Bounded powers of edge ideals: The strong exchange property

Seyed Amin Seyed Fakhari; Takayuki Hibi

REVIEW 2 major objections 4 minor 1 cited by

Bounded powers of edge ideals: The strong exchange property

T0 review · 2 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read For cycles, trees, and unicyclic graphs, the paper settles exactly when the bounded-power generator set W(c,G) has the strong exchange property; equivalently, when it is of Veronese type for every c.

desk verdict New classifications of cycles, trees, and unicyclic graphs for the strong exchange property, but the printed proof has a gap in the key three-variable lemma. read the letter →

arxiv 2506.03480 v1 pith:THFL5U4X submitted 2025-06-04 math.AC math.CO

Takayuki Hibi , Seyed Amin Seyed Fakhari This is my paper

classification math.ACmath.CO MSC 13F6513H1005E40

keywords boundedpowersofedgeidealsstrongexchangepropertyVeronesetypepolymatroidaltoricunicyclicgraphstreeclassificationcycle

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks when the minimal generator set of the top bounded power of an edge ideal satisfies the strong exchange property, not merely the symmetric exchange property that polymatroidality guarantees. It proves that this happens for all bound vectors c exactly for cycles of length 3 through 7, for trees that are either P6 or stars with at most one pendant edge per leaf, and for the four unicyclic families listed in Theorem 5.23. The interest is that strong exchange is equivalent to W(c,G) being of Veronese type, and it forces the associated toric ideal to have a quadratic Gröbner basis generated by symmetric exchange binomials. Thus the classification identifies exactly where bounded-power toric rings of edge ideals become the most tractable.

What carries the argument

The load-bearing object is W(c,G), the minimal monomial generating set of the top nonvanishing bounded power (I(G)^{δ_c(I(G))})_c; it is always polymatroidal, hence has symmetric exchange, and the paper asks when it has the stronger exchange property. The main equivalence used is that strong exchange holds exactly when W(c,G) is of Veronese type, i.e., a monomial multiple of the generator set of a Veronese algebra. The positive direction is carried by Lemma 3.3 (every triangle-free graph with independence number at most 3 has the property), Lemma 4.4 (deleting a leaf preserves the property), and Theorem 2.1 (complete multipartite graphs minus a matching are of Veronese type). The negative direction is carried by explicit pairs of monomials with incompatible exponents for C_n with n ≥ 8, P_n with n ≥ 7, and each excluded unicyclic graph.

What would settle it

Test the auxiliary ideal J in Lemma 3.3 on the triangle-free graphs used for C5, C7, and P6: if J fails the polymatroidal exchange axiom for any of the c vectors considered in the proof, then the positive classifications in Theorems 3.5, 4.10, and 5.23 are not supported.

Watch

Extended reading notes

Core claim

The central claim is that the strong exchange property for W(c,G) — the minimal monomial generating set of (I(G)^{δ_c(I(G))})_c — is a graph-level property that can be classified. For a cycle C_n it holds for all c exactly when 3 ≤ n ≤ 7. For a tree it holds exactly when the tree is P6 or is obtained from a star by attaching at most one pendant edge to each leaf. For a unicyclic graph with unique cycle of length ℓ, it holds exactly in the cases of Theorem 5.23: no graph with ℓ ≥ 8; for ℓ = 5, 6, 7 exactly those with independence number at most three; and for ℓ = 4 or ℓ = 3 exactly the four explicitly listed families. Throughout, 'strong exchange property for all c' is equivalent to 'W(c,G) is of Veronese type for all c,' so the classification is simultaneously a classification of when the bounded-power generator sets are Veronese-type algebras.

Load-bearing premise

The positive classifications of cycles, trees, and unicyclic graphs rest on Lemma 3.3, which asserts without proof or citation that the auxiliary ideal J left after factoring out pure powers is polymatroidal in three variables, and on Lemma 3.2, whose printed proof contains an unjustified inference; if either of those assertions fails, the positive direction loses its foundation.

