DT-GV correspondence on the Mukai-Umemura variety
Kiryong Chung, Joonyeong Won · 2026
Read DT-GV correspondence on the Mukai-Umemura variety by Kiryong Chung, Joonyeong Won (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-2603-05823.txt.
Introduction
Chung and Won compute Donaldson-Thomas fixed values, together with child versions, for a local Calabi-Yau 4-fold built over the Mukai-Umemura variety. Their math proceeds through several local-focus rules. Under the guess that genus-one Gopakumar-Vafa type fixed values are zero, the results verify forecasts made by Cao, Maulik, and Toda. The work thus confirms the hoped-for match in this setting through direct fixed-value math.
Catalog record
| CREATOR | Kiryong Chung, Joonyeong Won |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43125 |
| LICENSE | http://creativecommons.org/publicdomain/zero/1.0/ |
| RIGHTS | CC0 |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.05823
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
DT-GV correspondence on the Mukai-Umemura variety Kiryong Chung, Joonyeong Won Abstract . We compute Donaldson–Thomas (DT) invariants and their descendant invariants for the local Calabi–Yau 4 4 –fold over the Mukai–Umemura variety via several localization formulas. Assuming that the genus-one Gopakumar–Vafa (GV) type invariants vanish, our computations verify the predictions of Cao, Maulik, and Toda. 2020 Mathematics Subject Classification 14N35, 14C15, 14C35, 14J45 1. Introduction 1.1. Motivation and result Let Y be a Calabi–Yau 4 -fold and \beta\in H_{2}(Y,\mathbb{Z}) be a curve class. Let \mathbf{M}_{\beta}(Y)(=\mathbf{M}) denote the moduli space of one-dimensional stable sheaves F on Y satisfying [F]=\beta and \chi(F)=1 . Given a class \gamma\in\mathrm{H}^{4-2i}(Y,\mathbb{Z}) , the insertion is defined by \tau_{i}(\gamma)=\pi_{\mathbf{M}*}\!\left(\pi_{Y}^{*}\gamma\cup\mathrm{ch}_{i+3}(\mathcal{F})\right), where \mathcal{F}\in\mathrm{Coh}(\mathbf{M}_{\beta}(Y)\times Y) is the universal sheaf with the trivial determinant (i.e., \det(R\pi_{\mathbf{M}*}\mathcal{F})=\mathcal{O}_{\mathbf{M}} ). The primary (for i=0 ) or descendent (for i\geq 1 ) invariants are defined by the integration of the insertion over the virtual class [\mathbf{M}_{\beta}(Y)]^{\mathrm{vir}} : (1) \langle\tau_{i}(\gamma)\rangle_{\beta}:=\int_{[\mathbf{M}_{\beta}(Y)]^{\mathrm{vir}}}\tau_{i}(\gamma). The integral \langle\tau_{0}(\gamma)\rangle_{\beta} is called \mathrm{DT} -invariants (or primary invariant ). When i\geq 1 , we call \tau_{i}(\gamma) descendant insertions , and its integrations over the virtual class are called descendant invariants . The DT theory of Calabi–Yau 4-folds has recently attracted significant attention in enumerative geometry, motivated by its rich geometric structures and deep connections to other curve counting theories. In particular, Cao, Maulik, and Toda proposed a series of conjectures predicting a precise correspondence between DT invariants of Calabi–Yau 4–folds and the Gopakumar–Vafa (GV) type invariants introduced by Klemm and Pandharipande ( [ CMT18 , CT21 , KP08 ] ). More precisely, it states that Conjecture 1.1 ( [ CMT18 , Conjecture 0.2] and [ CT21 , Conjecture 0.2] ) . Let n_{0,\beta}(-) (resp. n_{1,\beta} ) denote