Hanlon-Hicks-Lazarev resolution revisited

Lev Borisov, Zengrui Han · 2025

Read Hanlon-Hicks-Lazarev resolution revisited by Lev Borisov, Zengrui Han (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-11077.txt.

Introduction

Lev Borisov and Zengrui Han revisit a result from algebraic geometry in this short 2025 paper. Hanlon, Hicks, and Lazarev had built resolutions of the structure sheaves of toric substacks using certain line bundles on the surrounding toric stacks. Borisov and Han supply a new proof of that construction which they describe as substantially simpler than the original argument, making the result easier to check and to use.

Catalog record

CREATORLev Borisov, Zengrui Han
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE27 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-11077

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Hanlon-Hicks-Lazarev resolution revisited Lev Borisov, Zengrui Han Abstract . Hanlon, Hicks and Lazarev constructed resolutions of structure sheaves of toric substacks by certain line bundles on the ambient toric stacks. In this paper, we give a new and substantially simpler proof of their result. 1. Introduction Mirror symmetry for toric varieties has been extensively studied over the last thirty years, going back to the pioneering paper of Batyrev [ 2 ] . In particular, there has been a lot of work on bounded derived categories of toric varieties, inspired by the Homological Mirror Symmetry conjectures of Kontsevich. It was first observed by Bondal in [ 3 ] , that these categories are closely related to the derived categories of constructible sheaves on certain real tori with respect to stratifications induced by toric hyperplanes defined by ray generators of the fan. This perspective was further developed by Fang, Liu, Treumann, and Zaslow [ 8 ] who related torus-equivariant coherent sheaves on the toric varieties to constructible sheaves on the real tori with microsupport on conic Lagrangian subsets of the cotangent bundles of the real tori called FLTZ skeleta. These results recently inspired Hanlon, Hicks and Lazarev [ 11 ] to construct resolutions of structure sheaves of toric substacks in smooth toric stacks by line bundles on the ambient stacks, which we call Hanlon-Hicks-Lazarev resolutions, or HHL resolutions for short. Their construction has in turn already inspired further papers by multiple authors, see [ 1 , 6 , 9 , 12 ] . As far as we understand it, the original argument in [ 11 ] is based on a hyperplane removal strategy. By using discrete Morse theory, the HHL complex can be proven to be homotopy equivalent to the complex associated to the coarser stratification obtained by removing one toric hyperplane on an open subset. Then one keeps removing hyperplanes until arriving at a Koszul-type stratification, thereby obtaining a homotopy equivalence to the standard Koszul resolution on (an open subset) of an affine chart. Running this process for all affine charts proves the result over the complement of a codimension two subset. One then needs extra effort to extend to the whole variety or stack. The main goal of this paper is twofold. First, we provide a new and substantially simpler proof of the main result of [ 11 ] . In particular, our argument avoids induction on the maximal cones, the hyperplane removal strategy and the discrete Morse machinery. Second, we restrict to a minimal amount of input data, thereby making our setup cleaner and more concise. We hope this makes this topic more accessible and that our paper could serve as a more user-friendly introduction. Now we explain the strategy of our proof and the organization of this paper. In Section 2 , we study the HHL complexes on affine spaces. This is a purely commutative-algebra construction that requires little knowledge of toric geometry. More precisely, let L\cong{\mathbb{Z}}^{n} be a lattice of rank n with a {\mathbb{Z}} -basis \{v_{i}\} and let \psi:L\rightarrow\Lambda be a morphism from L to another lattice \Lambda of rank k such that \mathrm{coker}\,\psi is finite. We then define a complex of free modules over the polynomial ring R={\mathbb{C}}[x_{1},\cdots,x_{n}] which we call the affine HHL complex. Our main result is the following Theorem 2.9 , where the semigroup algebra {\mathbb{C}}[C\cap M] is defined in Section 2 . Theorem 2.9 . The only nonzero homology of the affine HHL complex ( 2.1 ) is H_{0}={\mathbb{C}}[C\cap M] . The key ingredient of the proof is a decomposition of the affine HHL complex into a direct sum of complexes of {\mathbb{C}} -vector spaces, where each piece is isomorphic to the cellular chain complex of some polyhedral region on the universal cover of the real torus. To get more general HHL complexes, as in [ 11 ] , it remains to pass from modules to sheaves on {\mathbb{C}}^{n} , then restrict to toric open subsets U_{\Sigma} of {\mathbb{C}}^{n} and finally use the equivariant nature of the construction to pass to quotients of U_{\Sigma} by abelian reductive groups, see Theorem 3.1 and Theorem 3.4 in Section 3 . Acknowledgements. We thank Daniel Erman and Andrew Hanlon for useful conversations. The first author was also inspired by an Oberwolfach talk by Christine Berkesch. 