Editorial extensions

If this is right

For every cycle C_n with 3 ≤ n ≤ 7 and every bound vector c, the toric ideal Ker(π_G^c) has a quadratic Gröbner basis and is generated by symmetric exchange binomials; for n ≥ 8 this fails for the specific c built in Lemma 3.1.
A tree has the strong exchange property for all c exactly when it is P6 or a star with at most one pendant edge per leaf; every other tree admits some c for which W(c,G) fails the strong exchange property.
For a unicyclic graph whose unique cycle has length 5, 6, or 7, the property holds for all c exactly when the independence number is at most 3.
For a unicyclic graph with a 4-cycle or a triangle, the property holds exactly for the four families listed in Theorem 5.23; with cycle length at least 8 it never holds.
In every positive case W(c,G) is of Veronese type, so the bounded-power generator sets are monomial multiples of Veronese-algebra generator sets.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the counterexamples in Sections 3–5 use only small bound vectors (typically all ones or a handful of 2s), a natural testable extension is to turn Theorem 5.23 into a finite decision procedure: for each graph, check the property only on c vectors bounded by a small total degree.
The positive families share a common reduction: after factoring out forced pure powers, the generator set becomes a Veronese algebra in few variables; this suggests a recursive pruning criterion — delete leaves, strip pure powers, and check the residual graph's independence number — that might classify larger graph classes.
The paper's Conjecture 5.24 can be tested on exactly the unicyclic graphs that Theorem 5.23 excludes: decide whether their toric ideals are nonetheless generated by symmetric exchange binomials, as happens in Example 5.25.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Read the full text. The main new thing is a classification: for all c, W(c,G) has the strong exchange property iff G is in the listed families, for cycles (n≤7), trees (P6 or stars with at most one pendant per leaf), and unicyclics (the families in Theorem 5.23). That is a real advance over the earlier polymatroidality and Gorenstein results. The negative directions are done properly: explicit c vectors and pairs w1,w2 that violate strong exchange, and each counterexample is checkable by hand. Lemma 4.4 (leaf removal) is a clean reduction and makes the 'only if' arguments for trees and unicyclics coherent.

The problem is in the positive direction. Lemma 3.2, the three-variable polymatroidal ideal claim, has a printed step that does not follow. Starting from w1 = x^a y^b z^c and w2 = x^{a'} y^{b'} z^{c'} with a>a', b

Lemma 3.3 then asserts that after peeling off the variables outside A_v, the residual ideal J is polymatroidal in three variables, with no argument or citation. That is the second load-bearing step, because Lemma 3.3 feeds directly into Theorems 3.5, 4.10 and 5.23 for the positive cases. If J is in fact polymatroidal by projecting the known polymatroidal set (I(G)^δ)_c, that should be written down.

So: the negative half is solid, the classification statements look plausible and are likely true, but the proof of the positive half is not complete as printed. This is repairable—it is a gap, not a contradiction. The paper deserves a serious referee; I would send it out, with instructions to focus on Lemma 3.2 and the J-polymatroidality assertion. I would not cite it in its current form until that is patched.

Referee Report

2 major / 4 minor

Summary. The paper studies, for an edge ideal I(G) and a positive integer vector c, the minimal monomial generating set W(c,G) of the top-degree bounded power (I(G)^{δ_c(I(G))})_c. Using the known equivalence between the strong exchange property and being of Veronese type, it classifies which cycles, trees, and unicyclic graphs enjoy the strong exchange property for all c. The main results are Theorem 3.5 (cycles C_n with 3≤n≤7), Theorem 4.10 (trees: P_6 or stars with at most one pendant edge per leaf), and Theorem 5.23 (a case analysis for unicyclic graphs). The negative directions are supported by explicit monomial counterexamples, while the positive directions rely on a structural lemma for triangle-free graphs with independence number at most three (Lemma 3.3) and on a three-variable polymatroidal exchange lemma (Lemma 3.2).

Significance. If the classifications are correct, they give a clean and broad answer to a natural question about bounded powers of edge ideals, linking polymatroidal ideals, Veronese-type algebras, and quadratic Gröbner bases of toric ideals. The paper's negative results are concrete and checkable, and the overall strategy—reducing positive cases to a small list of graphs and to a three-variable exchange lemma—is attractive. The main caveat is that the two key positive lemmas are not fully proved as written: the induction in Lemma 3.2 has an unjustified step, and Lemma 3.3 asserts without proof that a certain residual ideal is polymatroidal. Since all positive classifications depend on these lemmas, the central claim is not yet established by the text.

major comments (2)

[Lemma 3.2] The proof of Lemma 3.2 is incomplete. After ruling out c≥c′, the membership x^{a′+1}y^{b′}z^{c′−1}∈I does follow from the polymatroid exchange axiom applied to (w2,w1) at the z-coordinate, because x is the only coordinate in which w2 is smaller than w1. The problematic step is the next sentence: 'Since a≥a′+1, c≤c′−1 and b<b′, one has a>a′+1.' The stated inequalities allow a=a′+1, and no argument is given for the strict inequality. The subsequent claims c<c′−1 and x^{a′+2}y^{b′}z^{c′−2}∈I, and the later iterations, depend on that unsupported strict inequality and on an unstated exchange/divisibility argument. As printed, 'Continuing these processes' is not a valid induction. Because Lemma 3.3 invokes Lemma 3.2, this gap affects the positive parts of Theorems 3.5, 4.10, and 5.23.
[Lemma 3.3] In Case 1 and again in Subcase 2.2 (and implicitly in Case 2), the proof factors (I(G)^δ)_c as J times a product of pure powers of the variables outside A_v or A′_v, and states that 'J is a polymatroidal ideal in three variables.' This is a load-bearing assertion: Lemma 3.2 is then applied to J. No proof or citation is given for the polymatroidality of the residual ideal, which is a projection-like object obtained after fixing the exponents of all variables outside A_v. The claim is plausible, but it is not established in the text. The same unproved reduction is used in Lemmas 5.14, 5.20, and 5.21. Until this is proved, the positive classifications of C_4–C_7, P_6, and the unicyclic families in Theorem 5.23 are not fully justified.

