Homological Shift Ideals Overview

• Homological shift ideals are monomial ideals generated by the multidegrees of syzygies in a minimal free resolution, encapsulating key homological features.
• They link minimal free resolution structures to combinatorial classes like Borel and polymatroidal ideals, facilitating analysis of linear quotients and quasi-additivity.
• Applications span explicit study of edge, Veronese, and almost Cohen–Macaulay ideals, providing insights into stabilization, regularity, and persistence phenomena.

Homological shift ideals are a central construction in the homological and combinatorial study of monomial ideals, encoding the supports of multigraded syzygies directly as new monomial ideals. They connect the structure of minimal free resolutions to combinatorial classes such as Borel, polymatroidal, and cover ideals, and provide a framework for understanding stable and asymptotic properties of syzygies, regularity, and associated primes in families of monomial ideals. Applications range from explicit analyses of edge, Borel, and Veronese-type ideals to the classification of almost Cohen–Macaulay codimension two ideals via prescribed shifts. The theory integrates with classical results from graph theory (including Dirac’s and Fröberg’s theorems), combinatorics of polymatroids, and asymptotic commutative algebra.

1. Definition and Fundamental Principles

Given a field $K$ and a polynomial ring $S=K[x_1,\dots,x_n]$, let $I\subset S$ be a monomial ideal. Its minimal multigraded free resolution takes the form

$$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$

with each $F_i \cong \bigoplus_{j} S(-a_{i,j})$ in multidegree $a_{i,j}\in \mathbb{N}^n$. The $i$th homological shift ideal is

$$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$

where $\beta_i(I)$ is the $i$th total Betti number of $S=K[x_1,\dots,x_n]$0. Thus, $S=K[x_1,\dots,x_n]$1 is the monomial ideal generated by all degrees in which $S=K[x_1,\dots,x_n]$2th syzygies of $S=K[x_1,\dots,x_n]$3 occur, capturing “the ideal of all $S=K[x_1,\dots,x_n]$4‐syzygy multidegrees” (Herzog et al., 2020, Ficarra et al., 2022, Ficarra, 2023). For $S=K[x_1,\dots,x_n]$5 one recovers $S=K[x_1,\dots,x_n]$6 itself. This definition extends to arbitrary powers $S=K[x_1,\dots,x_n]$7 and underlies the study of homological shift algebras,

$S=K[x_1,\dots,x_n]$8

viewed as a module over the Rees algebra $S=K[x_1,\dots,x_n]$9 (Ficarra et al., 2024, Ficarra et al., 15 Sep 2025).

The homological shift ideal $I\subset S$0 reflects the multigraded support of $I\subset S$1 or, in Betti language, is generated by all monomials $I\subset S$2 with $I\subset S$3.

2. Core Structural Properties

2.1 Linear Resolutions and Linear Quotients

If $I\subset S$4 is equigenerated and has linear quotients, then $I\subset S$5 inherits linear quotients, and thus a linear resolution. Explicitly, given an admissible order $I\subset S$6 on the minimal generators $I\subset S$7, one has

$I\subset S$8

and induction on $I\subset S$9 with mapping cone arguments yields the desired ordering (Herzog et al., 2020, Ficarra et al., 2022, Ficarra, 2023).

For higher shifts $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$0 or non-equigenerated $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$1, $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$2 need not have linear quotients, even if $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$3 is an edge ideal with linear resolution [(Ficarra et al., 2022), Example 1.4].

2.2 Polymatroidal and Borel Ideals

If $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$4 is polymatroidal, $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$5 is again polymatroidal (Ficarra, 2022), and this property extends to all shifts if $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$6 satisfies the strong exchange property or is generated in degree $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$7 (Ficarra et al., 2022, Bayati, 2023). In the principal squarefree Borel case $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$8, the $$F_\bullet: 0\to F_q \to \cdots \to F_1 \to F_0 \to S/I \to 0,$$9th shift is again principal Borel, $F_i \cong \bigoplus_{j} S(-a_{i,j})$0, with $F_i \cong \bigoplus_{j} S(-a_{i,j})$1 the largest $F_i \cong \bigoplus_{j} S(-a_{i,j})$2 gaps of $F_i \cong \bigoplus_{j} S(-a_{i,j})$3 (Herzog et al., 2021).

