File size: 2,646 Bytes
4a56cd6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 | import OAI.Geometry.Kakeya.Quadratic.Compression
import Mathlib.Topology.MetricSpace.HausdorffDimension
open Set Module
open scoped Topology ContDiff ENNReal
noncomputable section
namespace OAI.Geometry.Kakeya.Quadratic
variable {E D : Type*} [NormedAddCommGroup E] [NormedSpace β E]
[FiniteDimensional β E] [NormedAddCommGroup D] [NormedSpace β D]
[FiniteDimensional β D]
def ruledConstraint (P Q : E ββ[β] D) : Set (E Γ β) :=
{z | P z.1 = z.2 β’ Q z.1}
/-- The ruled constraint is parametrized by ker P times the real line.
This proves its dimension bound directly, without a manifold dimension
theorem or a Jordan normal form. -/
theorem dimH_ruledConstraint (P Q : E ββ[β] D)
(hs : Function.Surjective (P.prod Q)) :
dimH (ruledConstraint P Q) β€ (finrank β E - finrank β D + 1 : β) := by
obtain β¨R, hRβ© := (P.prod Q).exists_rightInverse_of_surjective
(LinearMap.range_eq_top.mpr hs)
have hP (a : D) : P (R (a, 0)) = a :=
congrArg (fun f : D Γ D ββ[β] D Γ D => (f (a, 0)).1) hR
have hQ (a : D) : Q (R (a, 0)) = 0 :=
congrArg (fun f : D Γ D ββ[β] D Γ D => (f (a, 0)).2) hR
let f : P.ker Γ β β E Γ β := fun z =>
((z.1 : E) + z.2 β’ R (Q (z.1 : E), 0), z.2)
have hf : ContDiff β β f := by
have hc : ContDiff β β (fun z : P.ker => (z : E)) := P.ker.subtypeL.contDiff
have hRc : ContDiff β β R := R.toContinuousLinearMap.contDiff
have hQc : ContDiff β β Q := Q.toContinuousLinearMap.contDiff
dsimp [f]
fun_prop
have hsub : ruledConstraint P Q β Set.range f := by
intro z hz
let v : P.ker := β¨z.1 - z.2 β’ R (Q z.1, 0), by
change P (z.1 - z.2 β’ R (Q z.1, 0)) = 0
rw [map_sub, map_smul, hP]
exact sub_eq_zero.mpr hzβ©
refine β¨(v, z.2), ?_β©
apply Prod.ext
Β· change (z.1 - z.2 β’ R (Q z.1, 0)) +
z.2 β’ R (Q (z.1 - z.2 β’ R (Q z.1, 0)), 0) = z.1
rw [map_sub, map_smul, hQ, smul_zero, sub_zero]
abel
Β· rfl
have hPs : Function.Surjective P := by
intro a
obtain β¨x, hxβ© := hs (a, 0)
exact β¨x, congrArg Prod.fst hxβ©
have hrank := P.finrank_range_add_finrank_ker
rw [LinearMap.range_eq_top.mpr hPs, finrank_top] at hrank
have hk : finrank β P.ker = finrank β E - finrank β D := by omega
calc
dimH (ruledConstraint P Q) β€ dimH (Set.range f) := dimH_mono hsub
_ β€ finrank β (P.ker Γ β) := (hf.differentiable (by norm_num)).dimH_range_le
_ = (finrank β E - finrank β D + 1 : β) := by
rw [Module.finrank_prod, hk, Module.finrank_self]
end OAI.Geometry.Kakeya.Quadratic
|