[Paper Review] Resolutions of modules with initially linear syzygies

Emil Sköldberg|arXiv (Cornell University)|Jun 9, 2011

Commutative Algebra and Its Applications4 references7 citations

TL;DR

This paper introduces modules with initially linear syzygies—a class generalizing ideals with linear quotients—and constructs their minimal free resolutions using discrete Morse theory. It proves that such resolutions admit a differential graded algebra (DGA) structure, extending a known result for stable ideals and establishing it for the first time for squarefree matroidal ideals, including the Fano matroid ideal, with explicit multiplication computations provided.

ABSTRACT

We introduce the class of modules with initially linear syzygies, which includes ideals with linear quotients, and study their minimal resolutions. Using a contracting homotopy for the resolutions, we see that the minimal resolution of a matroidal monomial ideal admits a DGA structure.

Motivation & Objective

• To define and study the class of modules with initially linear syzygies, generalizing ideals with linear quotients.
• To construct minimal free resolutions of such modules using discrete Morse theory and contracting homotopies.
• To prove that modules with initially linear syzygies are componentwise linear.
• To establish that the minimal resolution of $ S/I $ admits a differential graded algebra (DGA) structure when $ I $ is a stable or squarefree matroidal ideal.
• To illustrate the multiplicative structure via explicit computation on the resolution of the Fano matroid ideal.

Proposed method

• Uses a presentation of a graded $ S $-module with a term order such that the initial module of syzygies is generated by elements of the form $ x_j g_i $, defining initially linear syzygies.
• Applies discrete Morse theory to the Koszul complex via a matching on the associated digraph, leading to a minimal resolution as a direct summand.
• Employs a contracting homotopy to construct the minimal resolution and analyze the differential in terms of reductions.
• Defines a critical set $ \mathrm{c\mbox{-}crit}(e_I g_\alpha) $ to control non-zero components in the differential and product structure.
• Uses induction on $ |I| + |J| $ to verify the DGA condition: $ \mathrm{c\mbox{-}crit}(e_I g_\alpha \star e_J g_\beta) \subseteq \mathrm{c\mbox{-}crit}(e_I g_\alpha) \cap \mathrm{c\mbox{-}crit}(e_J g_\beta) $.
• Performs explicit multiplication computations in the resolution of the Fano matroid ideal restricted to $ k[x_1,\dots,x_4] $, computing products like $ g_{124} \star g_{134} = x_1x_4 e_2 g_{134} $.

Experimental results

Research questions

• RQ1 Can the minimal resolution of a module with initially linear syzygies be constructed using discrete Morse theory and a contracting homotopy?
• RQ2 Under what conditions does the minimal resolution of $ S/I $ admit a differential graded algebra (DGA) structure?
• RQ3 Does the DGA structure extend from stable ideals to squarefree matroidal ideals, including the Fano matroid?
• RQ4 How can the multiplicative structure of the resolution be computed explicitly for small examples?
• RQ5 What is the role of the critical set $ \mathrm{c\mbox{-}crit} $ in controlling non-zero components of the differential and product?

Key findings

• The minimal resolution of a module with initially linear syzygies can be constructed via discrete Morse theory and a contracting homotopy, yielding a minimal free resolution.
• Modules with initially linear syzygies are componentwise linear, generalizing known results for ideals with linear quotients.
• The minimal resolution of $ S/I $ admits a DGA structure when $ I $ is a stable ideal or a squarefree matroidal ideal, with a new proof for stable ideals and a novel result for matroidal ideals.
• For the Fano matroid ideal restricted to $ k[x_1,\dots,x_4] $, the product $ g_{124} \star g_{134} = x_1x_4 e_2 g_{134} $, $ g_{124} \star g_{234} = x_2x_4 e_1 g_{234} $, and $ g_{134} \star g_{234} = x_3x_4 e_1 g_{234} - x_3x_4 e_2 g_{134} $ are explicitly computed.
• The DGA condition is verified via induction on $ |I| + |J| $, relying on the inclusion $ \mathrm{c\mbox{-}crit}(c(d(e_I g_\alpha) \star e_J g_\beta)) \subseteq \mathrm{c\mbox{-}crit}(e_I g_\alpha) \cap \mathrm{c\mbox{-}crit}(e_J g_\beta) $.