Documentation

LeanPool.HighContrastHomogenization.Provider.Response.PreYoungWeakTop

High-contrast homogenization: Provider.Response.PreYoungWeakTop #

Imported from the Apache-2.0 HighContrastHomogenization development at commit 7a13dbcd8d6609264a713373f5c69ceeac870472.

The infinite weak-quantity branch #

The compact pre-Young estimate is automatic when its weak quantity is infinite and the response coefficient is positive. Separating this branch keeps finiteness assumptions local to the analytic component estimates.

theorem HCPolySupport.HighContrast.Response.le_compact_pre_young_of_weak_eq_top {J a b row weak expo : ENNReal} {Cresp : ℝ} (hCresp : 0 < Cresp) (hweak : weak = ⊤) :
J ≤ ENNReal.ofReal Cresp * a * (a + b + row) + ENNReal.ofReal Cresp * expo * b * (b + row) + ENNReal.ofReal Cresp * weak

The compact pre-Young right-hand side is top when its weak quantity is top and its coefficient is positive.

theorem HCPolySupport.HighContrast.Response.compact_pre_young_of_finite_weak {J a b row weak expo : ENNReal} {Cpre : ℝ} (hCpre : 1 ≤ Cpre) (hfinite : weak ≠ ⊤ → ∀ (Cresp : ℝ), Cpre ≤ Cresp → J ≤ ENNReal.ofReal Cresp * a * (a + b + row) + ENNReal.ofReal Cresp * expo * b * (b + row) + ENNReal.ofReal Cresp * weak) (Cresp : ℝ) :
Cpre ≤ Cresp → J ≤ ENNReal.ofReal Cresp * a * (a + b + row) + ENNReal.ofReal Cresp * expo * b * (b + row) + ENNReal.ofReal Cresp * weak

It suffices to prove the compact estimate under finiteness of the weak quantity; the complementary branch is automatic.

theorem HCPolySupport.HighContrast.Response.compact_pre_young_of_finite_component_bounds {J boundaryRow a b scaleRoot rowRoot weak expo : ENNReal} {Ccut Crow Cdiv Cpre : ℝ} (hCpre : 1 ≤ Cpre) (hcomponents : weak ≠ ⊤ → ∃ (divCurl : ENNReal) (cutoffEnergy : ENNReal), J ≤ divCurl + cutoffEnergy + boundaryRow ∧ divCurl ≤ ENNReal.ofReal Cdiv * weak ∧ cutoffEnergy ≤ ENNReal.ofReal 2 * a * a + ENNReal.ofReal 4 * a * b + ENNReal.ofReal Ccut * expo * b * b ∧ boundaryRow ≤ ENNReal.ofReal √2 * a * scaleRoot + ENNReal.ofReal Crow * expo * b * rowRoot) (hscaleRoot : scaleRoot ≤ rowRoot) (hfour : 4 ≤ Cpre) (hsqrtTwo : √2 ≤ Cpre) (hcut : Ccut ≤ Cpre) (hrow : Crow ≤ Cpre) (hdiv : Cdiv ≤ Cpre) (Cresp : ℝ) :
Cpre ≤ Cresp → J ≤ ENNReal.ofReal Cresp * a * (a + b + rowRoot) + ENNReal.ofReal Cresp * expo * b * (b + rowRoot) + ENNReal.ofReal Cresp * weak

Component estimates are needed only in the finite weak-quantity branch. The resulting compact estimate remains unconditional.