co-CM if Ic (C) is Cohen-Macaulay. A co-CM clutter is uniform because

Cohen-Macaulay clutters are unmixed.
Corollary 3.18. If C1 , . . . , Cs are co-CM clutters on the vertex set X, then
reg(R/I(∪si=1 Ci )) ≤ (d1 − 1) + · · · + (ds − 1),
where di is the number of elements in any edge of Ci .
Proof. By Theorem 3.2, we get that reg R/I(Ci ) = di − 1 for all i. Thus the result follows from
Corollary 3.17.
This result is especially useful for graphs. A graph G is weakly chordal if every induced cycle
in both G and Gc has length at most 4. It was pointed out in [117] that a weakly chordal graph
G can be covered by im(G) co-CM graphs (this fact was shown in [13]). Thus we have:
Theorem 3.19. [117] If G is a weakly chordal graph, then reg(R/I(G)) = im(G).
There are bounds for the regularity of R/I in terms of some other algebraic invariants of R/I.
Recall that the a-invariant of R/I, denoted by a(R/I), is the degree (as a rational function) of
the Hilbert series of R/I. Also recall that the independence complex of I(C), denoted by ∆C , is
the simplicial complex whose faces are the independent vertex sets of C. The arithmetic degree
of I = I(C), denoted by arith-deg(I), is the number of facets (maximal faces with respect to
inclusion) of ∆C . The arithmetical rank of I, denoted by ara(I), is the least number of elements
of R which generate the ideal I up to radical.
Theorem 3.20. [108, Corollary B.4.1] a(R/I) ≤ reg(R/I) − depth(R/I), with equality if R/I
is Cohen-Macaulay.
Theorem 3.21. ([79], [80, Proposition 3]) reg(R/I ∨ ) = pd(R/I) − 1 ≤ ara(I) − 1.
The equality reg(R/I ∨ ) = pd(R/I) − 1 was pointed out earlier in Theorem 3.2. There are
many instances where the equality pd(R/I) = ara(I) holds, see [1, 33, 76] and the references
there. For example, for paths, one has pd(R/I) = ara(I) [1]. Barile [1] has conjectured that the
equality holds for edge ideals of forests. We also have that n − mini {depth(R/I (i) )} is an upper
bound for ara(I), see [80]. This upper bound tends to be very loose. If I is the edge ideal of a
tree, then I is normally torsion free (see Section 4 together with Theorems 4.34 and 4.8). Then
mini {depth(R/I (i) )} = 1 by [87, Lemma 2.6]. But when I is the edge ideal of a path with 8
vertices, then the actual value of ara(I) is 5.
Theorem 3.22. [103, Theorem 3.1] If ht(I) ≥ 2, then reg(I) ≤ arith-deg(I).
The next open problem is known as the Eisenbud-Goto regularity conjecture [30].
Conjecture 3.23. If p ⊂ (x1 , . . . , xn )2 is a prime graded ideal, then
reg(R/p) ≤ deg(R/p) − codim(R/p).
A pure d-dimensional complex ∆ is called connected in codimension 1 if each pair of facets F, G
can be connected by a sequence of facets F = F0 , F1 , . . . , Fs = G, such that dim(Fi−1 ∩Fi ) = d−1
for 1 ≤ i ≤ s. According to [5, Proposition 11.7], every Cohen-Macaulay complex is connected
in codimension 1.
The following gives a partial answer to the monomial version of the Eisenbud-Goto regularity
conjecture.
Theorem 3.24. [103] Let I = I(C) be an edge ideal. If ∆C is connected in codimension 1, then
reg(R/I) ≤ deg(R/I) − codim(R/I).
The dual notion to the independence complex of I(C) is to start with a complex ∆ and
associate to it an ideal whose independence complex is ∆.
Definition 3.25. Given a simplicial complex ∆ with vertex set X = {x1 , . . . , xn }, the Stanley-
Reisner ideal of ∆ is defined as
I∆ = ({xi1 · · · xir | i1 < · · · < ir , {xi1 , . . . , xir } ∈
/ ∆}) ,
and its Stanley-Reisner ring K[∆] is defined as the quotient ring R/I∆ .
A simple proof the next result is given in [44].
Theorem 3.26. [101] Let C be a clutter and let ∆ = ∆C be its independence complex. Then
depth R/I(C) = 1 + max{i | K[∆i ] is Cohen-Macaulay},
where ∆i = {F ∈ ∆ | dim(F ) ≤ i} is the i-skeleton of ∆ and −1 ≤ i ≤ dim(∆).
13
A variation on the concept of the i-skeleton will facilitate an extension of the result above to
the sequentially Cohen-Macaulay case.
Definition 3.27. Let ∆ be a simplicial complex. The pure i-skeleton of ∆ is defined as:
∆[i] = h{F ∈ ∆| dim(F ) = i}i; −1 ≤ i ≤ dim(∆),
where hFi denotes the subcomplex generated by F.
Note that ∆[i] is always pure of dimension i. We say that a simplicial complex ∆ is sequentially
Cohen-Macaulay if its Stanley-Reisner ring has this property. The following results link the
sequentially Cohen-Macaulay property to the Cohen-Macaulay property and to the regularity
and projective dimension. The first is an interesting result of Duval.
Theorem 3.28. [26, Theorem 3.3] A simplicial complex ∆ is sequentially Cohen-Macaulay if
and only if the pure i-skeleton ∆[i] is Cohen-Macaulay for −1 ≤ i ≤ dim(∆).
Corollary 3.29. R/I(C) is Cohen-Macaulay if and only if R/I(C) is sequentially Cohen-Macaulay
and C is unmixed.
Lemma 3.30. Let C be a clutter and let ∆ = ∆C be its independence complex. If β0′ (C) is the
cardinality of a smallest maximal independent set of C, then ∆[i] = ∆i for i ≤ β0′ (C) − 1.
Proof. First we prove the inclusion ∆[i] ⊂ ∆i . Let F be a face of ∆[i] . Then F is contained in a
face of ∆ of dimension i, and so F is in ∆i . Conversely, let F be a face of ∆i . Then
dim(F ) ≤ i ≤ β0′ (C) − 1 =⇒ |F | ≤ i + 1 ≤ β0′ (C).
Since β0′ (C) is the cardinality of any smallest maximal independent set of C, we can extend F to
an independent set of C with i + 1 vertices. Thus F is in ∆[i] .
While β0′ regulates the equality of the i-skeleton and the pure i-skeleton of the independence
complex, its complement provides a lower bound for the regularity of the ideal of covers.
Theorem 3.31. Let C be a clutter, let Ic (C) be its ideal of vertex covers, and let α′0 (C) be the
cardinality of a largest minimal vertex cover of C. Then
reg R/Ic (C) ≥ α′0 (C) − 1,
with equality if R/I(C) is sequentially Cohen-Macaulay.
Proof. We set β0′ (C) = n − α′0 (C). Using Theorem 3.2 and the Auslander-Buchsbaum formula
(see Equation (3.2)), the proof reduces to showing: depth R/I(C) ≤ β0′ (C), with equality if
R/I(C) is sequentially Cohen-Macaulay.
First we show that depth R/I(C) ≤ β0′ (C). Assume ∆i is Cohen-Macaulay for some −1 ≤ i ≤
dim(∆), where ∆ is the independence complex of C. According to Theorem 3.26, it suffices to
prove that 1+i ≤ β0′ (C). Notice that β0′ (C) is the cardinality of any smallest maximal independent
set of C. Thus, we can pick a maximal independent set F of C with β0′ (C) vertices. Since ∆i is
Cohen-Macaulay, the complex ∆i is pure, that is, all maximal faces of ∆ have dimension i. If
1 + i > β0′ (C), then F is a maximal face of ∆i of dimension β0′ (C) − 1, a contradiction to the
purity of ∆i .
Assume that R/I(C) is sequentially Cohen-Macaulay. By Lemma 3.30 ∆[i] = ∆i for i ≤
β0 (C)−1. Then by Theorem 3.28, the ring K[∆i ] is Cohen-Macaulay for i ≤ β0′ (C)−1. Therefore,
′
applying Theorem 3.26, we get that the depth of R/I(C) is at least β0′ (C). Consequently, in this
case one has the equality depth R/I(C) = β0′ (C).
The inequality in Theorem 3.31 also follows directly from the definition of regularity because
reg(Ic (C)) is an upper bound for the largest degree of a minimal generator of Ic (C).
Remark 3.32. α′0 (C) is max{|e| : e ∈ E(C ∨ )} and α′0 (C ∨ ) is max{|e| : e ∈ E(C)}. This follows
by Alexander duality, see Eq. (3.3).
Corollary 3.33. If I(C) is an edge ideal, then pdR (R/I(C)) ≥ α′0 (C), with equality if R/I(C) is
sequentially Cohen-Macaulay.
Proof. It follows from the proof of Theorem 3.31.
There are many interesting classes of sequentially Cohen-Macaulay clutters where this formula
for the projective dimension applies (see Theorem 2.6). The projective dimension of edge ideals
of forests was studied in [22, 53], where some recursive formulas are presented. Explicit formulas
for the projective dimension for some path ideals of directed rooted trees can be found in [55,
Theorem 1.2]. Path ideals of directed graphs were introduced by Conca and De Negri [15]. Fix
an integer t ≥ 2, and suppose that D is a directed graph, i.e., each edge has been assigned a
direction. A sequence of t vertices xi1 , . . . , xit is said to be a path of length t if there are t − 1
distinct edges e1 , . . . , et−1 such that ej = (xij , xij+1 ) is a directed edge from xij to xij+1 . The
path ideal of D of length t, denoted by It (D), is the ideal generated by all monomials xi1 · · · xit
such that xi1 , . . . , xit is a path of length t in D. Note that when t = 2, then I2 (D) is simply the
edge ideal of D.
Example 3.34. Let K be any field and let G be the following chordal graph
x xr 10
xr7 r 8 rx11
x6 @ x5 x9 r r
x12
@
r r
@
@
x4 r r @r r x14
x 1 x13 @r
r x15
r
x3 r
x2 x16
Then, by Theorem 2.6 and Corollary 3.31, we get pdR (R/I(G)) = 6 and depth R/I(G) = 10.
Corollary 3.35. Let C be a clutter. If I(C) has linear quotients, then
reg R/I(C) = max{|e| : e ∈ E(C)} − 1.
Proof. The ideal of covers Ic (C) is sequentially Cohen-Macaulay by Theorem 2.6(d). Hence,
using Theorem 3.31, we get reg(R/I(C)) = α′0 (C ∨ ) − 1. To complete the proof notice that
α′0 (C ∨ ) = max{|e| : e ∈ E(C)} (see Remark 3.32).
The converse of Theorem 3.33 is not true.

