Remarks on diagonal dimension for algebraic stacks
Pat Lank, Fei Peng · 2026
Read Remarks on diagonal dimension for algebraic stacks by Pat Lank, Fei Peng (2026) 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-2605-13416.txt.
Introduction
Pat Lank and Fei Peng study diagonal dimension, a measure used to bound Rouquier dimension for algebraic stacks. It tracks how a space's diagonal morphism can be built from simpler pieces. They give an explicit bound for smooth morphisms with a regular target. They name classical generators of a fibre product, recovering Elagin's result, Lunts and Schnurer. Finally they show that a variety with mild singularities has diagonal dimension at most twice its Krull dimension. This holds in any characteristic.
Catalog record
| CREATOR | Pat Lank, Fei Peng |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44352 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.13416
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Remarks on diagonal dimension for algebraic stacks Pat Lank, Fei Peng Abstract . This note is concerned with the Rouquier dimension of the bounded derived category of coherent complexes on a Noetherian algebraic stack. Specifically, we study the diagonal dimension of a morphism, which can be used to produce upper bounds on Rouquier dimension. First, we obtain an explicit upper bound for smooth morphisms with a regular target. Second, we identify strong generators of a fiber product, recovering a result of Elagin–Lunts–Schnürer. Finally, we show that the diagonal dimension of a variety in arbitrary characteristic with mild singularities is at most twice its Krull dimension. 2020 Mathematics Subject Classification 14A30 (primary), 14A20, 13A35 1. Introduction 1.1. What is known Let \mathcal{X} be a Noetherian algebraic stack. Denote by D^{b}_{\operatorname{coh}}(\mathcal{X}) the bounded derived category of complexes with coherent cohomology. A basic problem is to understand the size and complexity of this triangulated category. We use the notion of generation introduced by Bondal–Van den Bergh [ 5 ] . This idea leads to numerical invariants that measure the complexity [ 39 , 2 ] . For G\in D^{b}_{\operatorname{coh}}(\mathcal{X}) , let \langle G\rangle_{n} denote the full subcategory consisting of objects obtained from G using finite coproducts, direct summands, shifts, and at most n-1 cones. The Rouquier dimension is the smallest n\geq 0 for which there exists G such that \langle G\rangle_{n+1}=D^{b}_{\operatorname{coh}}(\mathcal{X}) ; such an object is called a strong generator . An object G is a classical generator if \cup^{\infty}_{n=0}\langle G\rangle_{n}=D^{b}_{\operatorname{coh}}(\mathcal{X}) . Strong and classical generators provide a coarse description of D^{b}_{\operatorname{coh}}(\mathcal{X}) . Their existence is closely related to geometric properties such as the openness of the regular locus [ 10 , Proposition 4.6] . A central problem is to determine when such generators exist (ideally with an explicit description) and to obtain bounds on the Rouquier dimension. For schemes, these questions are well understood. In contrast, for algebraic stacks, they present additional challenges, and effective bounds on the Rouquier dimension remain subtle. Recent work has established the existence of generators in increasing generality (see e.g. [ 20 , 11 , 10 ] ), but explicit constructions are rare (see e.g. [ 26 ] ), and computable bounds are even more limited. 1.2. What we do Following [ 16 ] , we study a notion of diagonal dimension for morphisms of algebraic stacks. It provides a refinement to the Rouquier dimension and often tractable upper bounds for the Rouquier dimension. Specifically: Definition 1.1 . Let f\colon\mathcal{Y}\to\mathcal{X} be a separated morphism of Noetherian algebraic stacks. Denote by p_{i}\colon\mathcal{Y}\times_{\mathcal{X}}\mathcal{Y}\to\mathcal{Y} the i -th projection morphism and \Delta\colon\mathcal{Y}\to\mathcal{Y}\times_{\mathcal{X}}\mathcal{Y} the diagonal morphism. The diagonal dimension of f , denoted \dim_{\Delta}(f) , is the smallest n\geq 0 such that \mathbf{R}\Delta_{\ast}\mathcal{O}_{\mathcal{Y}}\in\langle\mathbf{L}p_{1}^{\ast}G_{1}\otimes^{\mathbf{L}}\mathbf{L}p_{2}^{\ast}G_{2}\rangle_{n+1} for some G_{1},G_{2}\in D^{b}_{\operatorname{coh}}(\mathcal{Y}) . In the case of schemes over a field, the diagonal dimension provides an upper bound for the Rouquier dimension. However, it is known that the two need not coincide even for smooth curves over a field; see [ 35 ] . In fact, for stacky curves, a similar observation has been made, see [ 4 ] . 