Admissible subcategories and metric techniques
Kabeer Manali Rahul · 2025
Read Admissible subcategories and metric techniques by Kabeer Manali Rahul (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-2504-11772.txt.
Introduction
Kabeer Manali Rahul builds a machine for making new decompositions of complex algebraic categories in this 2025 paper. A semiorthogonal decomposition splits a category into ordered pieces that barely interact, and such splits are prized in algebraic geometry for the structure they reveal.
The new tool leans on a special kind of distance measure on the category, which the author calls compressible. Starting from one decomposition on a category with such a measure, the method constructs a fresh decomposition on a new category built from the old one through that distance. For geometry, that yields new splits for the small triangulated categories attached to a scheme. The author links the work to earlier results by several groups and frames it as a general construction, not a one-case trick.
Catalog record
| CREATOR | Kabeer Manali Rahul |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 130 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-11772
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Admissible subcategories and metric techniques Kabeer Manali Rahul Abstract . In this work, we provide a way of constructing new semiorthogonal decompositions using metric techniques (à la Neeman). Given a semiorthogonal decomposition on a category with a special kind of metric, which we call a compressible metric, we can construct new semiorthogonal decomposition on a category constructed from the given one using the aforementioned metric. In the algebro-geometric setting, this gives us a way of producing new semiorthogonal decompositions on various small triangulated categories associated to a scheme, if we are given one. In the general setting, the work is related to that of Sun-Zhang, while its applications to algebraic geometry are related to the work of Bondarko and Kuznetsov-Shinder. 2020 Mathematics Subject Classification 14A30 (primary), 18G80, 14A22 1. Introduction Triangulated categories appear in diverse areas of mathematics, from geometry, to topology, to representation theory. Hence, there has been a lot of work on trying to understand the structure of triangulated categories. One of the most important tools for the same are semiorthogonal decompositions. These allow us to understand a complicated triangulated category by breaking it up into simpler pieces. These have found a lot of applications, especially in algebraic geometry and representation theory, including applications on the connections between these areas. With this in mind, it becomes an interesting problem to discover new semiorthogonal decompositions. There has been recent work in constructing new semiorthogonal decompositions from a given one with applications to algebraic geometry. The results are of the following form; we are given a semiorthogonal decomposition on a triangulated category \mathsf{S}_{1} , and we are able to produce a semiorthogonal decomposition on a different triangulated category \mathsf{S}_{2} which is related to \mathsf{S}_{1} in some manner. Two important recent papers which prove results of this form are [ KS23 ] and [ Bon23 ] . The main results of this work are also of this form. We will compare the results with the ones existing in the literature at the end of this section. To give the reader a more concrete idea about what these statements look like, let us state some applications of the main abstract results proved in this paper to the context of algebraic geometry. We give a consolidated result for schemes, algebras, and stacks. The conditions imposed in the theorem below are more stringent than required in some cases for ease of exposition. See the corresponding results in the main body of this work