Example 3.36. Let C6 be a cycle of length 6. Then R/I(C6 ) is not sequentially Cohen-Macaulay
by Proposition 2.4. Using Macaulay2, we get pd(R/I(C6 )) = α′0 (C6 ) = 4.
When R/I is not known to be Cohen-Macaulay, it can prove useful to have effective bounds
on the depth of R/I.
Theorem 3.37. Let G be a bipartite graph without isolated vertices. If G has n vertices, then
jnk
depth R/I(G) ≤ .
2
15
Proof. Let (V1 , V2 ) be a bipartition of G with |V1 | ≤ |V2 |. Note 2|V1 | ≤ n because |V1 | + |V2 | = n.
Since V1 is a maximal independent set of vertices one has β0′ (G) ≤ |V1 | ≤ n/2. Therefore, using
Corollary 3.33 and the Auslander-Buchsbaum formula, we get depth R/I(G) ≤ n/2.
n
Corollary 3.38. If G is a B-graph with n vertices, then depth R/I(G) ≤ dim R/I(G) ≤ 2 .
Proof. Recall that n = α0 (G) + β0 (G). By Theorem 2.9, β0 (G) ≤ α0 (G), and so β0 (G) ≤ n2 .

The result now follows because β0 (G) = dim R/I(C).
Lower bounds are given in [87] for the depths of R/I(G)t for t ≥ 1 when I(G) is the edge
ideal of a tree or forest. Upper bounds for the depth of R/I(G) are given in [46, Corollary 4.15]
when G is any graph without isolated vertices. The depth and the Cohen-Macaulay property of
ideals of mixed products is studied in [71].
We close this section with an upper bound for the multiplicity of edge rings. Let C be a
clutter. The multiplicity of the edge-ring R/I(C), denoted by e(R/I(C)), equals the number of
faces of maximum dimension of the independence complex ∆C , i.e., the multiplicity of R/I(C)
equals the number of independent sets of C with β0 (C) vertices. A related invariant that was
considered earlier is arith-deg(I(C)), the number of maximal independent sets of C.
Proposition 3.39. [46] If C is a d-uniform clutter and I = I(C), then e(R/I) ≤ dα0 (C) .
4. Stability of associated primes
One method of gathering information about an ideal is through its associated primes. Let I
be an ideal of a ring R. In this section, we will examine the sets of associated primes of powers
of I, that is, the sets
Ass(R/I t ) = {p ⊂ R | p is prime and p = (I t : c) for some c ∈ R}.
When I is a monomial ideal of a polynomial ring R = K[x1 , . . . , xn ], the associated primes will
be monomial ideals, that is, prime ideals which are generated by a subset of the variables. When
I is a square-free monomial ideal, the minimal primes of I, Min(R/I), correspond to minimal
vertex covers of the clutter C associated to I. In general Min(R/I) ⊂ Ass(R/I t ) for all positive
integers t. For a square-free monomial ideal, in the case where equality holds for all t, the ideal I
is said to be normally torsion-free. More generally, an ideal I ⊂ R is called normally torsion-free
if Ass(R/I i ) is contained in Ass(R/I) for all i ≥ 1 and I 6= R.
In [9], Brodmann showed that when R is a Noetherian ring and I is an ideal of R, the sets
Ass(R/I t ) stabilize for large t. That is, there exists a positive integer N such that Ass(R/I t ) =
Ass(R/I N ) for all t ≥ N . We will refer to a minimal such N as the index of stability of I.
There are two natural questions following from this result. In this article, we will focus on the
monomial versions of the questions.
Question 4.1. Given a monomial ideal I, what is an effective upper bound on the index of
stability for a given class of monomial ideals?
Question 4.2. Given a monomial ideal I, which primes are in Ass(R/I t ) for all sufficiently
large t?
An interesting variation on Questions 4.1 and 4.2 was posed in [99].