minor comments (4)

[Throughout] There are several typos: 'monmomial' in the Introduction, 'cardianlity' in Section 1, 'foe' in the Introduction, 'exponet' in the Introduction, 'assuption' in the proof of Lemma 3.3, and 'unicycle' in the Section 5 heading.
[Lemma 5.8] The vertex set is declared as {x_1,...,x_6} and the edge set uses V(C_4), but the text says 'where V(C_5)=...'; this should be V(C_4).
[Lemma 5.22] In the proof, the inequality 'c_{i+k} ≤ c_k' should read 'c_{i+k} ≤ c_i', and the displayed inequality 'c i+k ≤ c k' in the first paragraph should likewise be corrected.
[Theorem 5.23(iii)] In the proof of the 'if' part, item (2) is not explicitly addressed; the proof should cite Lemma 5.14, which exactly establishes that the graph described in (2) is of Veronese type.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the unicyclic classification is derived from independent prior results, not from its own conclusion; the main weakness is an unproved polymatroidality assertion in Lemma 3.3, which is a proof gap rather than a circular reduction.

full rationale

I walked the derivation chain from Definition 0.3 through Theorem 5.23. The load-bearing steps are: (a) Theorem 0.2, importing from [4] the equivalence between strong exchange and Veronese type; (b) Theorem 2.1, citing [6, Theorem 4.1] for complete multipartite graphs minus matchings; (c) Lemma 3.2, a text-internal statement that three-variable polymatroidal ideals have strong exchange; (d) Lemma 3.3, which reduces W(c,G) to J times a pure-power factor and applies Lemma 3.2; and (e) Lemma 4.4, used in the only-if directions. None of these steps defines the graph classes in terms of the strong-exchange property being proved, and none fits a parameter and then reports it as a prediction. The self-citations to [5] and [6] are to different prior results (polymatroidality of bounded powers; complete multipartite classification) that do not assume the cycle, tree, or unicyclic classifications, so they are independent support rather than a circular reduction. The one genuinely concerning passage is in Lemma 3.3, where after deriving (I(G)^delta)_c = J times a product of pure powers, the text asserts 'J is a polymatroidal ideal in three variables' without proof, and Lemma 3.2's exchange iteration is written too tersely. That is a missing justification a referee should check; it is not a circular reduction, because J's polymatroidality is not being assumed from the strong-exchange conclusion. Score 2 rather than 0 only acknowledges the presence of self-citations in the framework; no load-bearing equation reduces to its own input.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central claim rests on four external or asserted inputs: the polymatroid exchange property, the strong-exchange/Veronese equivalence from [4], the complete multipartite theorem from the authors' prior work [6], and an unproved assertion inside Lemma 3.3 that the residual three-variable ideal J is polymatroidal. There are no fitted constants or invented objects.

assumptions (4)

standard math Symmetric exchange property holds for polymatroidal ideals (Herzog and Hibi, [2]).
Used in the introduction to conclude that W(c,G) is polymatroidal and in the proof of Lemma 3.2.
standard math For W(c,G), the strong exchange property holds if and only if W(c,G) is of Veronese type (Herzog, Hibi and Vladoiu, [4, Theorem 1.1]).
Quoted as Theorem 0.2; it converts the strong-exchange classifications into Veronese-type classifications throughout the paper.
domain assumption Complete multipartite graphs minus a matching are of Veronese type (Hibi and Seyed Fakhari, [6, Theorem 4.1]).
Used in Theorem 2.1 and in Lemmas 5.13 and 5.14, which depend on the Kronecker graph K_{4,2} minus a matching.
ad hoc to paper In Lemma 3.3, the residual ideal J in the variables of an independent set A_v (or A'_v), obtained after factoring out pure powers of the other variables, is polymatroidal.
The proof states this without proof or citation, then applies Lemma 3.2 to J; this is the main unstated premise of the positive classifications.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Bounded powers of edge ideals: The strong exchange property." pith.science (2026). https://pith.science/paper/THFL5U4X

