Documentation

LeanPool.EllipticPDE.Regularity.Local.Reduction

Cutoff reduction of a local weak solution to an H₀¹ problem #

Evans, Partial Differential Equations (2nd ed.), §6.3.1, Theorem 1, step 1 tests the equation against -D_k^{-h}(ζ² D_k^h u), which is admissible for u ∈ H¹(U) because the cutoff ζ removes the boundary. This file uses the same cutoff differently: for a local weak solution U ∈ W12 Ω and a test function η of Ω, the product η U lies in H₀¹(Ω) (cutoffMul_mem_H01_of_mem_W12) and solves B[η U, w] = ∫ F w for every w ∈ H₀¹(Ω), with

F = η f − Σ a_{ij} ∂_j η U_i − Σ ∂_j a_{ij} ∂_i η U_0 − Σ a_{ij} ∂_i η U_j − Σ a_{ij} ∂_{ij} η U_0 + Σ b_i ∂_i η U_0.

Every term of F is in L²(Ω) with a bound in ‖f‖ and ‖U‖, so the interior chain for H₀¹ solutions applies to η U unchanged, and η = 1 near the set of interest makes the cutoff invisible in its conclusion.

The derivative ∂_j a_{ij} enters through one integration by parts, the only place a derivative of a coefficient appears. CoeffWeakGrad records what that step needs: a weak partial derivative of each entry, measurable and essentially bounded. Both a C¹ bundle (IsC1Coeff.coeffWeakGrad) and a W^{k+1,∞} bundle (IsWkInftyCoeff.coeffWeakGrad) supply it, so one reduction serves the H² estimate and the higher-order induction.

Main declarations #

Weak gradient of the principal coefficients. For each direction l and entry (i, j), a measurable function da l i j, essentially bounded by one constant, that is the weak partial derivative of a_{ij} in direction l. This is all the cutoff reduction asks of the principal part beyond FullEllipticOp.

Instances For
    theorem EllipticPdes.Regularity.IsC1Coeff.abs_partialD_le {d : ℕ} {A : Sobolev.EllipticCoeff d} (hA : IsC1Coeff A) (i j k : Fin d) (x : EuclideanSpace ℝ (Fin d)) :
    |Sobolev.partialD k (fun (y : EuclideanSpace ℝ (Fin d)) => A.a y i j) x| ≤ hA.A1

    Pointwise bound on a partial of a C¹ coefficient.

    A partial of a C¹ coefficient is measurable, being continuous.

    The classical partials of a C¹ coefficient are its weak partials.

    Equations
    • One or more equations did not get rendered due to their size.
    Instances For

      The first-order members of a W^{k+1,∞} family are weak partials of the coefficient.

      Equations
      • hA.coeffWeakGrad = { da := fun (l i j : Fin d) => hA.D [l] i j, measurable := ⋯, bound := hA.bound 1, ae_abs_le := ⋯, hasWeakPartial := ⋯ }
      Instances For
        theorem EllipticPdes.Regularity.integrable_weight_L2D_mul {d : ℕ} {Ω : Set (EuclideanSpace ℝ (Fin d))} {g : Sobolev.L2D Ω} {c : EuclideanSpace ℝ (Fin d) → ℝ} (hcm : Measurable c) {M : ℝ} (hc : ∀ᵐ (x : EuclideanSpace ℝ (Fin d)), |c x| ≤ M) {ψ : EuclideanSpace ℝ (Fin d) → ℝ} (hψ : Continuous ψ) (hψcs : HasCompactSupport ψ) :
        MeasureTheory.Integrable (fun (x : EuclideanSpace ℝ (Fin d)) => c x * ↑↑g x * ψ x) (MeasureTheory.volume.restrict Ω)

        A bounded measurable weight times an L²(Ω) class times a continuous compactly supported function is integrable on Ω.