Question 4.3. Suppose that N is the index of stability of an ideal I. Given a prime p ∈
Ass(R/I N ), can you find an integer Np for which p ∈ Ass(R/I t ) for t ≥ Np ?
Brodmann also showed that the sets Ass(I t−1 /I t ) stabilize. Thus in the general setting, one
could ask similar questions about these sets. However, for monomial ideals the following lemma
shows that in order to find information about Ass(R/I t ), one may instead study Ass(I t−1 /I t ).
Lemma 4.4. Let I be a monomial ideal. Then Ass(I t−1 /I t ) = Ass(R/I t ).
Proof. Suppose that p ∈ Ass(R/I t ). Then p = (I t : c) for some monomial c ∈ R. But since p
is necessarily a monomial prime, generated by a subset of the variables, then if xc ∈ I t for a
variable x ∈ p, then c ∈ I t−1 and so p ∈ Ass(I t−1 /I t ). The other inclusion is automatic.
Note that this method was used in [105] to show that the corresponding equality also holds
for the integral closures of the powers of I.
For special classes of ideals, there have been some results that use properties of the ideals
to find bounds on N . For example, if I is generated by a regular sequence, then by [67], I
is normally torsion-free, or Ass(R/I t ) = Min(R/I) for all t, and thus N = 1. If instead I is
generated by a d-sequence and is strongly Cohen-Macaulay, then it was shown in [88] that N
is bounded above by the dimension of the ring. In particular, N ≤ n − g + 1 where n is the
dimension of the ring and g is the height of the ideal. We are particularly interested in finding
similar bounds for classes of monomial ideals.
In [65], Hoa used integer programming techniques to give an upper bound on N for general
monomial ideals. Let n be the number of variables, s the number of generators of I, and d the
maximal degree of a generator.
Theorem 4.5. [65, Theorem 2.12] If I is a monomial ideal, then the index of stability is bounded
above by
n √ √ 2
o
max d(ns + s + d)( n)n+1 ( 2d)(n+1)(s−1) , s(s + n)4 sn+2 d2 (2d2 )s −s+1 .
Notice that this bound can be extremely large. For general monomial ideals, examples are
given in [65] to show that the bound should depend on d and n. However, if we restrict to special
classes of monomial ideals, much smaller bounds can be found. For example, an alternate bound
has been shown to hold for integral closures of powers of monomial ideals.
Theorem 4.6. [105, Theorem 16] If I is a monomial ideal, and N0 = n2n−1 dn−2 , then
Ass(R/I t ) = Ass(R/I N0 ) for t ≥ N0 when n ≥ 2.
Here again n is the number of variables and d is the maximal degree of a generator. For the
class of normal monomial ideals, this bound on the index of stability can be significantly lower
than the general bound given above. When n = 2, the index of stability of the integral closures
is lower still.
Lemma 4.7. [85] If n ≤ 2, then Ass(R/I t ) = Ass(R/I) for all t ≥ 1.
Note that this result is of interest for general monomial ideals; however, when n = 2 a square-
free monomial ideal will be a complete intersection. Of particular interest for this article are
results that use combinatorial and graph-theoretic properties to yield insights into the associated
primes and index of stability of monomial ideals. One pivotal result in this area establishes a
classification of all graphs for which N = 1.
Theorem 4.8. [100, Theorem 5.9] Let G be a graph and I its edge ideal. Then G is bipartite if
and only if I is normally torsion-free.
17
The result above shows that N = 1 for the edge ideal of a graph if and only if the graph is
bipartite. Since minimal primes correspond to minimal vertex covers, this completely answers
Questions 4.1 and 4.2 for bipartite graphs. In addition if I is the edge ideal of a balanced clutter,
then N = 1 [45].
Suppose now that G is a graph that is not bipartite. Then G contains at least one odd cycle.
For such graphs, a method of describing embedded associated primes, and a bound on where the
stability occurs, were given in [14]. The method of building embedded primes centered around
the odd cycles, so we first give an alternate proof of the description of the associated primes for
this base case.
Lemma 4.9. Suppose G is a cycle of length n = 2k + 1 and I is the edge ideal of G. Then
Ass(R/I t ) = Min(R/I) if t ≤ k and Ass(R/I t ) = Min(R/I) ∪ {m} if t ≥ k + 1. Moreover, when
t ≥ k + 1, m = (I t : c) for a monomial c of degree 2t − 1.
Proof. If p 6= m is a prime ideal, then Ip is the edge ideal of a bipartite graph, and thus by
Theorem 4.8 p ∈ Ass(R/I t ) (for any t ≥ 1) if and only if p is a minimal prime of I. Notice also
that the deletion of any vertex xi (which corresponds to passing to the quotient ring R/(xi ))
results in a bipartite graph as well. Thus by [51, CorollaryQ 3.6], m 6∈ Ass(R/I t ) for t ≤ k since a
maximal matching has k edges. For t ≥ k + 1, define b = ( ni=1 xi ) and c = b(x1 x2 )t−k−1 where
x1 x2 is any edge of G. Then since c has degree 2t − 1, c 6∈ I t , but G is a cycle, xi b ∈ I k+1 and
xi c ∈ I t . Thus m = (I t : c) and so m ∈ Ass(R/I t ) for t ≥ k + 1.
Corollary 4.10. Suppose G is a connected graph containing an odd cycle of length 2k + 1 and
suppose that every vertex of G that is not in the cycle is a leaf. Then Ass(R/I t ) = Min(R/I) ∪
{m} if t ≥ k + 1. Moreover, when t ≥ k + 1, m = (I t : c) for a monomial c of degree 2t − 1.
Proof. Let b and c be defined as in the proof of Lemma 4.9. Notice that if x is a leaf, then x
is connected to a unique vertex in the cycle and that xb ∈ I k+1 . The remainder of the proof
follows as in Lemma 4.9.
If G is a more general graph, the embedded associated primes of I = I(G) are formed by
working outward from the odd cycles. This was done in [14], including a detailed explanation
of how to work outward from multiple odd cycles. Before providing more concise proofs of
the process, we first give an informal, but illustrative, description. Suppose C is a cycle with
2k + 1 vertices x1 , . . . , x2k+1 . Color the vertices of C red and color any noncolored vertex that is
adjacent to a red vertex blue. The set of colored vertices, together with a minimal vertex cover
of the set of edges neither of whose vertices is colored, will be an embedded associated prime of
I t for all t ≥ k + 1. To find additional embedded primes of higher powers, select any blue vertex
to turn red and turn any uncolored neighbors of this vertex blue. The set of colored vertices,
together with a minimal vertex cover of the noncolored edges, will be an embedded associated
prime of I t for all t ≥ k + 2. This process continues until all vertices are colored red or blue.
The method of building new associated primes for a power of I from primes associated to
lower powers relies on localization. Since localization will generally cause the graph (or clutter)
to become disconnected, we first need the following lemma.
Lemma 4.11. ([51, Lemma 3.4], see also [14, Lemma 2.1]) Suppose I is a square-free monomial
ideal in S = K[x1 , . . . , xr , y1 , . . . , ys ] such that I = I1 S + I2 S, where I1 ⊂ S1 = K[x1 , . . . , xr ]
and I2 ⊂ S2 = K[y1 , . . . , ys ]. Then p ∈ Ass(S/I t ) if and only if p = p1 S + p2 S, where p1 ∈
Ass(S1 /I1t1 ) and p2 ∈ Ass(S2 /I2t2 ) with (t1 − 1) + (t2 − 1) = t − 1.
Note that this lemma easily generalizes to an ideal I = (I1 , I2 , . . . , Is ) where the Ii are edge
ideals of disjoint clutters. Then p ∈ Ass(R/I t ) if and only if p = (p1 , . . . , ps ) with pi ∈ Ass(R/Iiti )
where (t1 − 1) + (t2 − 1) + · · · + (ts − 1) = (t − 1).
We now fix a notation to show how to build embedded associated primes. Consider p ∈
Ass(R/I t ) for I the edge ideal of a graph G. Without loss of generality, Lemma 4.11 allows
us to assume G does not have isolated vertices. If p 6= m, then since p ∈ Ass(R/I t ) ⇔ pRp ∈
Ass(Rp /(Ip )t ), consider Ip . Write Ip = (Ia , Ib ) where Ia is generated by all generators of Ip of
degree two and Ib is the prime ideal generated by the degree one generators of Ip , which corre-
spond to the isolated vertices of the graph associated to Ip . Note that the graph corresponding
to Ia need not be connected. If Ia = (0), then p is a minimal prime of I, so assume Ia 6= (0).
Define pa to be the monomial prime generated by variables of Ia . Define N1 = ∪x∈pa N (x),
where N (x) is the neighbor set of x in G, and let p1 = pa ∪ N1 . Notice that if x ∈ pa , then
x is not isolated in Ip , so N1 ⊂ p and thus N1 ⊂ pa ∪ Ib . Define p2 = p\p1 = Ib \N1 , and
N2 = ∪x∈p1 N (x)\p. If G1 is the induced subgraph of G on the vertices in p1 ∪ N2 , G2 is the
induced subgraph of G on vertices in V \p1 , and Ii = I(Gi ) for i = 1, 2, then Ip = ((I1 )p1 , (I2 )p2 )
and p2 is a minimal vertex cover of I2 . By design, any vertex appearing in both G1 and G2 is
not in p, and thus (I1 )p1 and (I2 )p2 do not share a vertex. Thus by Lemma 4.11 and the fact
that associated primes localize, p ∈ Ass(Rp /(Ip )t ) if and only if p1 ∈ Ass(R1 /(I1 )t ) and p1 is the
maximal ideal of R1 = K[x | x ∈ p1 ]. For convenience, define Ra = K[x | x ∈ pa ].
Proposition 4.12. Let p ∈ Ass(R/I t ). Using the notation from above, assume p1 = (I1t : c) for
some monomial c ∈ Ra of degree at most 2t − 1. Let x ∈ p1 . Let p′1 = p1 ∪ N (x), and let p′2
be any minimal vertex cover of the edges of G′2 where G′2 is the induced subgraph of G on the
vertices V \p′1 . Let N2′ = ∪x∈p′1 N (x)\(p′1 ∪ p′2 ). Let G′1 be the induced subgraph of G with vertices
in p′1 ∪ N2′ . Then p′ = (p′1 , p′2 ) ∈ Ass(R/I t+1 ).
Proof. If v is an isolated vertex of G′2 then N (v) ⊂ p′1 and thus every edge of G containing
v is covered by p′1 . Hence p′ is a vertex cover of G. Since x ∈ p1 , there is an edge xy ∈ G1
with y ∈ pa . Consider c′ = cxy. Then the degree of c′ is at most 2t + 1, so c′ 6∈ (I1 )t+1 . If
I1′ = I(G′1 ), then c′ 6∈ (I1′ )t+1 as well. If z ∈ p1 , then z(cxy) = (zc)(xy) ∈ (I1′ )t+1 . If z ∈ N (x),
then z(cxy) = (cy)(zx) ∈ (I1′ )t+1 since y ∈ p1 . Thus p′1 ⊂ ((I1′ )t+1 : c′ ). Suppose z 6∈ p′1 is a
vertex of G′1 . Then z ∈ N2′ . Then z 6∈ N (x) and z 6∈ N (y) since y ∈ pa , so zx and zy are not
edges of G′1 . Also z 6∈ p1 and c ∈ Ra , so zc 6∈ (I1′ )t+1 . Thus the inclusion must be an equality.
Since p′2 is a minimal vertex cover of the edges of G′2 , then p′2 ∈ Ass(R/I2′ ) where I2′ is the edge
ideal of G′2 (where isolated vertices of G′2 are not included in I2′ ). Note that Ip′ = ((I1′ )p′1 , (I2′ )p′2 )
