By OpenAISourceGitHub

Manuscripts from openai/math

Fork
Clone
git clone https://openversions.com/git/index/openai-math.git
Cite
Version link
https://openversions.com/index/openai-math/commit/dfd8c46870b48511e697cbb43a3c8c91d375fa2f
Version identifier (commit hash)
dfd8c46870b48511e697cbb43a3c8c91d375fa2f

Fork to create your own copy with credit to the source. Clone to get a Git copy on your computer.

All papers

Claims

Findings

The spacetime Penrose inequality and enclosing area

For smooth one-ended initial-data exteriors in every spatial dimension n≥3n \geq 3 satisfying the dominant energy condition, weak future trapping on the compact boundary SS, integrable constraint densities, finite ADM limits, g−δ=O6(r−q)g - \delta = O_6(r^{-q}) and K=O5(r−1−q)K = O_5(r^{-1-q}) with n−22<q<n−2\frac{n-2}{2} < q < n-2, and positive enclosing area Amin⁡,g(S)>0A_{\min,g}(S) > 0, the ADM energy-momentum is automatically future timelike (E>∣P∣δE > |P|_\delta) and satisfies the sharp spacetime Penrose inequality m≥12(Amin⁡,g(S)ωn−1)n−2n−1m \geq \frac{1}{2} \left( \frac{A_{\min,g}(S)}{\omega_{n-1}} \right)^{\frac{n-2}{n-1}}.

In spatial dimensions n=3n = 3 and n=4n = 4, for smooth one-ended initial-data exteriors satisfying the dominant energy condition, weak future trapping, integrable constraint densities, finite ADM limits, and weaker asymptotic decay g−δ=O2(r−q)g - \delta = O_2(r^{-q}) and K=O1(r−1−q)K = O_1(r^{-1-q}) with q>n−22q > \frac{n-2}{2}, positive enclosing area Amin⁡,g(S)>0A_{\min,g}(S) > 0, future timelikeness E>∣P∣δE > |P|_\delta, and the sharp spacetime Penrose inequality (1.5) all hold.

For a connected orientable smooth complete three-dimensional initial data set without boundary having finitely many asymptotically flat ends with g−δ=O2(r−q)g - \delta = O_2(r^{-q}), K=O1(r−1−q)K = O_1(r^{-1-q}) (q>1/2q > 1/2), finite ADM limits, and integrable physical constraint densities satisfying μ≥∣J∣g\mu \geq |J|_g, every chosen end and every admissible weakly future outer trapped boundary S=∂D+S = \partial D^+ satisfies E>∣P∣δE > |P|_\delta and E2−∣P∣δ2≥Amin⁡(S)16π\sqrt{E^2 - |P|_\delta^2} \geq \sqrt{\frac{A_{\min}(S)}{16\pi}}.

For a connected orientable smooth four-dimensional initial data set with nonempty compact smooth weakly future trapped boundary SS, finitely many asymptotically flat ends with g−δ=O2(r−q)g - \delta = O_2(r^{-q}) and K=O1(r−1−q)K = O_1(r^{-1-q}) (q>1q > 1), finite ADM limits, and integrable constraint densities satisfying μ≥∣J∣g\mu \geq |J|_g, each end ee satisfies Ee>∣Pe∣δE_e > |P_e|_\delta and Ee2−∣Pe∣δ2≥12(Ae(S)ω3)2/3\sqrt{E_e^2 - |P_e|_\delta^2} \geq \frac{1}{2} \left( \frac{A_e(S)}{\omega_3} \right)^{2/3}. Moreover, in the one-ended maximal case (trgXKX=0\mathrm{tr}_{g_X} K_X = 0), the ADM momentum limit exists automatically and EX≥12(agX(SX)2π2)2/3E_X \geq \frac{1}{2} \left( \frac{a_{g_X}(S_X)}{2\pi^2} \right)^{2/3}.

Every prepared exterior (Ωn,g0,K0)(\Omega^n, g_0, K_0) in Definition 2.1 satisfies E0≥12(A0ω)(n−2)/(n−1)E_0 \geq \frac{1}{2} \left( \frac{A_0}{\omega} \right)^{(n-2)/(n-1)}.

