Skip to main content
QUICK REVIEW

[Paper Review] The cone of Betti diagrams of bigraded artinian modules of codimension two

Mats Boij, Gunnar Fløystad|arXiv (Cornell University)|Jan 19, 2010
Algebraic structures and combinatorial models4 references3 citations
TL;DR

This paper characterizes the positive cone of bigraded Betti diagrams for artinian modules of codimension two over a bigraded polynomial ring. It shows that when the total degree differences $e_1$ and $e_2$ are relatively prime, the extremal rays of the cone are parametrized by order ideals in $\mathbb{N}^2$ within the region $e_1x + e_2y < (e_1-1)(e_2-1)$, with each such ideal corresponding to a unique pure resolution. The construction generalizes both the equivariant resolution and the monomial quotient resolution, providing a complete classification of extremal rays in this setting.

ABSTRACT

We describe the positive cone generated by bigraded Betti diagrams of artinian modules of codimension two, whose resolutions become pure of a given type when taking total degrees. If the differences of these total degrees, p and q, are relatively prime, the extremal rays are parametrised by order ideals in N^2 contained in the region px + qy &lt; (p-1)(q-1). We also consider some examples concerning artinian modules of codimension three.

Motivation & Objective

  • To describe the positive cone generated by bigraded Betti diagrams of artinian modules of codimension two whose resolutions become pure upon taking total degrees.
  • To classify the extremal rays of this cone when the total degree differences $e_1$ and $e_2$ are relatively prime.
  • To generalize the construction of pure resolutions beyond the equivariant and monomial quotient resolutions, showing that all extremal rays arise from order ideals in a specific region of $\mathbb{N}^2$.
  • To explore whether similar classification results extend to codimension three modules, particularly regarding finiteness and structure of extremal rays.

Proposed method

  • The authors use multigraded Herzog-Kühl equations to constrain possible Betti diagrams of bigraded artinian modules.
  • They define a region $R(e_1,e_2)$ in $\mathbb{N}^2$ bounded by $e_1x + e_2y < (e_1-1)(e_2-1)$, which contains all relevant order ideals.
  • For each order ideal $\lambda \subset R(e_1,e_2)$, they construct a pure resolution of the form $S^{e_2} \leftarrow S^{e_2 + e_1} \leftarrow S^{e_1}$ with specific bigraded Betti numbers.
  • They prove that the Betti diagrams $\beta_\lambda(\mathbf{a})$ for $\mathbf{a} \in \mathbb{Z}^2$ generate the extremal rays of the positive cone $P(e_1,e_2)$, using degeneration techniques and kernel analysis on general matrices.
  • They verify that these diagrams correspond to actual resolutions by constructing explicit complexes via matrix degenerations and verifying the Herzog-Kühl equations.
  • For codimension three, they analyze the cone $P(1,2,1)$ and show that not all diagrams in the relaxed cone $P'(1,2,1)$ arise from actual modules, indicating a richer structure beyond the codimension two case.

Experimental results

Research questions

  • RQ1Are there only finitely many extremal rays in the positive cone of Betti diagrams for codimension two bigraded artinian modules with fixed total degree differences?
  • RQ2Can the extremal rays of the positive cone of codimension two Betti diagrams be parametrized by combinatorial objects such as order ideals in $\mathbb{N}^2$?
  • RQ3Does the construction of pure resolutions via order ideals in $R(e_1,e_2)$ yield all possible extremal rays in the positive cone $P(e_1,e_2)$ when $\gcd(e_1,e_2) = 1$?
  • RQ4In the codimension three case, is the positive cone $P(\mathbf{e})$ isomorphic to the relaxed cone $P'(\mathbf{e})$, or are there diagrams in $P'(\mathbf{e})$ not realizable by actual modules?
  • RQ5Are the extremal rays in the codimension three case still generated by a finite set of translation classes, and does a unique maximal or minimal element exist in the poset of such rays?

Key findings

  • When $e_1$ and $e_2$ are relatively prime, the extremal rays of the positive cone $P(e_1,e_2)$ are in bijection with order ideals $\lambda$ contained in the region $R(e_1,e_2) = \{ (x,y) \in \mathbb{N}^2 \mid e_1x + e_2y < (e_1-1)(e_2-1) \}$.
  • The maximal order ideal corresponds to the equivariant resolution, and the empty order ideal corresponds to the resolution of a quotient of monomial ideals constructed in [1].
  • Each order ideal $\lambda$ gives rise to a pure resolution of type $S^{e_2} \leftarrow S^{e_2 + e_1} \leftarrow S^{e_1}$, with bigraded Betti numbers determined by the shape of $\lambda$, and these generate the extremal rays of $P(e_1,e_2)$.
  • In the codimension three case, the cone $P(1,2,1)$ contains diagrams not realizable by actual modules—e.g., a non-negative linear combination $[(2,1,0)+(0,2,1)+(1,0,2)-(1,1,1)]\beta$ satisfies the Herzog-Kühl equations but does not arise from a module, showing $P(1,2,1) \subsetneq P'(1,2,1)$.
  • The Betti diagram $\alpha = [(2,1,0)+(2,0,1)+(1,2,0)+(0,2,1)+(1,0,2)+(0,1,2)-(1,1,1)]\beta$ is shown to be a Betti diagram of a resolution of an indecomposable artinian module and generates an extremal ray in $P(1,2,1)$, distinct from the equivariant diagram $\beta$.
  • The existence of such diagrams that are not multiples of a single ray indicates that the structure of the positive cone in codimension three is more complex than in codimension two, with no known complete classification of extremal rays.

Better researchstarts right now

From reading papers to final review, dramatically reduce your research time.

No credit card · Free plan available

This review was created by AI and reviewed by human editors.