ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let UU be a connected separated smooth scheme of dimension dd of finite type over a finite field FF, let XX be a proper normal scheme over FF containing UU as a dense open subscheme, and let FrF∈Gal⁡(F‾/F)Fr_F\in\operatorname{Gal}(\overline F/F) be geometric Frobenius. Let ρX:CH0(X)→π1(X)ab\rho_X:CH_0(X)\to\pi_1(X)^{\mathrm{ab}} be the reciprocity map sending a closed point class [x][x] to geometric Frobenius FrxFr_x. Let ℓ\ell be invertible in FF, let F\mathcal F be a smooth F‾ℓ\overline{\mathbb F}_\ell- or Q‾ℓ\overline{\mathbb Q}_\ell-sheaf on UU, and assume the integrality conjecture holds and Sw⁡X(F)∈CH0(X∖U)\operatorname{Sw}_X(\mathcal F)\in CH_0(X\setminus U) is defined. For a smooth sheaf G\mathcal G on XX, let det⁡G:π1(X)ab→F‾ℓ×\det\mathcal G:\pi_1(X)^{\mathrm{ab}}\to\overline{\mathbb F}_\ell^\times or Q‾ℓ×\overline{\mathbb Q}_\ell^\times be the character corresponding to its determinant. Define

det⁡(−FrF:Hc∗(UF‾,F))=∏q=02ddet⁡(−FrF:Hcq(UF‾,F))(−1)q.\det(-Fr_F:H_c^*(U_{\overline F},\mathcal F))=\prod_{q=0}^{2d}\det(-Fr_F:H_c^q(U_{\overline F},\mathcal F))^{(-1)^q}.

Swan trace formula conjecture. Then

det⁡(−FrF:Hc∗(UF‾,F⊗G))=det⁡(−FrF:Hc∗(UF‾,F))rank⁡G⋅det⁡G(ρX(Sw⁡X(F))).\det(-Fr_F:H_c^*(U_{\overline F},\mathcal F\otimes\mathcal G)) =\det(-Fr_F:H_c^*(U_{\overline F},\mathcal F))^{\operatorname{rank}\mathcal G}\cdot\det\mathcal G\bigl(\rho_X(\operatorname{Sw}_X(\mathcal F))\bigr).

This is presented as an expected refinement of the theorem cited in the source; the supplied text gives no resolution.

Record

Comments

Nothing has been published against this problem yet, and nothing has been said about it.

Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.

Sign in

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.