genus 0 (resp. 1 ) GV-type invariants and m_{\beta_{1},\beta_{2}} be the meeting invariants defined in [ KP08 ] . Under the suitable orientation on the moduli space \mathbf{M}_{\beta}(Y) , we have (a) n_{0,\beta}(\gamma_{0})=\langle\tau_{0}(\gamma_{0})\rangle_{\beta} for \gamma_{0}\in\mathrm{H}^{4}(Y,\mathbb{Z}) . (b) Furthermore, we have an linear identity: (2) \langle\tau_{1}(\gamma_{1})\rangle_{\beta}=\frac{n_{0,\beta}(\gamma_{1}^{2})}{2(\gamma_{1}\cdot\beta)}-\sum_{\beta_{1}+\beta_{2}=\beta}\frac{(\gamma_{1}\cdot\beta_{1})(\gamma_{1}\cdot\beta_{2})}{4(\gamma_{1}\cdot\beta)}\,m_{\beta_{1},\beta_{2}}-\sum_{k\geq 1,\,k|\beta}\frac{\gamma_{1}\cdot\beta}{k}\,n_{1,\beta/k} with respect to \gamma_{1}\in\mathrm{H}^{2}(Y,\mathbb{Z}) . In this paper, we study the case where Y=\mathrm{Tot}(K_{X}) is the total space of the canonical bundle of the Mukai-Umemura variety X (i.e. the unique smooth prime Fano threefold of genus 12 admitting a nontrivial \mathrm{SL}_{2}(\mathbb{C}) -action). For the defining relations of the meeting invariants m_{\beta_{1},\beta_{2}} for our local Fano 4 -fold Y , see the first paragraph of Section 3.4 . It was previously known that Conjecture 1.1 holds for curve classes \beta=d[\mathrm{line}] , 1\leq d\leq 3 and Y=\mathrm{Tot}(K_{X}) ( [ CLW24 , Theorem 4.2] ). For d\leq 3 , the explicit descriptions of the universal sheaf on the moduli space obtained in [ AF06 , Lemma 3.1] , [ Fae14 , Lemma 4.1] , and [ KS04 , Theorem 2.4] are sufficient to prove the conjecture. However, the case d=4 requires a more delicate analysis of the structure of stable sheaves and their contributions to the virtual cycle. In this paper, we extends our results degree up to d=4 . Theorem 1.2 ( = Theorem 3.16 ) . For Y=\mathrm{Tot}(K_{X}) with the Mukai-Umemura variety X , the equality ( 2 ) in Conjecture 1.1 holds for \beta=4[\mathrm{line}] whenever we assume that n_{1,d}=0 for 1\leq d\leq 4 . The proof of Theorem 1.2 is based on a detailed analysis of the torus-fixed stable sheaves in the moduli space (Proposition 3.4 and Proposition 3.9 ). We then apply various (virtual) localization formulas studied in [ AB84 , BV82 , GP99 , Tho92 , Tho93 ] . By organizing these contributions via localization, we explicitly compute the primary and descendent invariants (Proposition 3.13 ). This establishes the equality ( 2 ) in Conjecture 1.1 (Section 3.4 ). In the computations, one of the key features is that for any fixed curve other than the multiplicity 4 -lines, there always exists a weight-zero component in the obstruction space (Lemma 3.2 , Proposition 3.9 and Proposition 3.11 ), which prevents it from contributing to the localization formula. Moreover, the same phenomenon occurs for d\leq 3 and thus we can easily recover the our previous results ( [ CLW24 , Proposition 3.17] ). 2. Preliminary In this section, we present several localization formulas and recall results from [ CKK25 ] that will be used later. 2.1. Localization formulas for varieties with isolated fixed points Let T ( \cong\mathbb{C}^{*} ) be an algebraic torus and M be a T -space whose fixed locus M^{T} is finite. We work in equivariant cohomology H_{T}^{*}(-) and homology \mathrm{H}_{*}^{T}(-) . Let us denote e^{T}(-) by the equivariant Euler class. Theorem 2.1 (Virtual localization [ GP99 ] ) . Assume that M carries a T -equivariant perfect obstruction theory, with virtual fundamental class [M]^{\mathrm{vir}}\in H^{T}_{*}(M) . Then for any u\in H_{T}^{*}(M) , (3) \int_{[M]^{\mathrm{vir}}}u=\sum_{x\in M^{T}}\frac{u|_{x}}{e^{T}\!