Syzygies of Polymatroidal Ideals

• Syzygies of polymatroidal ideals are captured through homological shift ideals, which detail the multidegree structure and resolution behavior.
• Recent advances show that the first homological shift preserves the polymatroidal property via explicit constructions using linear quotients and exchange properties.
• Open problems remain for higher shifts and degrees, fueling ongoing research at the interface of combinatorial optimization and algebraic geometry.

A polymatroidal ideal is a monomial ideal in a polynomial ring whose set of minimal generators encodes the bases of a discrete polymatroid through a multivariate exchange property. The study of syzygies of polymatroidal ideals—specifically, their higher homological shift ideals—serves as a bridge between combinatorial commutative algebra and discrete submodular geometry. Recent advances focus on classifying homological shift ideals associated to polymatroidal ideals, their closure properties under these operations, explicit resolution behavior, and the exploitation of fine exchange properties in advancing structure theory.

1. Polymatroidal Ideals: Definition and Exchange Properties

Let $S = K[x_1, \ldots, x_n]$ be a polynomial ring over a field $K$. A monomial ideal $I \subset S$ generated in degree $d$ is polymatroidal if its minimal generating set $G(I)$ is closed under a discrete exchange property. More precisely, for $$u = x_1^{a_1}\cdots x_n^{a_n},\, v = x_1^{b_1}\cdots x_n^{b_n} \in G(I)$$ and any $i$ with $a_i > b_i$, there exists $j$ with $a_j < b_j$ such that $x_j \cdot(u / x_i) \in G(I)$. Equivalently, in exponent notation, if $$\mathbf{a} = (a_1,\ldots,a_n),\, \mathbf{b} = (b_1,\ldots,b_n)$$, whenever $a_i > b_i$ there exists $j$ with $a_j < b_j$ and $x^{\mathbf{a} - e_i + e_j} \in G(I)$ (Ficarra et al., 2022).

In the squarefree case (all $a_k \in \{0,1\}$ for all $k$), this reduces to the matroidal basis-exchange property, making matroidal ideals squarefree polymatroidal ideals. The class of polymatroidal ideals includes Borel ideals, Veronese-type ideals, and edge ideals of complete multipartite graphs, among others (Ficarra, 2022, Bayati, 2023). The exchange property has both symmetric and asymmetric (strong exchange property) refinements that further restrict the monomial combinatorics, especially for Veronese and certain Borel type subclasses.

2. Syzygies and Homological Shift Ideals

The syzygies of a monomial ideal are reflected in the minimal multigraded free resolution

$$F_\bullet: \quad F_0 \leftarrow F_1 \leftarrow \cdots \leftarrow F_p \leftarrow 0,$$

with $$F_k = \bigoplus_{j=1}^{\beta_k(I)} S(-\mathbf{a}_{k,j})$$, where $\beta_k(I)$ is the $k$th Betti number and each shift $\mathbf{a}_{k,j} \in \mathbb{N}^n$ denotes the multidegree of a $k$th syzygy.

For each $k \geq 0$, the $k$th homological shift ideal (HSI) is defined as

$$HS_k(I) = ( x^{\mathbf{a}_{k,1}}, x^{\mathbf{a}_{k,2}}, \ldots, x^{\mathbf{a}_{k,\beta_k(I)}} ) \subset S,$$

so $HS_0(I) = I$, and $HS_k(I)$ records the multidegrees where $k$th syzygies occur (Ficarra et al., 2022, Ficarra, 2022, Ficarra et al., 15 Sep 2025).

If $I$ has linear quotients, there is an explicit combinatorial description (Herzog–Takayama):

$$HS_k(I) = (u_i \cdot x_A : 1 \leq i \leq m,~A \subseteq \mathrm{set}(u_i),~|A| = k),$$

where $\mathrm{set}(u_i)$ is the set of variable indices appearing in $(u_1,\ldots,u_{i-1}):u_i$ (Ficarra et al., 2022, Bayati, 2023).

The $k$th homological shifts are central for tracking how the syzygies and the combinatorial data of a monomial ideal interact, both in terms of degree and multidegree.

3. The Bandari–Bayati–Herzog Conjecture and Its Status

The Bandari–Bayati–Herzog conjecture asserts that if $I$ is a polymatroidal ideal, then all homological shift ideals $HS_k(I)$ are themselves polymatroidal for every $k \geq 0$ (Ficarra et al., 2022, Ficarra, 2022). The rationale is that the hereditary nature of the base exchange property would be preserved under the passage from generators to syzygies and analogously higher-order syzygies.

Key status points, grounded in recent research, are as follows:

• $k=1$ Case: For every polymatroidal ideal $I$, $HS_1(I)$ is polymatroidal. This was established directly via linear quotients using both distance and adjacency arguments (Ficarra, 2022, Bayati, 2023). The first homological shift can be described as the adjacency ideal:

$$HS_1(I) = ( \mathrm{lcm}(u, v) : u, v \in G(I),\,d(u, v) = 1 ),$$

where $$d(u,v) = \frac{1}{2}\sum_{i=1}^n |\deg_i(u)-\deg_i(v)|$$ (Ficarra, 2022, Bayati, 2023).