and so the result follows from Lemma 4.11.
Note that if G contains an odd cycle of length 2k + 1, then embedded associated primes
satisfying the hypotheses of Proposition 4.12 exist for t ≥ k+1 by Lemma 4.9 and Corollary 4.10.
Starting with an induced odd cycle C one can now recover all the primes described in [14,
Theorem 3.3]. In addition, combining Corollary 4.10 with Lemma 4.11 as a starting place for
Proposition 4.12 recovers the result from [14, Theorem 3.7] as well. Define Ass(R/I t )∗ to be the
set of embedded associate primes of I t produced in Proposition 4.12 by starting from any odd
cycle, or collection of odd cycles, of the graph. Then Ass(R/I t )∗ ⊂ Ass(R/I s ) for all s ≥ t. To
see this, recall that if p is not a minimal prime, then there is a vertex x such that x ∪ N (x) ⊂ p.
Choosing such an x results in p1 = p′1 and the process shows that p ∈ Ass(R/I t+1 ). Notice also
that the sets Ass(R/I t )∗ stabilize. In particular, Ass(R/I t )∗ = Ass(R/I n )∗ for all t ≥ n where
n is the number of variables. Notice that choosing x ∈ N1 each time will eventually result in
19
m ∈ Ass(R/I t )∗ for some t. Counting the maximal number of steps this could take provides a
bound on the index of stability. Following the process above for a particular graph can often
yield a significantly lower power M for which Ass(R/I t )∗ stabilize. These results are collected
below.
Theorem 4.13. Let I be the edge ideal of a connected graph G that is not bipartite. Suppose G
has n vertices and s leaves, and N is the index of stability of I.
(a) [14, Theorem 4.1] The process used in Proposition 4.12 produces all embedded associated
primes in the stable set. That is, Ass(R/I N ) = Min(R/I) ∪ Ass(R/I N )∗ .
(b) [14, Corollary 4.3], (Proposition 4.12) If the smallest odd cycle of G has length 2k + 1,
then N ≤ n − k − s.
(c) [14, Theorem 5.6, Corollary 5.7] If G has a unique odd cycle, then Ass(R/I t ) = Min(R/I)∪
Ass(R/I t )∗ for all t. Moreover, the sets Ass(R/I t ) form an ascending chain.
(d) If p ∈ Ass(R/I N ), and N0 is the smallest positive integer for which p ∈ Ass(R/I N0 )∗ ,
then p ∈ Ass(R/I t ) for all t ≥ N0 .
To interpret Theorem 4.13 in light of our earlier questions, notice that (a) answers Ques-
tion 4.2, (b) answers Question 4.1, and (d) provides a good upper bound for Np in Question 4.3.
The significance of (c) is to answer a fourth question of interest. Before presenting that question,
we first discuss some extensions of the above results to graphs containing loops.
Corollary 4.14. Let I be a monomial ideal, not necessarily square-free, such that the generators
of I have degree at most two. Define Ass(R/I t )∗ to be the set of embedded associate primes of
I t produced in Proposition 4.12 by starting from any odd cycle, or collection of odd cycles where
generators of I that are not square-free are considered to be cycles of length one. Then the results
of Theorem 4.13 hold for I.
Proof. If I has generators of degree one, then write I = (I1 , I2 ) where I2 is generated in degree
one. Then I2 is a complete intersection, so by using Lemma 4.11 we may replace I by I1 and
assume I is generated in degree two. If I is not square-free, consider a generator x2 ∈ I. This
generator can be represented as a loop (cycle of length one) in the graph. Define pa = (x) and
N1 = N (x). Note that p1 = pa ∪ N1 = (I1 : c) where I1 is the induced graph on x ∪ N (x) and
c = x. Then p1 satisfies the hypotheses of Proposition 4.12. The results now follow from the
proof of Proposition 4.12.
Notice that ideals that are not square-free will generally have embedded primes starting with
t = 1 since the smallest odd cycle has length 1 = 2(0) + 1, so k + 1 = 1. The above corollary
can be extended to allow for any pure powers of variables to be generators of the ideal I.
Corollary 4.15. Let I = (I1 , I2 ) where I2 is the edge ideal of a graph G and I1 = (xsi11 , . . . , xsirr )
for any powers sj ≥ 1. Then the results of Theorem 4.13 hold for I with Ass(R/I t )∗ defined as
in Corollary 4.14.
Proof. As before, we may assume xj ≥ 2 for all j. Let K = (x2i1 , . . . , x2ir ) and let J = (K, I2 ).
Then J satisfies the hypotheses of Corollary 4.14. Let p ∈ Ass(R/J t )∗ be formed by starting
with pa = (xi1 ) and let p1 = (J1t : c) where J1 and c are defined as in Corollary 4.14. Suppose
xi1 , . . . , xiv ∈ p1 . Let qj ≥ 0 be the least integers such that c′ = xqi11 · · · xqivv · c 6∈ I1t . Then it
is straightforward to check that p1 = (I1t : c′ ) and so p ∈ Ass(R/I t )∗ . Thus higher powers of
variables can also be treated as loops and the results of Theorem 4.13 hold.
We now return to Theorem 4.13 (c). In general, the sets Ass(R/I t )∗ form an ascending chain.
Theorem 4.13 (c) gives a class of graphs for which Ass(R/I t )∗ describe every embedded prime
of a power of I t optimally. Thus Ass(R/I t ) will form a chain. This happens for many classes of
monomial ideals, and leads to the fourth question.
Question 4.16. If I is a square-free monomial ideal, is Ass(R/I t ) ⊂ Ass(R/I t+1 ) for all t?
For monomial ideals, Question 4.16 is of interest for low powers of I. For sufficiently large
powers, the sets of associated primes are known to form an ascending chain, and a bound beyond
which the sets Ass(I t /I t+1 ) form a chain has been shown by multiple authors (see [99], [85]).
This bound depends on two graded algebras which encode information on the powers of I, and
which will prove useful in other results. The first is the Rees algebra R[It] of I, which is defined
by
R[It] = R ⊕ It ⊕ I 2 t2 ⊕ I 3 t3 ⊕ · · ·
and the second is the associated graded ring of I,
grI (R) = R/I ⊕ I/I 2 ⊕ I 2 /I 3 ⊕ · · · .
Notice that while the result is for Ass(I t /I t+1 ), for monomial ideals Ass(R/I t+1 ) will also form
a chain.
Theorem 4.17. [99, 85] Ass(I t /I t+1 ) is increasing for t > a0R[It]+ (grI (R)).
Note that square-free is essential in Question 4.16. Examples of monomial ideals for which the
associated primes do not form an ascending chain have been given in [59, 65]. Those examples
were designed for other purposes and so are more complex than what is needed here. A simple
example can be found by taking the product of consecutive edges of an odd cycle.
Example 4.18. Let I = (x1 x22 x3 , x2 x23 x4 , x3 x24 x5 , x4 x25 x1 , x5 x21 x2 ). If m = (x1 , x2 , x3 , x4 , x5 ),
then m ∈ Ass(R/I t ) for t = 1, 4, but m 6∈ Ass(R/I t ) for t = 2, 3.
The ideal in Example 4.18 can be viewed as multiplying adjacent edges in a 5-cycle to form
generators of I, and so has a simple combinatorial realization. A similar result holds for longer
odd cycles, where the maximal ideal is not the only associate prime to appear and disappear.
However, if instead I = (x1 x2 x3 , x2 x3 x4 , x3 x4 x5 , x4 x5 x1 , x5 x1 x2 ) is the path ideal of the penta-
gon, then Ass(R/I t ) = Min(R/I) ∪ {m} for t ≥ 2 and thus Ass(R/I t ) form an ascending chain
(see [51, Example 3.14]).
There are some interesting cases where associated primes are known to form ascending chains.
The first listed is quite general, but has applications to square-free monomial ideals.
Theorem 4.19. ([84, Proposition 3.9], see also [56, Proposition 16.3]) If R is a Noetherian ring,
then Ass(R/I t ) form an ascending chain.
In order to present the next class of ideals for which the associated primes are known to form
ascending chains, we first need some some background definitions.
Definition 4.20. Let G be a graph. A colouring of the vertices of G is an assignment of colours
to the vertices of G such that adjacent vertices have distinct colours. The chromatic number
of G is the minimal number of colours in a colouring of G. A graph is called perfect if for
every induced subgraph H, the chromatic number of H equals the size of the largest complete
subgraph of H.
21
An excellent reference for the theory of perfect graphs is the book of Golumbic [48]. Using
perfect graphs, we now give an example to show how Theorem 4.19 can be applied to classes of
square-free monomial ideals. An alternate proof appears in [41, Corollary 5.11].
Example 4.21. If I is the ideal of minimal vertex covers of a perfect graph, then Ass(R/I t )
form an ascending chain.
Proof. By [112, Theorem 2.10], R[It] is normal. Thus I t = I t for all t, so by Theorem 4.19,
Ass(R/I t ) form an ascending chain.
Similar results hold for other classes of monomial ideals for which R[It] is known to be normal.
For example, in [15, Corollary 4.2] it is shown that a path ideal of a rooted tree has a normal
Rees algebra. Thus by Theorem 4.19, Ass(R/I t ) form an ascending chain for such ideals. Note
that a path ideal can be viewed as the edge ideal of a carefully chosen uniform clutter.
It is interesting to compare the result of Example 4.21 to [41, Theorem 5.9], where it is shown
that if I is the ideal of minimal vertex covers of a perfect graph, then the set of primes associated
to any fixed power has the saturated chain property. Here Ass(R/I t ) has the saturated chain
property if for every p ∈ Ass(R/I t ), either p is minimal or there is a Q ( p with Q ∈ Ass(R/I t )
and height Q = height p − 1.
Notice that Theorem 4.13 shows that in the case of a graph with a unique odd cycle, Ques-
tion 4.16 has an affirmative answer. This result can be generalized to any graph containing a
leaf. First we need a slight variation of a previously known result.
Lemma 4.22. [88, Lemma 2.3] Suppose I = I(G) is the edge ideal of a graph and a ∈ I/I 2 is a
regular element of the associated graded ring grI (R). Then the sets Ass(R/I t ) form an ascending
chain. Moreover, Ass(R/I t ) = Ass(I t−1 /I t ) for all t ≥ 1.
Proof. Let a ∈ I/I 2 be a regular element of grI (R). Assume p ∈ Ass(I t /I t+1 ). Then there
is a c ∈ I t /I t+1 with p = (0 :R/I c). But then p = (0 :R/I ac), and a lives in degree one,
so p ∈ Ass(I t+1 /I t+2 ). So these sets form an ascending chain. Now the standard short exact
sequence
0 → I t /I t+1 → R/I t+1 → R/I t → 0
gives
Ass(I t /I t+1 ) ⊂ Ass(R/I t+1 ) ⊂ Ass(R/I t ) ∪ Ass(I t /I t+1 )
and the result follows by induction.
Proposition 4.23. Let G be a graph containing a leaf x and let I = I(G) be its edge ideal.
Then Ass(R/I t ) ⊂ Ass(R/I t+1 ) for all t. That is, the sets of associated primes of the powers of
I form an ascending chain.
Proof. Since x is a leaf of G, there is a unique generator e = xy ∈ I divisible by x. Let a