2. HHL complexes on affine spaces In this section we study the affine case, i.e., HHL complexes that are defined on affine spaces {\mathbb{C}}^{n} . The context of this section should be accessible and interesting to commutative algebraists or readers who are not familiar with toric geometry. By the homogeneous coordinate ring construction, the general HHL result for toric stacks follows almost effortlessly from this seemingly special situation. 2.1. Affine HHL complex We start with a minimal amount of input data. Consider a pair (L,\psi:L\rightarrow\Lambda) where: • L\cong{\mathbb{Z}}^{n} is a lattice of rank n with a {\mathbb{Z}} -basis \{v_{i}\} , • \psi:L\rightarrow\Lambda is a morphism from L to another lattice \Lambda of rank k such that \mathrm{coker}\,\psi is finite. Definition 2.1 . We define stratifications on the real torus \displaystyle T:=\operatorname{Hom}(\Lambda,{\mathbb{R}}/{\mathbb{Z}})=\Lambda^{*}_{{\mathbb{R}}}/\Lambda^{*}\cong(S^{1})^{k} and its universal cover \displaystyle\Lambda^{*}_{{\mathbb{R}}}=\operatorname{Hom}(\Lambda,{\mathbb{R}})\cong{\mathbb{R}}^{k} which we call Bondal stratifications . Each basis vector v_{i} of L defines a map \displaystyle H_{i}:\ \Lambda^{*}_{{\mathbb{R}}}\rightarrow{\mathbb{R}},\quad f\mapsto f(\psi(v_{i})). The hyperplanes in \Lambda^{*}_{{\mathbb{R}}} defined by H_{i}=a for a\in{\mathbb{Z}} together define a periodic hyperplane arrangement on the space \Lambda^{*}_{{\mathbb{R}}} , and we denote the induced stratification by \widetilde{S} . Note that they naturally descend to the torus T thereby inducing a stratification there which we denote by S . We denote the set of m -dimensional strata in S and \widetilde{S} by S_{m} and \widetilde{S_{m}} , respectively. The fact that \psi has finite cokernel implies that each stratum \widetilde{\sigma} of \widetilde{S} is a bounded convex polytope. For a stratum \widetilde{\sigma} in \widetilde{S} , we denote by H_{i}(\widetilde{\sigma}) the value of H_{i} at an arbitrary point in the relative interior of \widetilde{\sigma} . Note that \left\lceil H_{i}(\widetilde{\sigma})\right\rceil is well-defined and does not depend on the choice of the point. We denote the coordinate ring of {\mathbb{C}}^{n} by R={\mathbb{C}}[x_{1},\cdots,x_{n}] . From the stratification S we define a complex of free R -modules \displaystyle 0\rightarrow\bigoplus_{\begin{subarray}{c}\sigma\in S_{k}\end{subarray}}R\,e_{\sigma}\rightarrow\cdots\rightarrow\bigoplus_{\begin{subarray}{c}\sigma\in S_{1}\end{subarray}}R\,e_{\sigma}\rightarrow\bigoplus_{\begin{subarray}{c}\sigma\in S_{0}\end{subarray}}R\,e_{\sigma}\rightarrow 0, (2.1) which we call the affine HHL complex , where e_{\sigma} is the generator corresponding to the stratum \sigma . The differentials are defined as follows. We fix an arbitrary choice of orientations for all strata in the Bondal stratification on the torus T . Note that this also induces orientations for all lifts of the strata in the universal cover \Lambda^{*}_{{\mathbb{R}}} . Let \sigma be an m -dimensional stratum in S , and \tau an (m-1) -dimensional facet of \sigma . We take an arbitrary lift \widetilde{\sigma} of \sigma in the universal cover \Lambda^{*}_{{\mathbb{R}}} , and look at all facets of \widetilde{\sigma} that are mapped to \tau under the quotient map. Note that there may be multiple facets that are mapped to the same stratum on the torus. Let \widetilde{\tau} be such a facet. Then for any i , the difference \epsilon_{i}:=\left\lceil H_{i}(\widetilde{\sigma})\right\rceil-\left\lceil H_{i}(\widetilde{\tau})\right\rceil is either 0 or 1 . Therefore we have a map \displaystyle R\,e_{\sigma}\rightarrow R\,e_{\tau},\quad e_{\sigma}\mapsto\left(\mathrm{sgn}(\widetilde{\sigma},\widetilde{\tau})\prod x_{i}^{\epsilon_{i}}\right)e_{\tau} where \mathrm{sgn}(\widetilde{\sigma},\widetilde{\tau}) equals to 1 if the orientations on \widetilde{\sigma} and \widetilde{\tau} are compatible and -1

Read the complete text