• Matroidal (Squarefree) Case: For matroidal $I$, Bayati showed all higher $HS_k(I)$ are also matroidal (Ficarra, 2022). The relation $HS_{k+1}(I) = HS_1(HS_k(I))$ (up to support considerations) holds.
• Strong Exchange Property: For polymatroidals with the strong exchange property, Herzog–Moradi–Rahimbeigi–Zhu showed all $HS_k(I)$ are polymatroidal (Ficarra et al., 2022).
• Degree-Two Polymatroidal Ideals: The theorem of Ficarra–Herzog (Ficarra et al., 2022) proves for degree-two polymatroidal ideals that all $HS_k(I)$ remain polymatroidal:

$$\text{If } I \subset K[x_1,\ldots,x_n] \text{ is polymatroidal, generated in degree 2, then } HS_k(I) \text{ is polymatroidal } \forall k \geq 0.$$

• General Case: For higher degrees, the conjecture is unresolved. The methods for the degree-two case, such as reduction to squarefree and variable-square pieces, do not directly generalize (Ficarra et al., 2022, Ficarra, 2022).

4. Quasi-Additivity and Linear Quotients

A crucial property intersecting the study of polymatroidal syzygies is quasi-additivity: for which pairs $(i,j)$ do

$HS_{i+j}(I) \subseteq HS_i(HS_j(I))$

hold? For polymatroidal ideals, it has been established that:

• $HS_{i+1}(I) \subseteq HS_i(HS_1(I))$ for all $i$ (Bayati, 2023). This follows because all polymatroidals and their first shifts have linear quotients.
• For polymatroidal ideals satisfying the strong-exchange property, and for all degree-two polymatroidals, the full inclusion $HS_{i+j}(I) \subseteq HS_i(HS_j(I))$ holds for all $i, j$ (Bayati, 2023).
• For matroidal (i.e., squarefree) polymatroidal ideals, indeed $HS_{i+j}(I) = HS_i(HS_j(I))$ for all $i,j$.

The table below summarizes the closure properties for classes of ideals under homological shift operations:

Ideal Class	All $HS_k(I)$ polymatroidal?	Quasi-additivity $HS_{i+j}(I) \subseteq HS_i(HS_j(I))$
Polymatroidal (all degrees)	$k=1$: Yes; $k>1$: Open	$j=1$: Yes
Polymatroidal (degree 2)	Yes	Yes, for all $i,j$
Matroidal (squarefree polymatroidal)	Yes	Equality for all $i, j$
Strong-exchange polymatroidal	Yes	Yes, for all $i,j$

Linear quotient structure is essential, as it enables explicit construction of resolutions and codifies quasi-additive behaviors (Ficarra et al., 2022, Bayati, 2023).

5. Asymptotic and Persistence Properties

Recent work extends the study of syzygies of polymatroidal ideals to their powers and Rees algebras (Ficarra et al., 15 Sep 2025). Let $\mathcal{R}(I) = \bigoplus_{k \geq 0} I^k t^k$ denote the Rees algebra, and define the $i$th homological shift algebra as $$HS_i(\mathcal{R}(I)) = \bigoplus_{k \geq 1} HS_i(I^k) t^k$$.

Key results and conjectures:

• For any polymatroidal $I$, $HS_1(\mathcal{R}(I))$ is generated over $\mathcal{R}(I)$ in $t$-degree one: $$HS_1(I^{k+1}) = I \cdot HS_1(I^k),~\forall k \geq 1$$ (Ficarra et al., 15 Sep 2025).
• Conjecture: For any $i$, $HS_i(\mathcal{R}(I))$ is generated in $t$-degrees $\leq i$ (i.e., eventually $HS_i(I^{k+1}) = I \cdot HS_i(I^k)$ for $k \geq i$), proven for principal Borel, strong-exchange, and matroidal ideals (Ficarra et al., 15 Sep 2025).
• Persistence of Associated Primes: For all $k$,

$Ass\,HS_i(I^k) \subseteq Ass\,HS_i(I^{k+1}),$

holds for $i=1$ in general, and for many ideals when $i>1$, providing a strong-homological persistence property (Ficarra et al., 15 Sep 2025).

• For componentwise polymatroidal ideals, $HS_1(I)$ is again componentwise polymatroidal, with graded shift ideals respecting the componentwise structure (Ficarra et al., 15 Sep 2025).

6. Minimal Resolutions and Connections to Subspace Arrangements

The syzygies of certain multiplicative monomial ideals associated to subspace arrangements are governed directly by polymatroid and matroid theory (Conca et al., 2019). If $J = I_1\cdots I_n$ where each $I_j$ is generated by linear forms arising from a subspace $V_j$, then the minimal free resolution of $J$ can be constructed via the Dilworth truncation of the associated representable polymatroid.

The Betti numbers and projective dimension are then controlled by the integer points in the truncated polymatroid polytope. Moreover, the resolution admits linear quotients, with syzygies corresponding to flats and circuits in the truncated polymatroid, thus providing a deep geometric and combinatorial interpretation of the syzygy modules.

7. Outlook and Open Problems

• The Bandari–Bayati–Herzog conjecture remains open for general polymatroidal ideals in degrees $d \geq 3$, as current methods are sensitive to the degree and do not generalize from the base case.
• Determining precise Betti tables, support, and additional homological invariants for $HS_k(I)$, especially for higher $k$, is an active area of investigation.
• The asymptotic properties and strong persistence results suggest that polymatroidal ideals are exceptionally stable from the viewpoint of syzygetic complexity.
• Extending these behaviors to the larger class of componentwise polymatroidal or componentwise linear ideals is a significant direction.
• Structure theory for non-squarefree and non-strong-exchange polymatroidals under repeated homological shifting is presently incomplete.

These results collectively position the study of syzygies of polymatroidal ideals as a central, highly structured instance of the interaction between combinatorial optimization, monomial geometry, and homological algebra (Ficarra et al., 2022, Ficarra, 2022, Bayati, 2023, Ficarra et al., 15 Sep 2025, Conca et al., 2019).