denote the image of e in I/I 2 . We claim a is a regular element of grI (R). To see this, it suffices
to show that if f e ∈ I t+1 for some t, then f ∈ I t . Since I is a monomial ideal and e is a
monomial, f e ∈ I t+1 if and only if every term of f e is in I t+1 . Thus we may assume f is a
monomial and f xy = e1 e2 · · · et+1 h for some edges ei of G and some monomial h. Suppose
f 6∈ I t . Then x divides ei for some i, say i = t + 1. Since x is a leaf, ei = xy and by cancellation
f = e1 · · · et h ∈ I t . Thus a is a regular element of grI (R) and by Lemma 4.22 the result
follows.
When extending the above results to more general square-free monomial ideals, one needs to
pass from graphs to clutters. An obstruction to extending the results is the lack of an analog to
Theorem 4.8 ([100, Theorem 5.9]). One possible analog appears as a conjecture of Conforti and
Cornuéjols, see [17, Conjecture 1.6], which we discuss later in this section. This conjecture is
stated in the language of combinatorial optimization. It says that a clutter C has the max-flow
min-cut (MFMC, see Definition 4.30) property if and only if C has the packing property. These
criterion for clutters have been shown in recent years to have algebraic translations [45] which
will be discussed in greater detail later in the section. An ideal satisfies the packing property if
the monomial grade of I (see Definition 4.29) is equal to the height of I and this same equality
holds for every minor of I [17, 45]. Here a minor is formed by either localizing at a collection of
variables, passing to the image of I in a quotient ring R/(xi1 , xi2 , . . . , xis ), or a combination of
the two. In [47, Corollary 3.14] and [62, Corollary 1.6], it was shown that C satisfies MFMC if
and only if the corresponding edge ideal I(C) is normally torsion-free. This allows the conjecture
to be restated (cf. [45, Conjecture 4.18]) as: if C has the packing property, then I(C) is normally
torsion-free.
Since a proof of this conjecture does not yet exist, the techniques used to describe the embed-
ded associated primes, the stable set of associated primes, and the index of stability for graphs
are difficult to extend. However some partial results are known. The first gives some conditions
under which it is known that the maximal ideal is, or is not, an associated prime. In special
cases, this can provide a seed for additional embedded associated primes using techniques such
as those in Proposition 4.12.
Theorem 4.24. If I is a square-free monomial ideal, every proper minor of I is normally
torsion-free, and β1 is the monomial grade of I, then
(a) [51, Corollary3.6] m 6∈ Ass(R/I t ) for t ≤ β1 .
(b) [51, Theorem 4.6] If I fails the packing property, then m ∈ Ass(R/I β1 +1 ).
(c) [51, Proposition 3.9] If I is unmixed and satisfies the packing property, then I is normally
torsion-free.
Other recent results have taken a different approach. Instead of working directly with the
edge ideal of a clutter C, one can work with its Alexander dual, which is again the edge ideal
of a clutter. Using this approach, the embedded associated primes of the Alexander dual have
been linked to colorings of a clutter. Recall that χ(C) is the minimal number d for which there
is a partition X1 , . . . , Xd of the vertices of C for which for all edges f of C, f 6⊂ Xi for every i.
A clutter is critically d-chromatic if χ(C) = d but χ(C\{x}) < d for every vertex x.
Theorem 4.25. (a) [41, Corollary 4.6] If I is the ideal of covers of a clutter C, and if
the induced subclutter Cp on the vertices in p is critically (d + 1)-chromatic, then p ∈
Ass(R/I d ) but p 6∈ Ass(R/I t ) for any t ≤ d − 1.
(b) [41, Theorem 5.9] If I is the ideal of covers of a perfect graph G, then p ∈ Ass(R/I t ) if
and only if the induced graph on the vertices in p is a clique of size at most t + 1.
If one restricts to a particular power, additional results on embedded associate primes are
known. For example, in [40, Corollary 3.4] it is shown that if I is the edge ideal of the Alexander
dual of a graph G, then embedded primes of R/I 2 are in one-to-one correspondence with induced
odd cycles of G. More precisely, p ∈ Ass(R/I 2 ) is an embedded prime if and only if the induced
subgraph of G on the vertices in p is an induced odd cycle of G.
An interesting class of ideals, which is in a sense dual to the edge ideals of graphs, is unmixed
square-free monomial ideals of height two. These are the Alexander duals of edge ideals of
23
graphs, which can be viewed as edge ideals of clutters where, instead of requiring that each edge
has two vertices, it is instead required that each minimal vertex cover has two vertices. For
such ideals it has been shown in [42, Theorem 1.2] that an affirmative answer to a conjecture on
graph colorings, [42, Conjecture 1.1], would imply an affirmative answer to Question 4.16. In [42,
Corollary 3.11] it is shown that this conjecture holds for cliques, odd holes, and odd antiholes.
Thus the Alexander duals of these special classes of graphs provide additional examples where
Question 4.16 has an affirmative answer.
We now provide a more detailed discussion of the Conforti-Cornuéjols conjecture, followed by
a collection of results which provide families of clutters where the conjecture is known to be true
(such a family was already given in Theorem 4.24(c)). We also discuss some algebraic versions
of this conjecture and how it relates to the depth of powers of edge ideals and to normality and
torsion-freeness.
Having defined the notion of a minor for edge ideals, using the correpondence between clutters
and square-free monomias ideals, we also have the notion of a minor of a clutter. We say that
C has the packing property if I(C) has this property.
Definition 4.26. Let A be the incidence matrix of a clutter C. The set covering polyhedron is
the rational polyhedron:
Q(A) = {x ∈ Rn | x ≥ 0, xA ≥ 1},
where 0 and 1 are vectors whose entries are equal to 0 and 1 respectively. Often we denote the
vectors 0, 1 simply by 0, 1. We say that Q(A) is integral if it has only integral vertices.
Theorem 4.27. (A. Lehman [78], [17, Theorem 1.8]) If a clutter C has the packing property,
then Q(A) is integral.
The converse is not true. A famous example is the clutter Q6 , given below. It does not pack
and Q(A) is integral.
Example 4.28. Let I = (x1 x2 x5 , x1 x3 x4 , x2 x3 x6 , x4 x5 x6 ). The figure:
x1q x2
p p p p p pp pp p p p p p p p pp p p p p p p p p p pp p p p p p q
p p pp pp p pp pp p pp p p pp p p p p p pp p pp p p p p
pp p p p p p p p p pp p pp p
pp p p p p pp ppp
pp p x5
ppp q p pp
pp p
pp p p p ppp p pp
p
p pp pp p p p p
p p pp pp ppx4 p p p p pp pp ppp pp p p p pp pp pp p
p pp p p p p p pp p
p p p p p pp p q x q p p p pp pp p p
p pp pp pp pp 6 p pp p
p p pp
p pp p p pp pp pp
p pp p pp pp p p p p p
p pp pp
p p pp p
p p
q
x3
corresponds to the clutter associated to I. This clutter will be denoted by Q6 . Using Normaliz
[11] we obtain that R[It] = R[It][x1 · · · x6 t2 ]. Thus R[It] is not normal. An interesting property
of this example is that Ass(R/I i ) = Ass(R/I) for all i (see [45]).
Definition 4.29. A set of edges of a clutter C is independent if no two of them have a common
vertex. We denote the maximum number of independent edges of C by β1 (C). We call β1 (C) the
edge independence number of C or the monomial grade of I.
Let A be the incidence matrix of C. The edge independence number and the covering number
are related to min-max problems because they satisfy:
α0 (C) ≥ min{h1, xi| x ≥ 0; xA ≥ 1}
= max{hy, 1i| y ≥ 0; Ay ≤ 1} ≥ β1 (C).
Notice that α0 (C) = β1 (C) if and only if both sides of the equality have integral optimum
solutions.
Definition 4.30. A clutter C, with incidence matrix A, satisfies the max-flow min-cut (MFMC)
property if both sides of the LP-duality equation
(4.1) min{hα, xi| x ≥ 0; xA ≥ 1} = max{hy, 1i| y ≥ 0; Ay ≤ α}
have integral optimum solutions x and y for each non-negative integral vector α. The system
x ≥ 0; xA ≥ 1 is called totally dual integral (TDI) if the maximum in Eq. (4.1) has an integral
optimum solution y for each integral vector α with finite maximum.
Definition 4.31. If α0 (C) = β1 (C) we say that the clutter C (or the ideal I) has the König
property.
Note that C has the packing property if and only if every minor of C satisfies the König
property. This leads to the following well-known result.
Corollary 4.32. [17] If a clutter C has the max-flow min-cut property, then C has the packing
property.
Proof. Assume that the clutter C has the max-flow min-cut property. This property is closed
under taking minors. Thus it suffices to prove that C has the König property. We denote the
incidence matrix of C by A. By hypothesis the LP-duality equation
min{h1, xi| x ≥ 0; xA ≥ 1} = max{hy, 1i| y ≥ 0; Ay ≤ 1}
has optimum integral solutions x, y. To complete the proof notice that the left hand side of this
equality is α0 (C) and the right hand side is β1 (C).
Conforti and Cornuéjols [16] conjecture that the converse is also true.
Conjecture 4.33. (Conforti-Cornuéjols) If a clutter C has the packing property, then C has the
max-flow min-cut property.
An algebraic description of the packing property has already been given. In order to use
algebraic techniques to attack this combinatorial conjecture, an algebraic translation is needed
for the max-flow min-cut property. There are several equivalent algebraic descriptions of the
max-flow min-cut property, as seen in the following result.
Theorem 4.34. [35, 47, 69] Let C be a clutter and let I be its edge ideal. The following conditions
are equivalent:
(i) grI (R) is reduced.
(ii) R[It] is normal and Q(A) is an integral polyhedron.
(iii) x ≥ 0; xA ≥ 1 is a TDI system.
(iv) C has the max-flow min-cut property.
(v) I i = I (i) for i ≥ 1, where I (i) is the ith symbolic power.
(vi) I is normally torsion-free.
25
By Theorems 4.34 and 4.27, Conjecture 4.33 reduces to:

Conjecture 4.35. [45] If I has the packing property, then R[It] is normal.
Several variations of condition (ii) of Theorem 4.34 are possible. In particular, there are
combinatorial conditions on the clutter that can be used to replace the normality of the Rees
algebra. One such condition is defined below.
Definition 4.36. Let C ∨ be the clutter of minimal vertex covers of C. The clutter C is called
diadic if |e ∩ e′ | ≤ 2 for e ∈ E(C) and e′ ∈ E(C ∨ )
Proposition 4.37. [45] If Q(A) is integral and C is diadic, then I is normally torsion-free.
Theorem 4.34 can be used to exhibit families of normally torsion-free ideals. Recall that a
matrix A is called totally unimodular if each i × i subdeterminant of A is 0 or ±1 for all i ≥ 1.
Corollary 4.38. If A is totally unimodular, then I and I ∨ are normally torsion-free.
Proof. By [97] the linear system x ≥ 0; xA ≥ 1 is TDI. Hence I is normally torsion-free by

Theorem 4.34. Let C ∨ be the blocker (or Alexander dual) of C. By [97, Corollary 83.1a(v),
p. 1441], we get that C ∨ satisfies the max-flow min-cut property. Hence I(C ∨ ) is normally
torsion-free by Theorem 4.34. Thus I ∨ is normally torsion-free because I(C ∨ ) = I ∨ .
In particular if I is the edge ideal of a bipartite graph, then I and I ∨ are normally torsion-free.
Theorem 4.34 shows that the Rees algebra and the associated graded ring play an important
role in the study of the max-flow min-cut property. An invariant related to the blowup algebras
will also be useful. The analytic spread of an edge ideal I is given by ℓ(I) = dim R[It]/mR[It].
If C is uniform, the analytic spread of I is the rank of the incidence matrix of C. The analytic
spread of a monomial ideal can be computed in terms of the Newton polyhedron of I, see [4].
The next result follows directly from [83, Theorem 3].
Proposition 4.39. If Q(A) is integral, then ℓ(I) < n = dim(R).
To relate this result on ℓ(I) to Conjecture 4.33 (or equivalently to Conjecture 4.35) we first
need to recall the following bound on the depths of the powers of an ideal I.
Theorem 4.40. inf i {depth(R/I i )} ≤ dim(R) − ℓ(I). If grI (R) is Cohen-Macaulay, then the
equality holds.
This inequality is due to Burch [12] (cf. [70, Theorem 5.4.7]), while the equality comes from
[31]. By a result of Brodmann [8], depth R/I k is constant for k ≫ 0. Broadmann improved
Burch’s inequality by showing that the constant value is bounded by dim(R) − ℓ(I). For a study
of the initial and limit behaviour of the numerical function f (k) = depth R/I k see [59].
Theorem 4.41. [68] Let R be a Cohen-Macaulay ring and let I be an ideal of R containing
regular elements. If R[It] is Cohen-Macaulay, then grI (R) is Cohen-Macaulay.
Proposition 4.42. Let C be a clutter and let I be its edge ideal. Let Ji be the ideal obtained
from I by making xi = 1. If Q(A) is integral, then I is normal if and only if Ji is normal for
all i and depth(R/I k ) ≥ 1 for all k ≥ 1.
Proof. Assume that I is normal. The normality of an edge ideal is closed under taking minors
[35], hence Ji is normal for all i. By hypothesis the Rees algebra R[It] is normal. Then R[It]
is Cohen-Macaulay by a theorem of Hochster [66]. Then the ring grI (R) is Cohen-Macaulay by
Theorem 4.41. Hence using Theorem 4.40 and Proposition 4.39 we get that depth(R/I i ) ≥ 1 for
all i. The converse follows readily adapting the arguments given in the proof of the normality
criterion presented in [35].
By Proposition 4.42 and Theorem 4.27, we get that Conjecture 4.33 also reduces to:
Conjecture 4.43. If I has the packing property, then depth(R/I i ) ≥ 1 for all i ≥ 1.
We conclude this section with a collection of results giving conditions under which Conjec-
ture 4.33, or its equivalent statements mentioned above, is known to hold. For uniform clutters
it suffices to prove Conjecture 4.33 for Cohen-Macaulay clutters [23].
Proposition 4.44. [45] Let C be the collection of bases of a matroid. If C satisfies the packing
property, then C satisfies the max-flow min-cut property.
When G is a graph, integrality of Q(A) is sufficient in condition (ii) of Theorem 4.34, and the
packing property is sufficient to imply the max-flow min-cut property, thus providing another
class of examples for which Conjecture 4.33 holds.
Proposition 4.45. [17, 45] If G is a graph and I = I(G), then the following are equivalent:
(a) grI (R) is reduced.
(b) G is bipartite.
(c) Q(A) is integral.
(d) G has the packing property.
(e) G has the max-flow min-cut property.
(f) I i = I (i) for i ≥ 1.
Definition 4.46. A clutter is binary if its edges and its minimal vertex covers intersect in an
odd number of vertices.
Theorem 4.47. [98] A binary clutter C has the max-flow min-cut property if and only if Q6 is
not a minor of C.
Corollary 4.48. If C is a binary clutter with the packing property, then C has the max-flow
min-cut property.
Proposition 4.49. [112] Let C be a uniform clutter and let A be its incidence matrix. If the
polyhedra
P (A) = {x| x ≥ 0; xA ≤ 1} and Q(A) = {x| x ≥ 0; xA ≥ 1}
are integral, then C has the max-flow min-cut property.
In light of Theorem 4.27, this result implies that if P (A) is integral and C has the packing
property, then C has the max-flow min-cut property. An open problem is to show that this result
holds for non-uniform clutters (see [82, Conjecture 1.1]).
A Meyniel graph is a simple graph in which every odd cycle of length at least five has at least
two chords. The following gives some support to [82, Conjecture 1.1] because Meyniel graphs
are perfect [97, Theorem 66.6].
Theorem 4.50. [82] Let C be the clutter of maximal cliques of a Meyniel graph. If C has the
packing property, then C has the max-flow min-cut property.
Let P = (X, ≺) be a partially ordered set (poset for short) on the finite vertex set X and let
G be its comparability graph. Recall that the vertex set of G is X and the edge set of G is the
set of all unordered pairs {xi , xj } such that xi and xj are comparable.
27
Theorem 4.51. [25] If G is a comparability graph and C is the clutter of maximal cliques of G,
then the edge ideal I(C) is normally torsion free.
Theorem 4.52. [24] Let C be a uniform clutter with a perfect matching such that C has the pack-
ing property and α0 (C) = 2. If the columns of the incidence matrix of C are linearly independent,
then C has the max-flow min-cut property.
5. ACKNOWLEDGMENTS
The software package Macaulay 2 [49] was used to compute many of the examples in this
paper, including Example 4.18, which was originally discovered by three undegraduate students
(Mike Alwill, Katherine Pavelek, and Jennifer von Reis) in an unpublished project in 2002. The
authors would like to thank an anonymous referee for providing us with useful comments and
suggestions.
References
[1] M. Barile, On the arithmetical rank of the edge ideals of forests, Comm. Algebra 36 (2008), no. 12, 4678 -
4703. 1, 12
[2] C. Berge, Some common properties for regularizable graphs, edge-critical graphs and B-graphs, in “Theory
and practice of combinatorics” (A. Rosa, G. Sabidussi and J. Turgeon, Eds.), North-Holland Math. Stud.
60, North-Holland, Amsterdam, 1982, pp. 31–44. 4, 5
[3] I. Bermejo and P. Gimenez, Saturation and Castelnuovo-Mumford regularity, J. Algebra 303 (2006), no. 2,
592-617. 5
[4] C. Bivià-Ausina, The analytic spread of monomial ideals, Comm. Algebra 31 (2003), no. 7, 3487–3496. 25
[5] A. Björner, Topological methods, Handbook of combinatorics, Vol. 1, 2, 1819–1872, Elsevier, Amsterdam,
1995. 3, 12
[6] A. Björner and M. Wachs, Shellable nonpure complexes and posets I, Trans. Amer. Math. Soc. 348 (1996),
1299–1327. 2
[7] B. Bollobás, Modern Graph Theory, Graduate Texts in Mathematics 184, Springer-Verlag, New York, 1998.
2
[8] M. Brodmann, The asymptotic nature of the analytic spread, Math. Proc. Cambridge Philos. Soc. 86 (1979),
35–39. 25
[9] M. Brodmann, Asymptotic stability of Ass(M/I n M ), Proc. Amer. Math. Soc. 74 (1979), 16–18. 15
[10] W. Bruns and J. Herzog, Cohen-Macaulay Rings, Cambridge University Press, Cambridge, Revised Edition,
1997. 2, 7
[11] W. Bruns and B. Ichim, Normaliz 2.0, Computing normalizations of affine semigroups 2008. Available from
http://www.math.uos.de/normaliz. 23
[12] L. Burch, Codimension and analytic spread, Proc. Camb. Phil. Soc. 72 (1972), 369-373. 25
[13] A. H. Busch, F. F. Dragan and R. Sritharan, New min-max theorems for weakly chordal and dually chordal
graphs, Lecture Notes in Computer Science 6509, 207–218, 2010, Springer Verlag. 11
[14] J. Chen, S. Morey and A. Sung, The stable set of associated primes of the ideal of a graph, Rocky Mountain
J. Math. 32 (2002), 71–89. 17, 18, 19
[15] A. Conca and E. De Negri, M -sequences, graph ideals, and ladder ideals of linear type, J. Algebra 211
(1999), 599–624. 14, 21
[16] M. Conforti and G. Cornuéjols, Clutters that pack and the Max-Flow Min-Cut property: A conjecture, The
Fourth Bellairs Workshop on Combinatorial Optimization (W. R. Pulleyblank, F. B. Shepherd, eds.), 1993.
24
[17] G. Cornuéjols, Combinatorial optimization: Packing and covering, CBMS-NSF Regional Conference Series
in Applied Mathematics 74, SIAM (2001). 2, 22, 23, 24, 26
[18] A. Corso and U. Nagel, Monomial and toric ideals associated to ferrers graphs, Trans. Amer. Math. Soc. 361
(2009), no. 3, 1371-1395. 5, 11
[19] K. Dalili, M. Kummini, Dependence of betti numbers on characteristic. Preprint, 2010, arXiv:1009.4243v1.
7
[20] K. Dalili, S. Faridi and W. Traves, The reconstruction conjecture and edge ideals, Discrete Math. 308 (2008),
2002-2010. 5
[21] G. A. Dirac, On rigid circuit graphs, Abh. Math. Sem. Univ. Hamburg 38 (1961), 71–76. 4
[22] A. Dochtermann and A. Engstrom, Algebraic properties of edge ideals via combinatorial topology, Electron.
J. Combin. 16 (2009), no. 2, R2. 1, 4, 14
[23] L. A. Dupont, E. Reyes and R. H. Villarreal, Cohen-Macaulay clutters with combinatorial optimization
properties and parallelizations of normal edge ideals, São Paulo J. Math. Sci. 3 (2009), no. 1, 61-75. 26
[24] L. A. Dupont and R. H. Villarreal, Algebraic and combinatorial properties of ideals and algebras of uniform
clutters of TDI systems, J. Comb. Optim., to appear. 1, 27
[25] L. A. Dupont and R. H. Villarreal, Edge ideals of clique clutters of comparability graphs and the normality
of monomial ideals, Math. Scand. 106 (2010), no. 1, 88–98. 1, 27
[26] A. M. Duval, Algebraic shifting and sequentially Cohen-Macaulay simplicial complexes. Electron. J. Combin.
3 (1996), no. 1, Research Paper 21. 13
[27] J. A. Eagon and V. Reiner, Resolutions of Stanley-Reisner rings and Alexander duality, J. Pure Appl. Algebra
130 (1998), 265–275. 6, 8
[28] D. Eisenbud, Commutative Algebra with a view toward Algebraic Geometry, Graduate Texts in Mathematics
150, Springer-Verlag, 1995. 2, 5, 9
[29] D. Eisenbud, The geometry of syzygies: A second course in commutative algebra and algebraic geometry,
Graduate Texts in Mathematics 229, Springer, New York, 2005. 5
[30] D. Eisenbud and S. Goto, Linear free resolutions and minimal multiplicity, J. Algebra 88 (1984), no. 1,
189–133. 12
[31] D. Eisenbud and C. Huneke, Cohen-Macaulay Rees algebras and their specializations, J. Algebra 81 (1983),
202–224. 25
[32] E. Emtander, A class of hypergraphs that generalizes chordal graphs, Math. Scand. 106 (2010), no. 1, 50–66.
1, 7
[33] V. Ene, O. Olteanu and N. Terai, Arithmetical rank of lexsegment edge ideals, Bull. Math. Soc. Sci. Math.
Roumanie (N.S.) 53 (2010), no. 4, 315–327. 11, 12
[34] P. Erdös and T. Gallai, On the minimal number of vertices representing the edges of a graph, Magyar Tud.
Akad. Mat. Kutató Int. Közl. 6 (1961), 181–203. 5
[35] C. Escobar, R. H. Villarreal and Y. Yoshino, Torsion freeness and normality of blowup rings of monomial