A conjecture due to Bandari–Bayati–Herzog posits $F_i \cong \bigoplus_{j} S(-a_{i,j})$4 is polymatroidal for all $F_i \cong \bigoplus_{j} S(-a_{i,j})$5 whenever $F_i \cong \bigoplus_{j} S(-a_{i,j})$6 is polymatroidal (Ficarra et al., 2022, Ficarra, 2022, Ficarra et al., 15 Sep 2025).

2.3 Quasi-Additivity and Maximal Shifts

For key classes (principal Borel, degree-2 polymatroidal, strong exchange polymatroidal, and squarefree Borel ideals), subadditivity/quasi-additivity holds: $F_i \cong \bigoplus_{j} S(-a_{i,j})$7 and is an equality for squarefree Borel ideals (Bayati, 2023). This formation controls the propagation of Betti degrees in higher syzygies.

2.4 Asymptotics and Strong Persistence

The homological shift algebra $F_i \cong \bigoplus_{j} S(-a_{i,j})$8 is a finitely generated $F_i \cong \bigoplus_{j} S(-a_{i,j})$9–module (Ficarra et al., 2024, Ficarra et al., 15 Sep 2025), so for $a_{i,j}\in \mathbb{N}^n$0, stability phenomena emerge, including stabilization of associated primes ($a_{i,j}\in \mathbb{N}^n$1), depth, regularity (eventually linear in $a_{i,j}\in \mathbb{N}^n$2), and the $a_{i,j}\in \mathbb{N}^n$3-number. For polymatroidal and edge ideals, the $a_{i,j}\in \mathbb{N}^n$4st shift satisfies

$a_{i,j}\in \mathbb{N}^n$5

implying the $a_{i,j}\in \mathbb{N}^n$6st homological strong persistence property and that the associated prime chain is increasing (Ficarra et al., 13 Jan 2025, Ficarra et al., 15 Sep 2025).

3. Homological Shift Ideals and Graph Ideals

3.1 Edge Ideals, Chordality, and Dirac’s Theorem

For a graph $a_{i,j}\in \mathbb{N}^n$7 with edge ideal $a_{i,j}\in \mathbb{N}^n$8, Fröberg’s theorem asserts $a_{i,j}\in \mathbb{N}^n$9 has a $i$0-linear resolution if and only if $i$1 is chordal (Ficarra et al., 2022). In this case, $i$2 is generated by all squarefree monomials on $i$3 vertices, corresponding to certain connectivity constraints in $i$4 (Herzog et al., 2020, Ficarra et al., 2022, Chau et al., 14 Mar 2025).

If $i$5 is a reversible chordal graph (e.g., a proper interval graph or forest), then all shifts $i$6 have linear quotients (Ficarra et al., 2022); this is not true for arbitrary co-chordal graphs (Chau et al., 14 Mar 2025), with precise forbidden subgraph obstructions ($i$7 for $i$8).

3.2 Vertex Cover and Complementary Edge Ideals

For certain cover ideals $i$9 (vertex cover ideals) and complementary edge ideals $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$0, the shifts are structurally tractable; under combinatorial decompositions (e.g., Betti splitting, partitioning), all shifts may have linear quotients or even be weakly polymatroidal, particularly for chordal, Cameron–Walker, or clique-corona graphs (Roy et al., 2 Jun 2025, Lu et al., 17 Nov 2025).

