The DT–PT virtual structure sheaf localization conjecture

From papers

Let MM be the relevant moduli space with obstruction theory cmathbbEcmathbb{E}^\bullet, equipped with an action of the torus A\mathsf{A} and a further C×\mathbb{C}^\times-action, and let MAM^{\mathsf{A}} and MA×C×M^{\mathsf{A}\times\mathbb{C}^\times} denote the corresponding fixed loci. Write ι ⁣:MAM\iota\colon M^{\mathsf{A}}\hookrightarrow M for the inclusion, let O^Mvir\widehat{\mathcal{O}}^\mathrm{vir}_M be the Oh–Thomas virtual structure sheaf, and let O^vir,BF\widehat{\mathcal{O}}^{\mathrm{vir},\mathrm{BF}} and the virtual Euler classes in the formula be the Behrend–Fantechi-style localized classes associated with the induced obstruction theories. A canonical halving is a class E1/2\mathbb{E}_{1/2} satisfying

E1/2+E1/2=EMA×C×.\mathbb{E}_{1/2}+\mathbb{E}_{1/2}^\vee=\mathbb{E}^\bullet\big|_{M^{\mathsf{A}\times\mathbb{C}^\times}}.

The DT–PT virtual structure sheaf localization conjecture. There exists a canonical halving E1/2\mathbb{E}_{1/2}, given explicitly in the source's Definition~, such that

(ι)1O^Mvir=[O^MA×C×vir,BFe^C×(Nvir,BF)e^C×(TMAvir,BF)e^C×(E1/2A-fix)1e^A×C×(E1/2A-mov)]A.\left(\iota_*\right)^{-1}\widehat{\mathcal{O}}^\mathrm{vir}_M=\left[\frac{\widehat{\mathcal{O}}^{\mathrm{vir},\mathrm{BF}}_{M^{\mathsf{A}\times\mathbb{C}^\times}}}{\widehat{\mathsf{e}}_{\mathbb{C}^\times}(\mathcal{N}^{\mathrm{vir},\mathrm{BF}})}\frac{\widehat{\mathsf{e}}_{\mathbb{C}^\times}(\mathcal{T}^{\mathrm{vir},\mathrm{BF}}_{M^\mathsf{A}})}{\widehat{\mathsf{e}}_{\mathbb{C}^\times}(\mathbb{E}_{1/2}^{\mathsf{A}\text{-fix}})}\frac{1}{\widehat{\mathsf{e}}_{\mathsf{A}\times\mathbb{C}^\times}(\mathbb{E}_{1/2}^{\mathsf{A}\text{-mov}})}\right]\Bigg|_{\mathsf{A}}.

The second fraction is a sign ±1\pm1 on each connected component FMA×C×F\subset M^{\mathsf{A}\times\mathbb{C}^\times}, and the restriction appearing in the conjectural formula is required to be well-defined. The claim proposes an explicit relation between the Oh–Thomas virtual structure sheaf and the localized Behrend–Fantechi cycle on fixed loci; the corresponding restriction for the Behrend–Fantechi localization is well-defined by properness of MAM^{\mathsf{A}}, while well-definedness of the restriction in the displayed conjectural identity remains part of the conjecture.

Progress summary

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Sources & referencesView supporting material

Primary source

Henry Liu, “The 4-fold Pandharipande–Thomas vertex”, arXiv:2306.12923 (2023).

Solutions 0

No solutions have been posted yet.