Independent cokernel elements of a Cauchy–Riemann operator from approximate solutions of the adjoint equation
ProvedCauchyRiemannIndex.linearIndependent_cokernel_of_adjoint_smallLet be continuous and bounded, and let , acting from to , with formal adjoint . Suppose that for some
and let be smooth sections, -periodic in and vanishing for large , such that every nonzero combination () satisfies
Then the cokernel contains linearly independent elements; in particular .
Formalization Note is W1p 2 n, is the norm of crApply T on a triple, and the norms in the estimate are written with eLpNorm (values in , with entering through ENNReal.ofReal). The test sections are IsTestSection, and is written out with partialS, partialT and the transpose (T x)ᵀ. Boundedness of is the bound for all and . The image is the submodule crRange 2 T of LpCyl n 2, and the conclusion is a linearly independent family Fin k → LpCyl n 2 ⧸ crRange 2 T.
import Definitions.Def_CauchyRiemannIndex_Setting import Mathlib.MeasureTheory.Integral.Bochner.Basic
namespace CauchyRiemannIndex
open MeasureTheory Matrix Filter Topology
open scoped ENNReal NNReal ContDiff RealInnerProductSpace
/-- Lower bound for the cokernel of `D_T = ∂ₛ + J₀ ∂ₜ + T : W^{1,2} → L²` on the cylinder, for a
continuous bounded coefficient `T`: suppose `‖η‖²_{L²} ≤ c ‖D_T η‖²_{L²}` on
`W^{1,2}`, and let `v₁, …, v_k` be test sections such that every nonzero combination `v = Σ gᵢ vᵢ`
satisfies `c ‖D_T* v‖²_{L²} < ‖v‖²_{L²}`, where `D_T* v = -∂ₛ v + J₀ ∂ₜ v + Tᵀ v` is the formal
adjoint. Then the cokernel `L² / im D_T` contains `k` linearly independent elements. -/
theorem linearIndependent_cokernel_of_adjoint_small (n : ℕ) (T : ℝ × ℝ → Mat n)
(hT : Continuous T)
(hTb : ∃ M : ℝ, ∀ x (v : Vec n), ‖Matrix.toEuclideanLin (T x) v‖ ≤ M * ‖v‖)
(c : ℝ) (hc : 0 ≤ c)
(hest : ∀ w ∈ W1p 2 n, eLpNorm (w.1 : ℝ × ℝ → Vec n) 2 cylMeasure ^ 2 ≤
ENNReal.ofReal c * eLpNorm (crApply T w) 2 cylMeasure ^ 2)
(k : ℕ) (v : Fin k → ℝ × ℝ → Vec n) (hv : ∀ i, IsTestSection (v i))
(hsmall : ∀ g : Fin k → ℝ, g ≠ 0 →
c * ∫ x, ‖∑ i, g i • (-partialS (v i) x + Matrix.toEuclideanLin (J₀ n) (partialT (v i) x) +
Matrix.toEuclideanLin (T x)ᵀ (v i x))‖ ^ 2 ∂cylMeasure <
∫ x, ‖∑ i, g i • v i x‖ ^ 2 ∂cylMeasure) :
∃ u : Fin k → LpCyl n 2 ⧸ crRange 2 T, LinearIndependent ℝ u := by
sorry
end CauchyRiemannIndex