roseking's picture
Add lean source and binary files (part 11)
110f4f3 verified
Raw History Blame Contribute Delete
4.63 kB
import Mathlib
import OAI.Analysis.BiholderTransport.Geodesics.MinimizingSubinterval
import OAI.Analysis.BiholderTransport.Geodesics.GlobalGauss
namespace OAI
noncomputable section
open Set Filter Manifold Bundle Metric
open scoped Topology ContDiff NNReal
namespace WeakMTWTransport
variable {n : ℕ} {M : Type*} [MetricSpace M] [CompactSpace M]
[ChartedSpace (Model n) M] [IsManifold 𝓘(ℝ,Model n) ∞ M]
[RiemannianBundle (fun x : M => TangentSpace 𝓘(ℝ,Model n) x)]
[IsContMDiffRiemannianBundle 𝓘(ℝ,Model n) ∞ (Model n)
(fun x : M => TangentSpace 𝓘(ℝ,Model n) x)]
attribute [local instance] normedAddCommGroupTangentSpaceVectorSpace
normedSpaceTangentSpaceVectorSpace
lemma chart_spray_hasDerivAt (a : M) (z : TangentBundle 𝓘(ℝ,Model n) M) {t : ℝ}
(ht : (sprayFlow t z).1∈(extChartAt 𝓘(ℝ,Model n) a).source) :
HasDerivAt (fun s => extChartAt 𝓘(ℝ,Model n) a (sprayFlow s z).1)
(mfderiv 𝓘(ℝ,Model n) 𝓘(ℝ,Model n) (extChartAt 𝓘(ℝ,Model n) a)
(sprayFlow t z).1 (sprayFlow t z).2) t := by
have hsrc : (sprayFlow t z).1∈(chartAt (Model n) a).source := by
simpa only [extChartAt_source] using ht
have projectedDeriv :=
(ContinuousLinearMap.fst ℝ (Model n) (Model n)).hasFDerivAt.comp_hasDerivAt t
(sprayFlow_chart_hasDerivAt (⟨a,0⟩ : TangentBundle 𝓘(ℝ,Model n) M) z ht)
change HasDerivAt (fun s => extChartAt 𝓘(ℝ,Model n) a (sprayFlow s z).1)
(tangentCoordChange 𝓘(ℝ,Model n) (sprayFlow t z).1 a
(sprayFlow t z).1 (sprayFlow t z).2) t at projectedDeriv
apply projectedDeriv.congr_deriv
symm
change (mfderiv 𝓘(ℝ,Model n) 𝓘(ℝ,Model n) (chartAt (Model n) a)
(sprayFlow t z).1) (sprayFlow t z).2 = _
rw [mfderiv_chartAt_eq_tangentCoordChange hsrc]
rfl
variable [IsRiemannianManifold 𝓘(ℝ,Model n) M]
lemma exists_lipschitzOn_chart (a : M) :
∃ C : ℝ≥0, ∃ r>0, ball a r⊆(extChartAt 𝓘(ℝ,Model n) a).source ∧
LipschitzOnWith C (extChartAt 𝓘(ℝ,Model n) a) (ball a r) := by
have : IsContinuousRiemannianBundle (Model n)
(fun x : M => TangentSpace 𝓘(ℝ,Model n) x) :=
continuousRiemannianBundle_of_smooth (IB := 𝓘(ℝ,Model n))
obtain ⟨C,hC,H⟩ := eventually_norm_mfderiv_extChartAt_lt 𝓘(ℝ,Model n) a
obtain ⟨R,hR,hsub⟩ := Metric.mem_nhds_iff.mp
(inter_mem (extChartAt_source_mem_nhds (I := 𝓘(ℝ,Model n)) a) H)
let Cnn : ℝ≥0 := ⟨C,hC.le⟩
refine ⟨Cnn,R/3,by positivity,?_,?_⟩
· exact fun x hx => (hsub ((ball_subset_ball (by linarith : R/3≤R)) hx)).1
apply LipschitzOnWith.of_dist_le_mul
intro x hx y hy
obtain ⟨p,hp,he⟩ := exists_minimizing_vector (n := n) x y
let z : TangentBundle 𝓘(ℝ,Model n) M := ⟨x,p⟩
have hpN : ‖p‖=dist x y := by rw [←he]; exact hp.symm
have hpath : ∀ t∈Icc (0:ℝ) 1, (sprayFlow t z).1∈ball a R := by
intro t ht
have hD := dist_sprayFlow_le z 0 t
simp only [sprayFlow_zero,zero_sub,abs_neg,abs_of_nonneg ht.1] at hD
change dist x (sprayFlow t z).1≤t*‖p‖ at hD
have hD' : dist x (sprayFlow t z).1≤dist x y :=
hD.trans (by rw [←hpN]; exact mul_le_of_le_one_left (norm_nonneg _) ht.2)
have hxy := dist_triangle x a y
have hax : dist x a<R/3 := hx
have hay : dist y a<R/3 := hy
have htri := dist_triangle (sprayFlow t z).1 x a
rw [dist_comm (sprayFlow t z).1 x] at htri
rw [dist_comm a y] at hxy
change dist (sprayFlow t z).1 a<R
linarith
let γ := fun t : ℝ => extChartAt 𝓘(ℝ,Model n) a (sprayFlow t z).1
let dγ : ℝ → Model n := fun t => mfderiv 𝓘(ℝ,Model n) 𝓘(ℝ,Model n)
(extChartAt 𝓘(ℝ,Model n) a) (sprayFlow t z).1 (sprayFlow t z).2
have hd : ∀ t∈Icc (0:ℝ) 1, HasDerivWithinAt γ (dγ t) (Icc 0 1) t := by
intro t ht
exact (chart_spray_hasDerivAt a z (hsub (hpath t ht)).1).hasDerivWithinAt
have hb : ∀ t∈Icc (0:ℝ) 1, ‖dγ t‖≤C*‖p‖ := by
intro t ht
apply (ContinuousLinearMap.le_opNorm _ _).trans
rw [sprayFlow_speed]
exact mul_le_mul_of_nonneg_right (hsub (hpath t ht)).2.le (norm_nonneg _)
have hfin := norm_image_sub_le_of_norm_deriv_le_segment_01' hd (fun t ht => hb t ⟨ht.1,ht.2.le⟩)
have he1 : (sprayFlow 1 z).1=y := by
simpa only [z,←riemannianExp_eq_sprayFlow] using he
have he0 : (sprayFlow 0 z).1=x := by rw [sprayFlow_zero]
change dist (extChartAt 𝓘(ℝ,Model n) a x) (extChartAt 𝓘(ℝ,Model n) a y)≤C*dist x y
simpa only [γ,he1,he0,←dist_eq_norm,dist_comm,hpN] using hfin
end WeakMTWTransport
end
end OAI