for the more general statements. Theorem A . Let X be a noetherian, separated, finite-dimensional scheme, and \mathcal{A} a coherent \mathcal{O}_{X} -algebra. Let \mathcal{X} be a concentrated noetherian stack (concentrated means that the canonical map \mathcal{X}\to\operatorname{Spec}(\mathbb{Z}) is concentrated, see [ HR17 , Definition 2.4] ) such that one of the conditions below hold, • \mathcal{X} has quasi-finite and separated diagonal. • \mathcal{X} is a Deligne-Mumford stack of characteristic zero. Then, for any semiorthogonal decomposition \langle\mathsf{A},\mathsf{B\rangle} on \mathsf{S}_{1} , we get a semiorthogonal decomposition \langle\operatorname{Coprod}(\mathsf{A})\cap\mathsf{S}_{2},\operatorname{Coprod}(\mathsf{B})\cap\mathsf{S}_{2}\rangle on \mathsf{S}_{2} for any (\mathsf{S}_{1},\mathsf{S}_{2}) as in the table below. Note that \operatorname{Coprod}(\mathsf{A}) denotes the localising subcategory generated by \mathsf{A} inside the derived category of quasicoherent sheaves (over the scheme/algebra/stack in question). Extra conditions \mathsf{S}_{1} \mathsf{S}_{2} Reference – \mathbf{D}^{\operatorname{perf}}(X) \mathbf{D}^{-}_{\operatorname{coh}}(X) Remark 5.9 \mathsf{B} is admissible \mathbf{D}^{\operatorname{perf}}(X) \mathbf{D}^{b}_{\operatorname{coh}}(X) Remark 5.9 X is J-2 \mathbf{D}^{b}_{\operatorname{coh}}(X) \mathbf{D}^{+}_{\operatorname{coh}}(X) Theorem 5.10 – \mathbf{D}^{\operatorname{perf}}(\mathcal{A}) \mathbf{D}^{-}_{\operatorname{coh}}(\mathcal{A}) Theorem 5.12 \mathsf{B} is admissible \mathbf{D}^{\operatorname{perf}}(\mathcal{A}) \mathbf{D}^{b}_{\operatorname{coh}}(\mathcal{A}) Theorem 5.12 X is J-2 \mathbf{D}^{b}_{\operatorname{coh}}(\mathcal{A}) \mathbf{D}^{+}_{\operatorname{coh}}(\mathcal{A}) Theorem 5.13 X is J-2 & \mathsf{B} is admissible \mathbf{D}^{b}_{\operatorname{coh}}(\mathcal{A}) \mathbf{D}^{b}_{\operatorname{coh}}(\operatorname{Inj}\mathcal{A}) Theorem 5.13 – \mathbf{D}^{\operatorname{perf}}(\mathcal{X}) \mathbf{D}^{-}_{\operatorname{coh}}(\mathcal{X}) Theorem 5.14 \mathsf{B} is admissible \mathbf{D}^{\operatorname{perf}}(\mathcal{X}) \mathbf{D}^{b}_{\operatorname{coh}}(\mathcal{X}) Theorem 5.14 This theorem follows from more general results proved in Section 3 . We now get into the details of those results, but again, not in their full generality. We begin with the following definition. Note that some of the notation below is only used in the introduction. Definition A . Let \mathsf{T} be a compactly generated triangulated category with a single compact generator G . Let \big({}^{\mathrm{I}}\mathsf{T}_{G}^{\leq 0},{}^{\mathrm{I}}\mathsf{T}_{G}^{\geq 0}\big) and \big({}^{\mathrm{II}}\mathsf{T}_{G}^{\geq 0},{}^{\mathrm{II}}\mathsf{T}_{G}^{\leq 0}\big) denote the standard t-structure and co-t-structure compactly generated by G respectively, see Definition 2.5 and Definition 2.7 . We define the following full subcategories of \mathsf{T} , • \mathsf{T}^{-}_{c}\mathrel{\mathchoice{\vbox{\hbox{$\displaystyle:$}}}{\vbox{\hbox{$\textstyle:$}}}{\vbox{\hbox{$\scriptstyle:$}}}{\vbox{\hbox{$\scriptscriptstyle:$}}}{=}}\bigcap_{n\in\mathbb{Z}}\mathsf{T}^{c}\star{}^{\mathrm{I}}\mathsf{T}_{G}^{\leq-n} , that is, \mathsf{T}^{-}_{c} is the full subcategory of \mathsf{T} consisting of all objects F such that for all n\in\mathbb{Z} , there exists a triangle E_{n}\to F\to D_{n}\to\Sigma E_{n} for a compact object E_{n} and with D_{n}\in{}^{\mathrm{I}}\mathsf{T}_{G}^{\leq-n} . • {}^{\mathrm{I}}\mathsf{T}^{b}_{c}=\{F\in\mathsf{T}^{-}_{c}:\operatorname{Hom}_{\mathsf{T}}(\Sigma^{i}G,F)=0 for i>>0\} • \mathsf{T}^{+}_{c}\mathrel{\mathchoice{\vbox{\hbox{$\displaystyle:$}}}{\vbox{\hbox{$\textstyle:$}}}{\vbox{\hbox{$\scriptstyle:$}}}{\vbox{\hbox{$\scriptscriptstyle:$}}}{=}}\bigcap_{n\in\mathbb{Z}}\mathsf{T}^{c}\star{}^{\mathrm{II}}\mathsf{T}_{G}^{\geq n} , that is, \mathsf{T}^{+}_{c} is the full subcategory of \mathsf{T} consisting