        theorem EllipticPdes.Regularity.principal_leibniz {d : ℕ} {Ω : Set (EuclideanSpace ℝ (Fin d))} (Op : Sobolev.FullEllipticOp d) (D : CoeffWeakGrad Op.toEllipticCoeff) {U : Sobolev.H1amb Ω} (hU : U ∈ Sobolev.W12 Ω) (i j : Fin d) {η v : EuclideanSpace ℝ (Fin d) → ℝ} (hη : Sobolev.IsTestFn Ω η) (hv : Sobolev.IsTestFn Ω v) :
        ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, Op.a x i j * ↑↑(U.ofLp 0) x * (Sobolev.partialD i η x * Sobolev.partialD j v x) = ((-∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, D.da j i j x * ↑↑(U.ofLp 0) x * (Sobolev.partialD i η x * v x)) - ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, Op.a x i j * ↑↑(U.ofLp j.succ) x * (Sobolev.partialD i η x * v x)) - ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, Op.a x i j * ↑↑(U.ofLp 0) x * (Sobolev.partialD j (Sobolev.partialD i η) x * v x)

        Integration by parts on the principal term. The term ∫ a_{ij} U₀ ∂_i η ∂_j v with the derivative moved off v. It is the only place a derivative of a coefficient enters, and it is taken in the weak sense through HasWeakDerivOn.mul_isWkInfty_left, so the test function stays smooth and Ω need not have finite measure.

        theorem EllipticPdes.Regularity.reduction_testFn {d : ℕ} {Ω : Set (EuclideanSpace ℝ (Fin d))} (Op : Sobolev.FullEllipticOp d) (hΩo : IsOpen Ω) (D : CoeffWeakGrad Op.toEllipticCoeff) {U : Sobolev.H1amb Ω} {f : Sobolev.L2D Ω} (hsol : IsLocalWeakSolution Op Ω U f) {η : EuclideanSpace ℝ (Fin d) → ℝ} (hη : Sobolev.IsTestFn Ω η) {v : EuclideanSpace ℝ (Fin d) → ℝ} (hv : Sobolev.IsTestFn Ω v) :
        ((Op.fullBilin Ω) ⟨(cutoffMul hη) U, ⋯⟩) ⟨hv.testGraph, ⋯⟩ = (((((∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, ↑↑f x * (η x * v x)) - ∑ i : Fin d, ∑ j : Fin d, ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, Op.a x i j * ↑↑(U.ofLp i.succ) x * (Sobolev.partialD j η x * v x)) - ∑ i : Fin d, ∑ j : Fin d, ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, D.da j i j x * ↑↑(U.ofLp 0) x * (Sobolev.partialD i η x * v x)) - ∑ i : Fin d, ∑ j : Fin d, ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, Op.a x i j * ↑↑(U.ofLp j.succ) x * (Sobolev.partialD i η x * v x)) - ∑ i : Fin d, ∑ j : Fin d, ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, Op.a x i j * ↑↑(U.ofLp 0) x * (Sobolev.partialD j (Sobolev.partialD i η) x * v x)) + ∑ i : Fin d, ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, Op.b x i * ↑↑(U.ofLp 0) x * (Sobolev.partialD i η x * v x)

        Cutoff reduction against test functions. For a local weak solution U ∈ W12 Ω and a test function η of Ω, the element η U ∈ H₀¹(Ω) satisfies B[η U, v] = ∫ F v for every test function v, with F spelled out as separate integrals.

        The datum as an L²(Ω) class #

        noncomputable def EllipticPdes.Regularity.weightL {d : ℕ} (Ω : Set (EuclideanSpace ℝ (Fin d))) {c ψ : EuclideanSpace ℝ (Fin d) → ℝ} (hcm : Measurable c) {M : ℝ} (hc : ∀ᵐ (x : EuclideanSpace ℝ (Fin d)), |c x| ≤ M) (hψ : Continuous ψ) (hψcs : HasCompactSupport ψ) :

        Multiplication by c · ψ on L²(Ω), for c essentially bounded and ψ continuous with compact support.