4. Asymptotic Syzygies, Persistence, and Golodness

The shift algebras $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$1 for $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$2 with linear powers satisfy robust asymptotic properties:

• Regularity: $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$3 is eventually linear in $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$4 (Ficarra et al., 2024, Ficarra et al., 15 Sep 2025).
• Associated primes and depth: stabilize for $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$5 (Ficarra et al., 13 Jan 2025, Ficarra et al., 15 Sep 2025).
• Strong persistence property: For edge and polymatroidal ideals, $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$6 sequences of associated primes are ascending, governed by the generation behavior $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$7 (Ficarra et al., 13 Jan 2025, Ficarra et al., 15 Sep 2025).
• Golod property: For all $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$8, $$\operatorname{HS}_i(I) = \left( x^{a_{i,j}} : j = 1,\dots,\beta_i(I) \right),$$9 is Golod whenever $\beta_i(I)$0 has linear powers; this implies vanishing of Massey products and simplifies Poincaré series computations (Ficarra et al., 2024).

5. Applications and Explicit Descriptions

5.1 Borel, Veronese, and Polymatroidal Ideals

Principal $\beta_i(I)$1-Borel ideals, Veronese-type ideals, and principal Borel ideals have explicitly describable shifts. For $\beta_i(I)$2, $\beta_i(I)$3th shifts are $\beta_i(I)$4 where $\beta_i(I)$5 are the largest gaps, yielding recursive descriptions of height, multiplicity, and analytic spread (Herzog et al., 2021).

Veronese-type and polymatroidal ideals with strong exchange properties have shifts coinciding with certain truncations or ideals of bounded support, and all such shifts retain polymatroidality (Herzog et al., 2020).

5.2 Graph Applications and Almost Cohen–Macaulay Ideals

Graph-theoretic interpretations allow explicit computations for edge ideals of paths, trees, cycles, and multipartite graphs. For almost Cohen–Macaulay, 3-generated codimension 2 ideals, the entire sequence of shifts is numerically codified as latent shifts, with such ideals classified by level matrices whose maximal minors realize the prescribed shifts (Burity et al., 19 Mar 2026).

6. Open Problems and Future Directions

Principal open questions include:

• The Bandari–Bayati–Herzog conjecture for all polymatroidal ideals remains unresolved for $\beta_i(I)$6 (Ficarra et al., 2022, Ficarra, 2022).
• Classification of all monomial ideals (or graph classes) for which every shift ideal is (weakly) polymatroidal or has linear quotients. Negative examples exist for Cohen–Macaulay whiskered/bipartite graphs (Roy et al., 2 Jun 2025), but positive results and characterizations in special graph classes are ongoing.
• Generalization of subadditivity and quasi-additivity to broader classes beyond Borel and polymatroidal ideals (Bayati, 2023, Adiprasito et al., 2024).
• Structural consequences for other invariants, including projective dimension and Betti sequence tail bounds (Herzog et al., 2020).
• Relations to singularity invariants and geometric modeling in codimension two and beyond (Burity et al., 19 Mar 2026).

These directions highlight the interplay between discrete combinatorics, minimal free resolutions, and homological algebra in the study of homological shift ideals.

---

Key References:

• Dirac's theorem and multigraded syzygies (Ficarra et al., 2022)
• Homological shift ideals (Herzog et al., 2020)
• Polymatroidal ideals and their asymptotic syzygies (Ficarra et al., 15 Sep 2025)
• Homological shifts of polymatroidal ideals (Ficarra, 2022)
• Some homological properties of Borel type ideals (Herzog et al., 2021)
• Homological shifts of a complementary edge ideal (Lu et al., 17 Nov 2025)
• Edge ideals with linear quotients and without homological linear quotients (Chau et al., 14 Mar 2025)
• Characterizing almost Cohen-Macaulay $\beta_i(I)$7-generated ideals of codimension $\beta_i(I)$8 in terms of prescribed shift (Burity et al., 19 Mar 2026)
• The homological shift algebra of a monomial ideal (Ficarra et al., 2024)
• A quasi-additive property of homological shift ideals (Bayati, 2023)