of all objects F such that for all n\in\mathbb{Z} , there exists a triangle E_{n}\to F\to D_{n}\to\Sigma E_{n} for a compact object E_{n} and with D_{n}\in{}^{\mathrm{II}}\mathsf{T}_{G}^{\geq n} . • {}^{\mathrm{II}}\mathsf{T}^{b}_{c}=\{F\in\mathsf{T}^{+}_{c}:\operatorname{Hom}_{\mathsf{T}}(\Sigma^{i}G,F)=0 for i<<0\} With this in mind, we introduce the following hypotheses. Hypothesis . Let \mathsf{T} be a compactly generated triangulated category with a single compact generator G . Then, (I) \mathsf{T} satisfies Hypothesis (I) if \operatorname{Hom}_{\mathsf{T}}(\Sigma^{-i}G,G)=0 for i>>0 . (II) \mathsf{T} satisfies Hypothesis (II) if \operatorname{Hom}_{\mathsf{T}}(\Sigma^{i}G,G)=0 for i>>0 . We note here that there are plenty of examples where these hypotheses hold. Example A . Hypothesis (I) holds for the following categories, • \mathbf{D}_{\operatorname{Qcoh}}(X) for a noetherian scheme X . • \mathbf{D}_{\operatorname{Qcoh}}(\mathcal{A}) for a quasicoherent \mathcal{O}_{X} -algebra over a noetherian scheme X . • \mathbf{D}_{\operatorname{Qcoh}}(\mathcal{X}) for a 1-Thomason stack, see [ HR17 , Definition 8.1] . Hypothesis (II) holds for the following categories, • \mathbf{K}(\operatorname{Inj}X) for a noetherian J-2 scheme X . • \mathbf{K}_{m}(\operatorname{Proj}X) for a noetherian, separated, J-2 scheme X , where \mathbf{K}_{m}(\operatorname{Proj}X) denote the mock homotopy category of projectives as defined in [ Mur08 ] . • \mathbf{K}(\operatorname{Inj}\mathcal{A}) for a quasicoherent \mathcal{O}_{X} -algebra over a noetherian J-2 scheme X . With these hypotheses, we have the following general results on producing semiorthogonal decompositions. Theorem B ( Corollary 3.6 & Corollary 3.15 ) . Let \mathsf{T} satisfy Hypothesis (I) . If \langle\mathsf{A},\mathsf{B}\rangle is a semiorthogonal decomposition on \mathsf{T}^{c} , then \langle\operatorname{Coprod}(\mathsf{A})\cap\mathsf{T}^{-}_{c},\operatorname{Coprod}(\mathsf{B})\cap\mathsf{T}^{-}_{c}\rangle is a semiorthogonal decomposition on \mathsf{T}^{-}_{c} . If \mathsf{B} is further assumed to be admissible, then \langle\operatorname{Coprod}(\mathsf{A})\cap{}^{\mathrm{I}}\mathsf{T}^{b}_{c},\operatorname{Coprod}(\mathsf{B})\cap{}^{\mathrm{I}}\mathsf{T}^{b}_{c}\rangle is a semiorthogonal decomposition on {}^{\mathrm{I}}\mathsf{T}^{b}_{c} . Theorem C ( Corollary 3.7 & Corollary 3.16 ) . Let \mathsf{T} satisfy Hypothesis (II) . If \langle\mathsf{A},\mathsf{B}\rangle is a semiorthogonal decomposition on \mathsf{T}^{c} , then \langle\operatorname{Coprod}(\mathsf{A})\cap\mathsf{T}^{+}_{c},\operatorname{Coprod}(\mathsf{B})\cap\mathsf{T}^{+}_{c}\rangle is a semiorthogonal decomposition on \mathsf{T}^{+}_{c} . If \mathsf{B} is further assumed to be admissible, then \langle\operatorname{Coprod}(\mathsf{A})\cap{}^{\mathrm{II}}\mathsf{T}^{b}_{c},\operatorname{Coprod}(\mathsf{B})\cap{}^{\mathrm{II}}\mathsf{T}^{b}_{c}\rangle is a semiorthogonal decomposition on {}^{\mathrm{II}}\mathsf{T}^{b}_{c} . Then, note that to get Theorem A from Theorem B and Theorem C , we need to compute the subcategories defined in Definition A . It turns out that some of the work has already been done for us. For the examples related to Hypothesis (I) , the computation of \mathsf{T}^{-}_{c} and {}^{\mathrm{I}}\mathsf{T}^{b}_{c} follows from [ Nee24 , DLMR24 , HLLP25 , DLMRP25 ] . For the examples related to Hypothesis (II) , more work is required to compute \mathsf{T}^{+}_{c} and {}^{\mathrm{II}}\mathsf{T}^{b}_{c} . The main issue is that a priori there is not much control on the co-t-structure compactly generated by a compact object which is used to define these categories, see Definition A . Let us take the case of \mathsf{T}=\mathbf{K}(\operatorname{Inj}X) for a noetherian J-2 scheme X to illustrates our main strategy to solve this issue. \mathbf{K}(\operatorname{Inj}X) has a standard co-t-structure, given by