\bigl(T_{x}^{\mathrm{vir}}M\bigr)}. Remark 2.2 (Smooth case [ AB84 , BV82 ] ) . If M is smooth, for any u\in H_{T}^{*}(M) , we have (4) \int_{[M]}u=\sum_{x\in M^{T}}\frac{u|_{x}}{e^{T}(T_{x}M)}. 2.2. Some facts in equivariant K-theory In [ Tho92 , Tho93 ] , the author proves an equivariant self-intersection formula and Lefschetz-type localization formula in algebraic K-theory for torus actions. For later use, we collect some results. Let us denote \lambda_{-1}(E|_{p})=\sum_{i=0}^{\mathrm{rank}\,E}(-1)^{i}\,[\wedge^{i}E|_{p}] by the K -theoretic \lambda –class of a T -equivariant locally free sheaf E on X at a fixed point p\in X . Denote by R(T) the representation ring of T , which is canonically isomorphic to R(\mathbb{C}^{*})\cong\mathbb{Z}[t^{\pm 1}] , where t denotes the standard one-dimensional character of \mathbb{C}^{*} . We define the \mathbb{C}^{*} -equivariant Euler characteristic (or pairing ) by \chi(E,F):=\sum_{i}(-1)^{i}[\mathrm{Ext}^{i}(E,F)]\in R(\mathbb{C}^{*}) for E and F are T -equivariant coherent sheaves on X . Theorem 2.3 (Theorem 3.5 [ Tho92 ] ) . Let X be a smooth projective variety with a torus action T and let F,G\in\text{Coh}^{T}(X) . Assume that the fixed locus X^{T} consists of isolated points. Then for i_{p}:\{p\}\hookrightarrow X , we have \chi(F,G)=\sum_{p\in X^{T}}\frac{i_{p}^{*}([F])^{\vee}\cdot i_{p}^{*}([G])}{\lambda_{-1}(T_{p}^{*}X)}. We recall the (equivariant) self-intersection formula in the K -theory. Theorem 2.4 ( [ Tho93 , Theorem 3.1] ) . Let i\colon C\hookrightarrow X be an embedding map such that C is a locally complete intersection. Let p\in C^{T} be a T -fixed point. Then for any F\in\text{Coh}^{T}(C) , i_{p}^{*}\bigl([i_{*}F]\bigr)=[F]_{p}\cdot\lambda_{-1}\!\bigl(N^{*}_{C/X,p}\bigr)\in K_{T}(\mathrm{pt}). Here [F]_{p}=[F\otimes^{\mathbf{L}}_{\mathcal{O}_{C}}\mathbb{C}_{p}]\in K_{T}(\mathrm{pt}) denotes the derived fiber of F at the point p . Remark 2.5 . In Theorem 2.4 , if F is locally free on C around the closed point p , then [F]_{p} coincides with the class of the T -representation on the usual fiber F|_{p} . Corollary 2.6 . Suppose that X is smooth and that the fixed point locus X^{T} is isolated. Then for any \mathbb{C}^{*} -equivariant locally free sheaf E on X , we have \chi(E)=\sum_{p\in X^{T}}\frac{[E]_{p}}{\lambda_{-1}(T_{p}^{*}X)}. Proof. Let F=\mathcal{O}_{X} and G=E . Then by Theorem 2.3 and Theorem 2.4 , we get the result. ∎ Corollary 2.7 . Let X be a smooth projective variety with a T -action, and let i\colon C\hookrightarrow X be a smooth T -invariant curve. Then (5) \chi(i_{*}\mathcal{O}_{C},i_{*}\mathcal{O}_{C})=\sum_{i=0}^{\dim X-1}(-1)^{i}\chi(C,\,\wedge^{i}N_{C/X}), where \wedge^{0}N_{C/X}=\mathcal{O}_{C} . Proof. Applying F=G=i_{*}\mathcal{O}_{C} in Theorem 2.3 , a direct calculation provides the result by Theorem 2.4 and Corollary 2.6 . ∎ 2.3. Geometric data of the variety X Let us recall the definition of the Mukai-Umemura variety X . Let