1.2.1. Bounds Now we state our first result. Theorem 1.2 . Let f\colon\mathcal{Y}\to\operatorname{Spec}(R) be a concentrated separated smooth finitely presented quasi-DM morphism to a regular affine scheme of finite Krull dimension from an algebraic stack. If U\to\mathcal{Y} is a smooth surjective morphism from a scheme, then \dim_{\Delta}(f)\leq(\dim R+1)(2\max\{\dim U,\operatorname{cd}(\mathcal{Y})\}+1)-1<\infty. In particular, \dim D^{b}_{\operatorname{coh}}(\mathcal{Y}) is bounded above by the upper bound. This yields an explicit and computable upper bound for the Rouquier dimension of D^{b}_{\operatorname{coh}}(\mathcal{Y}) . Here, \operatorname{cd}(-) denotes the cohomological dimension of an algebraic stack (see Section 2.2 ). When R is a field, Theorem 1.2 recovers known bounds for smooth varieties [ 39 , Proposition 7.9] , smooth algebraic spaces [ 34 , Lemma 1.2.9] , and smooth Deligne–Mumford stacks [ 3 , Lemma 2.20] . In contrast, our result applies in greater generality. For example, it applies when \mathcal{Y} is a separated algebraic stack which is smooth and quasi-DM over a field of characteristic zero with affine stabilizers (see [ 21 , Theorem 2.1(1)] ). It also applies when \mathcal{Y} is a separated quasi-DM algebraic stack which is smooth over a DVR (possibly of mixed characteristic), provided the stabilizers are affine and nice in the sense of [ 21 ] . 1.2.2. Fiber products Our second result identifies generators for fiber products in terms of those of the factors. Proposition 1.3 . Let \mathcal{S} be a concentrated quasi-DM regular algebraic stack. Consider surjective morphisms of finite presentation between algebraic stacks {\lx@inpgf@ignorespaces{\mathcal{Y}^{\prime}_{1}}} {\lx@inpgf@ignorespaces{\mathcal{Y}_{1}}} {\lx@inpgf@ignorespaces{\mathcal{Y}_{2}}} {\lx@inpgf@ignorespaces{\mathcal{Y}^{\prime}_{2}}} {\lx@inpgf@ignorespaces{\mathcal{S}}} \scriptstyle{\lx@inpgf@ignorespaces h_{1}} \scriptstyle{\lx@inpgf@ignorespaces f_{1}} \scriptstyle{\lx@inpgf@ignorespaces f_{2}} \scriptstyle{\lx@inpgf@ignorespaces h_{2}} where the f_{i}\circ h_{i} are concentrated smooth quasi-DM and the h_{i} are proper. If G_{i} is a classical generator for D^{b}_{\operatorname{coh}}(\mathcal{Y}_{i}) for each i , then \mathbf{L}g_{1}^{\ast}G_{1}\otimes^{\mathbf{L}}\mathbf{L}g_{2}^{\ast}G_{2} is a classical generator for D^{b}_{\operatorname{coh}}(\mathcal{Y}_{1}\times_{\mathcal{S}}\mathcal{Y}_{2}) where g_{i}\colon\mathcal{Y}_{1}\times_{\mathcal{S}}\mathcal{Y}_{2}\to\mathcal{Y}_{i} are the natural projections. Notably, Proposition 1.3 applies broadly and does not impose separatedness or finite Krull dimension on the base. For instance, it applies to separated integral schemes of finite type over a perfect field via [ 12 , Theorem 4.1] , and to reduced Deligne–Mumford stacks of finite presentation over a field of characteristic zero via [ 48 , §5.1] and [ 14 , §3.6] . Moreover, our approach recovers [ 15 , Theorem 4.18] for varieties over a perfect field and provides a proof without dg methods. Unlike loc. cit., we work directly at the level of algebraic stacks and extend to Deligne–Mumford stacks admitting suitable regular covers. It would be interesting for future work to study analogs of our results for coherent noncommutative algebras over an algebraic stack (see e.g. [ 15 , Theorem 4.15] ). 1.2.3. Comparisons Our final result compares the diagonal dimension along morphisms whose unit for derived pullback/pushforward is an isomorphism. Theorem 1.4 . Let \mathcal{S} be a Noetherian algebraic stack. Consider concentrated finitely presented flat morphisms f_{i}\colon\mathcal{Y}_{i}\to\mathcal{S} . If there exist an \mathcal{S} -morphism f\colon\mathcal{Y}_{2}\to\mathcal{Y}_{1} such that \mathcal{O}_{\mathcal{Y}_{1}}\xrightarrow{ntrl.}\mathbf{R}f_{\ast}\mathcal{O}_{Y_{2}} is an isomorphism and \mathbf{R}f_{\ast} preserves coherent cohomology, then \dim_{\Delta}(f_{1})\leq\dim_{\Delta}(f_{2}) . There are important situations in which Theorem 1.4 applies. Corollary 1.5 . (1) (characteristic zero) Let f\colon X\to\operatorname{Spec}(k) be a morphism from a separated integral scheme to a field of characteristic zero. If X has rational singularities, then \dim_{\Delta}(f)\leq 2\dim X . (2) (arbitrary characteristic) Let f\colon X\to\operatorname{Spec}(k) be a separated irreducible scheme of finite presentation over a perfect field k . If X has linearly reductive singularities, then \dim_{\Delta}(f)\leq 2\dim X . These results appear to provide a new general bound on the diagonal dimension for varieties with mild singularities. In the case of characteristic zero, [ 23 ] and [ 17 ] have shown that the diagonal dimension for many toric varieties is precisely the Krull dimension. However, the example above holds for cases which need not be toric and provides an upper bound (e.g. quotient singularities, see Example 3.9 ). Moreover, in positive characteristics, our results are applicable to varieties with tame quotient singularities. See 3.8 . Acknowledgments . The authors thank Anirban Bhaduri and Timothy De Deyn for discussions on an earlier version of our work. Also, we greatly appreciate discussions with Alexey Elagin which motivated thinking about Theorem 1.4 . Fei Peng would like