ideals, Commutative Algebra, Lect. Notes Pure Appl. Math. 244, Chapman & Hall/CRC, Boca Raton, FL,
2006, pp. 69-84. 24, 25, 26
[36] M. Estrada and R. H. Villarreal, Cohen-Macaulay bipartite graphs, Arch. Math. 68 (1997), 124–128. 3
[37] M. Farber, Characterizations of strongly chordal graphs, Discrete Math. 43 (1983), no. 2-3, 173-189. 4
[38] S. Faridi, Cohen-Macaulay properties of square-free monomial ideals, J. Combin. Theory Ser. A 109 (2005),
no. 2, 299–329. 2, 3
[39] S. Faridi, Simplicial trees are sequentially Cohen-Macaulay, J. Pure Appl. Algebra 190 (2004), 121–136. 1, 4
[40] C. Francisco, H.T. Hà, and A. Van Tuyl, Associated primes of monomial ideals and odd holes in graphs, J.
Algebraic Combin. 32 (2010), 287–301. 1, 22
[41] C. Francisco, H.T. Hà and A. Van Tuyl, Colorings of hypergraphs, perfect graphs, and associated primes of
powers of monomial ideals, J. Algebra 331 (2011), no. 1, 224-242. 1, 21, 22
[42] C. Francisco, H.T. Hà and A. Van Tuyl, A conjecture on critical graphs and connections to the persistence
of associated primes, Discrete Math. 310 (2010), 2176–2182. 23
[43] C. A. Francisco and A. Van Tuyl, Sequentially Cohen-Macaulay edge ideals, Proc. Amer. Math. Soc. 135
(2007), 2327–2337. 3, 4
[44] R. Fröberg, On Stanley-Reisner rings, in Topics in algebra (S. Balcerzyk et. al., Eds.), Part 2. Polish Scientific
Publishers, 1990, pp. 57–70. 8, 12
[45] I. Gitler, E. Reyes and R. H. Villarreal, Blowup algebras of square–free monomial ideals and some links to
combinatorial optimization problems, Rocky Mountain J. Math. 39 (2009), no. 1, 71–102. 1, 17, 22, 23, 25,
26
[46] I. Gitler and C. Valencia, Bounds for invariants of edge-rings, Comm. Algebra 33 (2005), 1603–1616. 1, 5, 15
[47] I. Gitler, C. Valencia and R. H. Villarreal, A note on Rees algebras and the MFMC property, Beiträge
Algebra Geom. 48 (2007), no. 1, 141–150. 1, 22, 24
[48] M. C. Golumbic, Algorithmic graph theory and perfect graphs, second edition, Annals of Discrete Mathe-
matics 57, Elsevier Science B.V., Amsterdam, 2004. 21
29
[49] D.R. Grayson and M.E. Stillman, Macaulay2, a software system for research in algebraic geometry, 1996.
http://www.math.uiuc.edu/Macaulay2/. 7, 9, 27
[50] G. M. Greuel and G. Pfister, A Singular introduction to commutative algebra, Springer, Berlin, Second
Extended Edition, 2008. 5
[51] H. T. Hà and S. Morey, Embedded associated primes of powers of square-free monomial ideals, J. Pure Appl.
Algebra 214 (2010), no. 4, 301–308. 1, 17, 20, 22
[52] H. T. Hà, S. Morey and R.H. Villarreal, Cohen-Macaulay admissible clutters, J. Commut. Algebra 1 (2009),
no. 3, 463–480. 4
[53] H. T. Hà and A. Van Tuyl, Monomial ideals, edge ideals of hypergraphs, and their graded Betti numbers, J.
Algebraic Combin. 27 (2008), 215–245. 1, 5, 7, 8, 10, 14
[54] F. Harary, Graph Theory, Addison-Wesley, Reading, MA, 1972. 2
[55] J. He and A. Van Tuyl, Algebraic properties of the path ideal of a tree, Comm. Algebra 38 (2010), no. 5,
1725-1742. 4, 14
[56] M. Herrmann, S. Ikeda, U. Orbanz, Equimultiplicity and blowing up. An algebraic study, Springer-Verlag,
Berlin, 1988. 20
[57] J. Herzog, A Generalization of the Taylor complex construction, Comm. Algebra 35 (2007), no. 5, 1747–1756.
11
[58] J. Herzog and T. Hibi, Componentwise linear ideals, Nagoya Math. J. 153 (1999), 141–153. 6, 8
[59] J. Herzog and T. Hibi, The depth of powers of an ideal, J. Algebra 291 (2005), 534–550. 20, 25
[60] J. Herzog and T. Hibi, Distributive lattices, bipartite graphs and Alexander duality, J. Algebraic Combin.
22 (2005), no. 3, 289–302. 3
[61] J. Herzog and T. Hibi, Monomial Ideals, Graduate Texts in Mathematics 260, Springer-Verlag, 2011. 1, 6, 8
[62] J. Herzog, T. Hibi, N.V. Trung and X. Zheng, Standard graded vertex cover algebras, cycles and leaves,
Trans. Amer. Math. Soc. 360 (2008), 6231–6249. 1, 4, 22
[63] J. Herzog, T. Hibi and X. Zheng, Dirac’s theorem on chordal graphs and Alexander duality, European J.
Combin. 25 (2004), no. 7, 949–960. 4
[64] J. Herzog, T. Hibi and X. Zheng, Monomial ideals whose powers have a linear resolution, Math. Scand. 95
(2004), 23–32. 4
[65] L.T. Hoa, Stability of associated primes of monomial ideals, Vietnam J. Math. 34 (2006), no. 4, 473–487.
16, 20
[66] M. Hochster, Rings of invariants of tori, Cohen-Macaulay rings generated by monomials, and polytopes, Ann.
of Math. 96 (1972), 318–337. 25
[67] M. Hochster, Criteria for the equality of ordinary and symbolic powers of primes, Math. Z. 133 (1973),
53–65. 16
[68] C. Huneke, On the associated graded ring of an ideal, Illinois J. Math. 26 (1982), 121-137. 25
[69] C. Huneke, A. Simis and W. V. Vasconcelos, Reduced normal cones are domains, Contemp. Math. 88 (1989),
95–101. 24
[70] C. Huneke and I. Swanson, Integral Closure of Ideals Rings, and Modules, London Math. Soc., Lecture Note
Series 336, Cambridge University Press, Cambridge, 2006. 25
[71] C. Ionescu and G. Rinaldo, Some algebraic invariants related to mixed product ideals, Arch. Math. 91 (2008),
20-30. 11, 15
[72] S. Jacques, Betti Numbers of Graph Ideals, Ph.D. Thesis, University of Sheffiels 2004,
arXiv:math/0410107v1.
[73] G. Kalai and R. Meshulam, Unions and intersections of Leray complexes, J. Combin. Theory Ser. A 113
(2006), 1586–1592. 11
[74] M. Katzman, Characteristic-independence of Betti numbers of graph ideals, J. Combin. Theory Ser. A 113
(2006), no. 3, 435–454. 5, 7, 8
[75] D. Kiani and S. Moradi, Bounds for the regularity of edge ideal of vertex decomposable and shellable graphs.
Preprint, 2010, arXiv:1007.4056v1. 5
[76] K. Kimura, N. Terai and K. Yoshida, Arithmetical rank of squarefree monomial ideals of small arithmetic
degree, J. Algebraic Combin. 29 (2009), 389-404. 12
[77] M. Kummini, Regularity, depth and arithmetic rank of bipartite edge ideals, J. Algebraic Combin. 30 (2009),
no. 4, 429–445. 1, 5, 11
[78] A. Lehman, On the width-length inequality and degenerate projective planes, in Polyhedral Combinatorics
(W. Cook and P. Seymour Eds.) DIMACS Series in Discrete Mathematics and Theoretical Computer Science
1, Amer. Math. Soc., 1990, pp. 101-105 23
i
[79] G. Lyubeznik, On the local cohomology modules HA (R) for ideals A generated by monomials in an R-
sequence. In: Greco, S., Strano, R., eds. Complete Intersections. Lectures given at the 1st 1983 Session of the
Centro Internazionale Matematico Estivo (C.I.M.E.), Acireale, Italy, June 13-21, 1983; Berlin-Heidelberg:
Springer, pp. 214-220. 12
[80] G. Lyubeznik, On the arithmetical rank of monomial ideals, J. Algebra 112 (1988), 86–89. 12
[81] M. Mahmoudi, A. Mousivand, M. Crupi, G. Rinaldo, N. Terai, and S. Yassemi, Vertex decomposability and
regularity of very well-covered graphs, to appear: J. Pure Appl. Algebra. 5
[82] J. Martı́nez-Bernal, E. O’Shea and R. H. Villarreal, Ehrhart clutters: regularity and max-flow min-cut,
Electron. J. Combin. 17 (2010), no. 1, R52. 26
[83] S. McAdam, Asymptotic prime divisors and analytic spreads, Proc. Amer. Math. Soc. 80 (1980), 555–559.
25
[84] S. McAdam, Asymptotic Prime Divisors, Lecture Notes in Mathematics 103, Springer–Verlag, New York,
1983. 20
[85] S. McAdam and P. Eakin, The asymptotic Ass, J. Algebra 61 (1979), no. 1, 71–81. 16, 20
[86] E. Miller and B. Sturmfels, Combinatorial Commutative Algebra, Graduate Texts in Mathematics 227,
Springer, 2004. 6
[87] S. Morey, Depths of powers of the edge ideal of a tree, Comm. Algebra 38 (2010), no. 11, 4042–4055. 1, 12,
15
[88] S. Morey, Stability of associated primes and equality of ordinary and symbolic powers of ideals, Comm.
Algebra 27 (1999), 3221–3231. 16, 21
[89] S. Morey, E. Reyes and R. H. Villarreal, Cohen-Macaulay, shellable and unmixed clutters with a perfect
matching of König type, J. Pure Appl. Algebra 212 (2008), no. 7, 1770–1786. 1, 2, 3
[90] E. Nevo, Regularity of edge ideals of C4 -free graphs via the topology of the lcm-lattice, J. Combin. Theory
Ser. A 118 (2011), no.2, 491–501. 1, 5, 11
[91] I. Peeva and M. Stillman, Open problems on syzygies and Hilbert functions, J. Commut. Algebra 1 (2009),
no. 1, 159–195. 1
[92] M. D. Plummer, Some covering concepts in graphs, J. Combin. Theory 8 (1970), 91–98. 2
[93] M. D. Plummer, Well-covered graphs: a survey, Quaestiones Math. 16 (1993), no. 3, 253–287. 2
[94] G. Ravindra, Well-covered graphs, J. Combinatorics Information Syst. Sci. 2 (1977), no. 1, 20–21. 2, 3
[95] P. Renteln, The Hilbert series of the face ring of a flag complex, Graphs Combin. 18 (2002), no. 3, 605–619.
11
[96] A. Schrijver, Theory of Linear and Integer Programming, John Wiley & Sons, New York, 1986. 3
[97] A. Schrijver, Combinatorial Optimization, Algorithms and Combinatorics 24, Springer-Verlag, Berlin, 2003.
2, 3, 25, 26
[98] P. D. Seymour, The matroids with the max-flow min-cut property, J. Combin. Theory Ser. B 23 (1977),
189-222. 26
[99] R. Sharp, Convergence of sequences of sets of associated primes. Proc. Amer. Math. Soc. 131 (2003), 3009–
3017. 15, 20
[100] A. Simis, W. V. Vasconcelos and R. H. Villarreal, On the ideal theory of graphs, J. Algebra 167 (1994),
389–416. 2, 16, 22
[101] D. E. Smith, On the Cohen-Macaulay property in commutative algebra and simplicial topology, Pacific J.
Math. 141 (1990), 165–196. 12
[102] R. P. Stanley, Combinatorics and Commutative Algebra. Second edition. Progress in Mathematics 41.
Birkhäuser Boston, Inc., Boston, MA, 1996. 2, 3
[103] N. Terai, Alexander duality theorem and Stanley-Reisner rings, Sūrikaisekikenkyūsho Kōkyūroku 1078
(1999), 174–184, Free resolutions of coordinate rings of projective varieties and related topics (Kyoto 1998).
5, 6, 8, 12
[104] B. Toft, Colouring, stable sets and perfect graphs, in Handbook of Combinatorics I (R. L. Graham et. al.,
Eds.), Elsevier, 1995, pp. 233–288. 4, 10
[105] T. N. Trung, Stability of associated primes of integral closures of monomial ideals, J. Combin. Theory Ser.
A 116 (2009), 44–54. 16
[106] A. Van Tuyl, Sequentially Cohen-Macaulay bipartite graphs: vertex decomposability and regularity, Arch.
Math. 93 (2009), no. 5, 451–459. 1, 4, 5, 10
[107] A. Van Tuyl and R. H. Villarreal, Shellable graphs and sequentially Cohen-Macaulay bipartite graphs, J.
Combin. Theory Ser. A 115 (2008), no. 5, 799–814. 3, 4, 11
31
[108] W. V. Vasconcelos, Computational Methods in Commutative Algebra and Algebraic Geometry, Springer-
Verlag, 1998. 12
[109] R. H. Villarreal, Cohen-Macaulay graphs, Manuscripta Math. 66 (1990), 277–293. 1, 3
[110] R. H. Villarreal, Monomial Algebras, Monographs and Textbooks in Pure and Applied Mathematics 238,
Marcel Dekker, New York, 2001. 1, 2, 4
[111] R. H. Villarreal, Unmixed bipartite graphs, Rev. Colombiana Mat. 41 (2007), no. 2, 393–395. 2, 3
[112] R. H. Villarreal, Rees algebras and polyhedral cones of ideals of vertex covers of perfect graphs, J. Algebraic
Combin. 27 (2008), 293–305. 21, 26
[113] V. I. Voloshin, Coloring mixed hypergraphs: theory, algorithms and applications, Fields Institute Monographs
17, American Mathematical Society, Providence, RI, 2002. 7
[114] G. Whieldon, Jump sequences of edge ideals. Preprint, 2010, arXiv:1012.0108v1. 5
[115] R. Woodroofe, Chordal and sequentially Cohen-Macaulay clutters. Preprint, 2009, arXiv:0911.4697v2. 4,
7
[116] R. Woodroofe, Vertex decomposable graphs and obstructions to shellability, Proc. Amer. Math. Soc. 137
(2009), 3235–3246. 1, 3, 4
[117] R. Woodroofe, Matchings, coverings, and Castelnuovo-Mumford regularity. Preprint, 2010,
arXiv:1009.2756. 1, 6, 10, 11
[118] X. Zheng, Resolutions of facet ideals, Comm. Algebra 32 (2004), 2301–2324. 1, 8, 11
Department of Mathematics, Texas State University, 601 University Drive, San Marcos, TX
78666.
E-mail address: morey@txstate.edu
Departamento de Matemáticas, Centro de Investigación y de Estudios Avanzados del IPN,