For a fixed prepared exterior and 0<ϵ<10 < \epsilon < 1, for all sufficiently large integers NN there exists a connected enclosing-set exterior XNX_N, complete with its compact outer-minimizing and locally perimeter-minimizing frontier ΣN\Sigma_N included, in a smooth Riemannian ambient metric g~N\tilde{g}_N with Rg~N≥0R_{\tilde{g}_N} \geq 0, Rg~N∈L1(XN)R_{\tilde{g}_N} \in L^1(X_N), g~N−δ=O2(r2−n)\tilde{g}_N - \delta = O_2(r^{2-n}), Rg~N=O(r−n−β′)R_{\tilde{g}_N} = O(r^{-n-\beta'}), E(g~N)≤E0+PN(e−Nϵ+e−Nϵ/2)E(\tilde{g}_N) \leq E_0 + P_N(e^{-N\epsilon} + e^{-N\epsilon/2}), and Hg~Nn−1(ΣN)≥e−(n−1)ϵA0\mathcal{H}^{n-1}_{\tilde{g}_N}(\Sigma_N) \geq e^{-(n-1)\epsilon} A_0.

Any smooth one-ended exterior (Ω,g,K)(\Omega, g, K) with O6/O5O_6/O_5 decay, (n−2)/2<q<n−2(n-2)/2 < q < n-2, integrable constraint densities, DEC, E>∣P∣E > |P|, θ+(S)≤0\theta_+(S) \leq 0, and a=Amin⁡g(S)>0a = A_{\min}^g(S) > 0 admits prepared data (gj,Kj)(g_j, K_j) on Ω\Omega with Pj=0P_j = 0, Ej→m=(E2−∣P∣2)1/2E_j \to m = (E^2 - |P|^2)^{1/2}, Φj∗g≤(1+ϵj)gj\Phi_j^* g \leq (1 + \epsilon_j) g_j, and Amin⁡gj(S)≥(1+ϵj)−k/2aA_{\min}^{g_j}(S) \geq (1 + \epsilon_j)^{-k/2} a.

For smooth prepared data satisfying (2.1) and all sufficiently large N>n−1N > n-1, there exist smooth functions f,Zf, Z on the filled manifold MN\mathcal{M}_N solving the coupled system (4.18) with −ϵ<t≤Z≤P(N)-\epsilon < t \leq Z \leq P(N), ∣f∣+∣df∣g≤exp⁡(P(N))|f| + |df|_g \leq \exp(P(N)), uniform height bounds ∣h∣≤C|h| \leq C and ∣h(x)∣≤Cr(x)−b1|h(x)| \leq C r(x)^{-b_1}, and exponential decay of ff on the end.

Methods

Asymptotic end preparation by matching ADM charges with a boosted Schwarzschild-Tangherlini slice, gluing via compact Bogovskii-type right inverses with sublinear first moments, bending the vacuum slice to rest, and repairing the constraint deficit with a radial conformal factor.

Source excerpts

“We first match the original energy and momentum with a boosted vacuum end. A compact annular correction joins the two ends with a constraint error whose total scale tends to zero. Within the vacuum region we then bend the slice to a static one, giving zero momentum. A positive conformal factor repairs the small annular error without decreasing the metric.” · Show in paper

Conformal strictification of the dominant energy condition and boundary future expansion using a regularized spacetime drift-Poisson equation with negative inner Neumann boundary data.

Source excerpts

“We solve a scalar equation with positive source and negative inner normal derivative. The source gives a quantitative strict energy margin; the normal derivative makes the future expansion strictly negative.” · Show in paper

Smooth boundary filling and solution of a coupled warped-graph and conformal-factor system with a penalized divergence source, using Leray-Schauder degree on moving domains, Michael-Simon graph Sobolev iteration, a gradient-aligned rank-one regularity lemma, and a natural-growth scalar estimate.

Source excerpts

“The gradient-aligned trace equation and the bounded natural-growth divergence equation are treated separately, as reusable scalar lemmas. Finally we solve the trace equation for every trial Z, use it to define a compact map, and obtain a fixed point by two homotopies. Exhausting the outer spheres gives the complete solution and the end estimates needed for its ADM flux.” · Show in paper

Height-separation conformal plateau and obstacle-constrained perimeter minimization with two-sided volume density estimates to detach the minimizing frontier into the nonnegative scalar curvature region.

Source excerpts

“We use this separation to place a large constant conformal factor around the filling. The cutoff vanishes on a distant tail, so it leaves the ADM energy unchanged. Its transition derivatives are small by weighted coercivity. We then minimize perimeter subject to a distance obstacle inside the plateau. Contact with that obstacle would give positive volume densities on both sides of the frontier; the plateau would turn the resulting perimeter into an area larger than a distant competitor.” · Show in paper