Width deadline across downward jumps
ProvedOpenGA.le_width_deadline_across_downward_jumpsdifferential-inequalitiesgeometric-analysisricci-flow
Let , , and with . Let be continuous on , with and . Suppose that, for ,
and that at each internal joining time. Then
Here is the lower right Dini derivative, represented by frequent strict upper bounds on forward slopes. The limsup upper bound in the geometric application implies this hypothesis. Nonnegativity is needed only at the final endpoint.
This is the explicit analytic extinction estimate, extended to a finite sequence of downward jumps. It does not assert that a geometric flow or its width has been constructed.
Preamble
import Mathlib.Analysis.Calculus.MeanValue import Mathlib.Analysis.SpecialFunctions.Pow.Deriv set_option autoImplicit false open Set Topology open scoped Topology
Formal statement
theorem OpenGA.le_width_deadline_across_downward_jumps
{C W₀ : ℝ} (hC : 0 < C)
(n : ℕ) (hn : 0 < n) (times : ℕ → ℝ) (widths : ℕ → ℝ → ℝ)
(hzero : times 0 = 0)
(htimes : ∀ i < n, times i < times (i + 1))
(hcont : ∀ i < n, ContinuousOn (widths i) (Icc (times i) (times (i + 1))))
(hinitial : widths 0 (times 0) ≤ W₀)
(hfinal : 0 ≤ widths (n - 1) (times n))
(hslope : ∀ i < n, ∀ x ∈ Ico (times i) (times (i + 1)), ∀ r : ℝ,
-(4 * Real.pi) + 3 / (4 * (x + C)) * widths i x < r →
∃ᶠ z in 𝓝[>] x, slope (widths i) x z < r)
(hjumps : ∀ i, i + 1 < n →
widths (i + 1) (times (i + 1)) ≤ widths i (times (i + 1))) :
times n ≤ (C ^ ((1 : ℝ) / 4) + W₀ / (16 * Real.pi * C ^ ((3 : ℝ) / 4))) ^ (4 : ℕ) - C := by sorrySource
Colding-Minicozzi, Width and Finite Extinction Time of Ricci Flow, https://arxiv.org/pdf/0707.0108, p.4, Theorem 1.7 and equations (1.8)-(1.10). Analytic consequence of (1.10), extended to finitely many downward jumps by interval comparison; lower-Dini hypotheses are sufficient. The geometric surgery application is separate.