@misc{pith2026250603480,
  author       = {Pith},
  title        = {Pith review of: Bounded powers of edge ideals: The strong exchange property},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/THFL5U4X}},
  note         = {Machine review of arXiv:2506.03480}
}

abstract

Let $S=K[x_1, \ldots,x_n]$ denote the polynomial ring in $n$ variables over a field $K$ and $I \subset S$ a monomial ideal. Given a vector $\mathfrak{c}\in\mathbb{Z}_{>0}^n$, the ideal $I_{\mathfrak{c}}$ is the ideal generated by those monomials belonging to $I$ whose exponent vectors are componentwise bounded above by $\mathfrak{c}$. Let $\delta_{\mathfrak{c}}(I)$ be the largest integer $q$ for which $(I^q)_{\mathfrak{c}}\neq 0$. Let $I(G) \subset S$ denote the edge ideal of a finite graph $G$ on the vertex set $V(G) = \{x_1, \ldots, x_s\}$. In our previous work, it is shown that $(I(G)^{\delta_{\mathfrak{c}}(I)})_{\mathfrak{c}}$ is a polymatroidal ideal. Let $\mathcal{W}(\mathfrak{c},G)$ denote the minimal system of monomial generators of $(I(G)^{\delta_{\mathfrak{c}}(I)})_{\mathfrak{c}}$. It follows that $\mathcal{W}(\mathfrak{c},G)$ satisfies the symmetric exchange property. In the present paper, the question when $\mathcal{W}(\mathfrak{c},G)$ enjoys the strong exchange property, or equivalently, when $\mathcal{W}(\mathfrak{c},G)$ is of Veronese type is studied.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Bounded powers of edge ideals: symmetric exchange binomials
math.AC 2025-07 conditional novelty 6.0 of 10

For paths and for graphs with maximal bounded-power degree 2, the toric ideal of the polymatroid B(G,c) is generated by symmetric exchange binomials.

Reference graph

Works this paper leans on

7 extracted references · 4 canonical work pages · cited by 1 Pith paper

[1]

De Negri and T

E. De Negri and T. Hibi, Gorenstein algebras of Veronese type,J. Algebra193(1997), 629–639

work page 1997
[2]

Herzog and T

J. Herzog and T. Hibi, Discrete polymatroids,J. Algebraic Combin.16(2002), 239–268

work page 2002
[3]

Monomial Ideals,

J. Herzog and T. Hibi, “Monomial Ideals,” GTM 260, Springer, 2011

2011
[4]

Herzog, T

J. Herzog, T. Hibi and M. Vladoiu, Ideals of fiber type and polymatroids,Osaka J. Math.42 (2005), 807–829

work page 2005
[5]

Hibi and S

T. Hibi and S. A. Seyed Fakhari, Bounded powers of edge ideals: regularity and linear quo- tients,Proc. Amer. Math. Soc., to appear, arXiv:2502.01768

arXiv
[6]

Hibi and S

T. Hibi and S. A. Seyed Fakhari, Bounded powers of edge ideals: Gorenstein toric rings, arXiv:2504.21760

arXiv
[7]

Ohsugi and T

H. Ohsugi and T. Hibi, Compressed polytopes, initial ideals and complete multipartite graphs, Illinois J. Math.44(2000), 391–406. (Takayuki Hibi) Department of Pure and Applied Mathematics, Graduate School of Information Science and Technology, Osaka University, Suita, Osaka 565–0871, Japan Email address:hibi@math.sci.osaka-u.ac.jp (Seyed Amin Seyed F akh...

work page 2000

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

De Negri and T

E. De Negri and T. Hibi, Gorenstein algebras of Veronese type,J. Algebra193(1997), 629–639

work page 1997

[2] [2]

Herzog and T

J. Herzog and T. Hibi, Discrete polymatroids,J. Algebraic Combin.16(2002), 239–268

work page 2002

[3] [3]

Monomial Ideals,

J. Herzog and T. Hibi, “Monomial Ideals,” GTM 260, Springer, 2011

2011

[4] [4]

Herzog, T

J. Herzog, T. Hibi and M. Vladoiu, Ideals of fiber type and polymatroids,Osaka J. Math.42 (2005), 807–829

work page 2005

[5] [5]

Hibi and S

T. Hibi and S. A. Seyed Fakhari, Bounded powers of edge ideals: regularity and linear quo- tients,Proc. Amer. Math. Soc., to appear, arXiv:2502.01768

arXiv

[6] [6]

Hibi and S

T. Hibi and S. A. Seyed Fakhari, Bounded powers of edge ideals: Gorenstein toric rings, arXiv:2504.21760

arXiv

[7] [7]

Ohsugi and T

H. Ohsugi and T. Hibi, Compressed polytopes, initial ideals and complete multipartite graphs, Illinois J. Math.44(2000), 391–406. (Takayuki Hibi) Department of Pure and Applied Mathematics, Graduate School of Information Science and Technology, Osaka University, Suita, Osaka 565–0871, Japan Email address:hibi@math.sci.osaka-u.ac.jp (Seyed Amin Seyed F akh...

work page 2000