ScalingStacks

3.1. Exponential decay lemma [05CM]

Original official author HTML, exact retained edition. Historical TeX conversion verdicts remain unchanged. Cited-edition alignment and mathematical self-containment are not assessed.

Complete original source context · Original author HTML

3.1. Exponential decay lemma

We would like to analyze the behavior of Gibbons-Hawking solutions when the tori (both in the fibers and in the base) are shrinking. First, let us introduce a non-linear differential equation of the Monge-Ampère type.

Definition.

We refer to a pair of real positive definite matrix functions Vi​j,Wp​qV^{ij},W^{pq} as a solution to the split Monge-Ampère equation in an open subset R⊂ℝn×ℝlR\subset\mathbb{R}^{n}\times\mathbb{R}^{l} if detVi​j=detWp​q\det V^{ij}=\det W^{pq} and they are locally given by a smooth potential function KK:

(17) Vi​j=∂2K∂si​∂sj,Wp​q=−∂2K∂tp​∂tq,1≤i,j≤n,k+1≤p,q≤n+l.V^{ij}=\frac{\partial^{2}K}{\partial s_{i}\partial s_{j}},\quad W^{pq}=-\frac{\partial^{2}K}{\partial t_{p}\partial t_{q}},\quad 1\leq i,j\leq n,\quad k+1\leq p,q\leq n+l.

To describe the asymptotics at the discriminant we would like to treat the simplex σ⊂N∗\sigma\subset N^{*} on the same footing as τ\tau. Namely, we let Σ⊂Nℝ∗\Sigma\subset N^{*}_{\mathbb{R}} be the cone over σ\sigma, and let Σ∨\Sigma^{\vee} be its dual cone in NℝN_{\mathbb{R}}. The polyhedral complex Π⁡(σ)\Pi(\sigma) provides a polyhedral decomposition of Nℝ/τN_{\mathbb{R}}/\tau into cells QiσQ_{i}^{\sigma}. Denote by γσ\gamma_{\sigma} the Nτ∗N^{*}_{\tau}-valued 1-current defined in the same way as γτ\gamma_{\tau}.

Definition.

Given a domain RR in Nℝ∗/σ×Nℝ/τ≅ℝn×ℝlN^{*}_{\mathbb{R}}/\sigma\times N_{\mathbb{R}}/\tau\cong\mathbb{R}^{n}\times\mathbb{R}^{l} a (σ,τ)(\sigma,\tau)-type singular solution to the split Monge-Ampère equation in RR is a pair of matrix functions Vi​j,Wp​qV^{ij},W^{pq} which are local Monge-Ampère solutions in R∖(Π⁡(τ)×Π⁡(σ))R\setminus(\Pi(\tau)\times\Pi(\sigma)) with asymptotics at the discriminant locus governed by the distributional equation

(18) 12​π​(∂2Wp​q∂si​∂sj+∂2Vi​j∂tp​∂tq)​d​si∧d​tp=γτj​(s)​γσq​(t).\frac{1}{2\pi}\left(\frac{\partial^{2}W^{pq}}{\partial s_{i}\partial s_{j}}+\frac{\partial^{2}V^{ij}}{\partial t_{p}\partial t_{q}}\right)ds_{i}\wedge dt_{p}=\gamma^{j}_{\tau}(s)\gamma^{q}_{\sigma}(t).
Conjecture 3.1 (Exponential decay lemma).

Given a convex domain RR in ℝk×ℝl\mathbb{R}^{k}\times\mathbb{R}^{l} and a (σ,τ)(\sigma,\tau)-type solution V,WV,W of the split Monge-Ampère equation in RR there is a real one-parameter family of (σ,τ)(\sigma,\tau)-solutions Vλ,WλV_{\lambda},W_{\lambda} to the Gibbons-Hawking ansatz in λ​R×(S1)l\lambda R\times(S^{1})^{l} such that

  • •

    The diameter of the circles both in the fiber TnT^{n} and in the torus part (S1)l(S^{1})^{l} of the base away from the discriminant is roughly given by λ−1\lambda^{-1}.

  • •

    The zero Fourier modes of the GH solutions Vλ0​(u,x),Wλ0​(u,x)V_{\lambda}^{0}(u,x),W_{\lambda}^{0}(u,x) as functions of the rescaled variables s,ts,t, where u=λ​s,x=λ​tu=\lambda s,x=\lambda t, will converge (in some properly weighted norm on the function space) to V⁡(s,t),W⁡(s,t)V(s,t),W(s,t) as λ→∞\lambda\to\infty.

  • •

    The higher Fourier modes decay exponentially away from the discriminant Π⁡(τ)×Π⁡(σ)\Pi(\tau)\times\Pi(\sigma) in λ​R\lambda R, uniformly in λ\lambda. That is, if β⁡(u,x)\beta(u,x) denotes the Euclidean distance from the point (u,x)(u,x) to the discriminant, then

    |Vλm​(u,x)|≤C1​e−β⁡(u,x)​|m|,|Wλm​(u,x)|≤C2​e−β⁡(u,x)​|m|,|V_{\lambda}^{m}(u,x)|\leq C_{1}e^{-\beta(u,x)|m|},\quad|W_{\lambda}^{m}(u,x)|\leq C_{2}e^{-\beta(u,x)|m|},

    for some constants C1,C2C_{1},C_{2}, and large enough λ\lambda and β\beta.

We would like to give some easy examples and a rough argument based on those why we believe this conjecture is true. Note, however, that once justified, it will have an important consequence for the metric collapse program for the toric hypersurfaces and complete intersections:

Corollary 3.2.

The metric space (Zσ,τ​(λ−1​R),λ−2​gλ)(Z_{\sigma,\tau}(\lambda^{-1}R),\lambda^{-2}g_{\lambda}), where gλg_{\lambda} is the Riemannian (orbifold) metric from the Gibbons-Hawking ansatz, converges in the Gromov-Hausdorff sense to (R,g∞i​j)(R,g^{ij}_{\infty}), with the limiting metric g∞i​j=Vi​j​d​si​d​sj+Wp​q​d​tp​d​tqg^{ij}_{\infty}=V^{ij}ds_{i}ds_{j}+W^{pq}dt_{p}dt_{q}.

Original mathematics by the credited authors. Source-backed reader collection; mathematical self-containment is not assessed.