Apartado Postal 14–740, 07000 Mexico City, D.F.
E-mail address: vila@math.cinvestav.mx

Key Words and Phrases

Uploaded by

Copyright:

Available Formats

Key Words and Phrases

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Key Words and Phrases

Uploaded by

Copyright:

Available Formats

EDGE IDEALS: ALGEBRAIC AND COMBINATORIAL PROPERTIES

SUSAN MOREY AND RAFAEL H. VILLARREAL

2. Algebraic and combinatorial properties of edge ideals

We call a clutter C sequentially Cohen-Macaulay if R/I(C) is sequentially Cohen-Macaulay.

If G is a graph, then G is balanced if and only if G is bipartite and G is totally balanced if

In [2] the integer α0 (G) is called the transversal number of G.

3. Invariants of edge ideals: regularity, projective dimension, depth

Therefore, using Theorem 3.2 and Lemma 3.5, we get

Proof. The proof is by induction on the number of vertices. Let G be a graph in F. If G

C = ∪si=1 Ci equals ∪si=1 E(Ci ). By Proposition 3.16, it suffices to notice

A clutter C is called co-CM if Ic (C) is Cohen-Macaulay. A co-CM clutter is uniform because

Proof. It follows from the proof of Theorem 3.31.

The converse of Theorem 3.33 is not true.

The result now follows because β0 (G) = dim R/I(C).

4. Stability of associated primes

An interesting variation on Questions 4.1 and 4.2 was posed in [99].

Proof. Since x is a leaf of G, there is a unique generator e = xy ∈ I divisible by x. Let a

By Theorems 4.34 and 4.27, Conjecture 4.33 reduces to:

Proof. By [97] the linear system x ≥ 0; xA ≥ 1 is TDI. Hence I is normally torsion-free by

Departamento de Matemáticas, Centro de Investigación y de Estudios Avanzados del IPN,

You might also like