        Equations
        Instances For
          theorem EllipticPdes.Regularity.weightL_bound_nonneg {d : ℕ} {ψ : EuclideanSpace ℝ (Fin d) → ℝ} (hψ : Continuous ψ) (hψcs : HasCompactSupport ψ) {M : ℝ} :
          0 ≤ max M 0 * ⋯.choose
          theorem EllipticPdes.Regularity.norm_weightL_le {d : ℕ} {Ω : Set (EuclideanSpace ℝ (Fin d))} {c ψ : EuclideanSpace ℝ (Fin d) → ℝ} (hcm : Measurable c) {M : ℝ} (hc : ∀ᵐ (x : EuclideanSpace ℝ (Fin d)), |c x| ≤ M) (hψ : Continuous ψ) (hψcs : HasCompactSupport ψ) (g : Sobolev.L2D Ω) :
          ‖(weightL Ω hcm hc hψ hψcs) g‖ ≤ max M 0 * ⋯.choose * ‖g‖
          theorem EllipticPdes.Regularity.inner_weightL_testCls {d : ℕ} {Ω : Set (EuclideanSpace ℝ (Fin d))} {c ψ : EuclideanSpace ℝ (Fin d) → ℝ} (hcm : Measurable c) {M : ℝ} (hc : ∀ᵐ (x : EuclideanSpace ℝ (Fin d)), |c x| ≤ M) (hψ : Continuous ψ) (hψcs : HasCompactSupport ψ) (g : Sobolev.L2D Ω) {v : EuclideanSpace ℝ (Fin d) → ℝ} (hv : Sobolev.IsTestFn Ω v) :
          inner ℝ ((weightL Ω hcm hc hψ hψcs) g) hv.testCls = ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, c x * ↑↑g x * (ψ x * v x)
          theorem EllipticPdes.Regularity.inner_testCls_eq {d : ℕ} {Ω : Set (EuclideanSpace ℝ (Fin d))} {g : Sobolev.L2D Ω} {v : EuclideanSpace ℝ (Fin d) → ℝ} (hv : Sobolev.IsTestFn Ω v) :
          inner ℝ g hv.testCls = ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, ↑↑g x * v x
          noncomputable def EllipticPdes.Regularity.redDatum {d : ℕ} {Ω : Set (EuclideanSpace ℝ (Fin d))} (Op : Sobolev.FullEllipticOp d) (D : CoeffWeakGrad Op.toEllipticCoeff) {η : EuclideanSpace ℝ (Fin d) → ℝ} (hη : Sobolev.IsTestFn Ω η) (U : Sobolev.H1amb Ω) (f : Sobolev.L2D Ω) :

          Reduction datum F for η U, as an L²(Ω) class: each term is a bounded weight supported in tsupport η against f, U₀ or a gradient coordinate of U.

          Equations
          • One or more equations did not get rendered due to their size.
          Instances For
            theorem EllipticPdes.Regularity.reduction_weakForm {d : ℕ} {Ω : Set (EuclideanSpace ℝ (Fin d))} (Op : Sobolev.FullEllipticOp d) (D : CoeffWeakGrad Op.toEllipticCoeff) {η : EuclideanSpace ℝ (Fin d) → ℝ} (hη : Sobolev.IsTestFn Ω η) (hΩo : IsOpen Ω) {U : Sobolev.H1amb Ω} {f : Sobolev.L2D Ω} (hsol : IsLocalWeakSolution Op Ω U f) (w : ↥(Sobolev.H01 Ω)) :
            ((Op.fullBilin Ω) ⟨(cutoffMul hη) U, ⋯⟩) w = ∫ (x : EuclideanSpace ℝ (Fin d)) in Ω, ↑↑(redDatum Op D hη U f) x * ↑↑((↑w).ofLp 0) x

            Cutoff reduction on H₀¹(Ω). For a local weak solution U ∈ W12 Ω and a test function η of Ω, B[η U, w] = ∫ F w for every w ∈ H₀¹(Ω), with F = redDatum.