Euclidean varifolds and their weight measures
DefinitionOpenGA_EuclideanVarifoldgeometric-measure-theoryopenga-varifoldpoincare-foundationsvarifold
A -varifold on a finite-dimensional real inner-product space is a nonnegative Radon measure on . Its weight is the spatial pushforward and its mass is the total measure, which may be infinite. Compactness of makes the weight measure Radon. This interface includes compactly supported test integrals, support, addition and nonnegative scaling. It imposes neither rectifiability nor stationarity. A general manifold requires its tangent Grassmann bundle, not this product.
Definition code
import Definitions.Def_OpenGA_UnorientedGrassmannian
import Mathlib.MeasureTheory.Measure.Regular
import Mathlib.MeasureTheory.Measure.Support
import Mathlib.MeasureTheory.Integral.CompactlySupported
import Mathlib.MeasureTheory.Integral.Bochner.Basic
/-!
# Euclidean varifolds and their weight measures
A `k`-varifold on `E` is a nonnegative Radon measure on
`E × Grassmannian E k`. The `Measure.Regular` field includes finiteness on
compact sets and regularity; total mass need not be finite. On this locally
compact metric space this gives the Radon measures used by Simon.
Reference: Leon Simon, *Introduction to Geometric Measure Theory*, 2018,
Chapter 8, Section 1, pp. 235-236. This module treats the Euclidean ambient
space. It does not identify the product with the Grassmann bundle of an
arbitrary manifold, and does not impose rectifiability or stationarity.
-/
noncomputable section
open MeasureTheory Set
open scoped ENNReal NNReal Topology CompactlySupported
namespace OpenGA
/-- A Radon measure on positions and unoriented `k`-planes. -/
structure Varifold (E : Type*) [NormedAddCommGroup E] [InnerProductSpace ℝ E]
[FiniteDimensional ℝ E] [MeasurableSpace E] [BorelSpace E] (k : ℕ) where
measure : Measure (E × Grassmannian E k)
regular : measure.Regular
attribute [instance] Varifold.regular
namespace Varifold
variable {E : Type*} [NormedAddCommGroup E] [InnerProductSpace ℝ E]
[FiniteDimensional ℝ E] [MeasurableSpace E] [BorelSpace E] {k : ℕ}
@[ext] theorem ext {V W : Varifold E k} (h : V.measure = W.measure) : V = W := by
cases V
cases W
cases h
rfl
/-- Spatial projection of a varifold, conventionally denoted `‖V‖`. -/
def weightMeasure (V : Varifold E k) : Measure E := V.measure.map Prod.fst
theorem weightMeasure_apply (V : Varifold E k) {A : Set E} (hA : MeasurableSet A) :
V.weightMeasure A = V.measure (A ×ˢ Set.univ) := by
rw [weightMeasure, Measure.map_apply measurable_fst hA]
congr 1
ext z
simp
instance weightMeasure_finiteOnCompacts (V : Varifold E k) :
IsFiniteMeasureOnCompacts V.weightMeasure := by
constructor
intro K hK
rw [weightMeasure_apply V hK.measurableSet]
exact (hK.prod isCompact_univ).measure_lt_top
/-- Compactness of the Grassmannian ensures that spatial projection is again Radon. -/
instance weightMeasure_regular (V : Varifold E k) : V.weightMeasure.Regular :=
inferInstance
/-- Total mass, allowed to be infinite for a general varifold. -/
def mass (V : Varifold E k) : ℝ≥0∞ := V.measure Set.univ
theorem mass_eq_weightMeasure_univ (V : Varifold E k) :
V.mass = V.weightMeasure Set.univ := by
simp [mass, weightMeasure_apply V MeasurableSet.univ]
theorem weightMeasure_compact_lt_top (V : Varifold E k) {K : Set E} (hK : IsCompact K) :
V.weightMeasure K < ∞ := hK.measure_lt_top
/-- The support in the ambient space is the support of the weight measure. -/
def support (V : Varifold E k) : Set E := V.weightMeasure.support
theorem isClosed_support (V : Varifold E k) : IsClosed V.support :=
Measure.isClosed_support
theorem mem_support_iff (V : Varifold E k) (x : E) :
x ∈ V.support ↔ ∀ U ∈ nhds x, 0 < V.weightMeasure U :=
V.weightMeasure.mem_support_iff_forall x
instance : Zero (Varifold E k) := ⟨⟨0, inferInstance⟩⟩
instance : Add (Varifold E k) := ⟨fun V W => by
haveI : IsFiniteMeasureOnCompacts (V.measure + W.measure) := ⟨fun K hK => by
rw [Measure.add_apply]
exact ENNReal.add_lt_top.mpr ⟨hK.measure_lt_top, hK.measure_lt_top⟩⟩
exact ⟨V.measure + W.measure, inferInstance⟩⟩
instance : SMul ℝ≥0 (Varifold E k) := ⟨fun c V => ⟨c • V.measure, inferInstance⟩⟩
@[simp] theorem measure_zero : (0 : Varifold E k).measure = 0 := rfl
@[simp] theorem measure_add (V W : Varifold E k) : (V + W).measure = V.measure + W.measure := rfl
@[simp] theorem measure_smul (c : ℝ≥0) (V : Varifold E k) :
(c • V).measure = c • V.measure := rfl
@[simp] theorem weightMeasure_zero : (0 : Varifold E k).weightMeasure = 0 := by
simp [weightMeasure]
@[simp] theorem weightMeasure_add (V W : Varifold E k) :
(V + W).weightMeasure = V.weightMeasure + W.weightMeasure := by
exact Measure.map_add _ _ measurable_fst
@[simp] theorem mass_zero : (0 : Varifold E k).mass = 0 := by simp [mass]
@[simp] theorem mass_add (V W : Varifold E k) : (V + W).mass = V.mass + W.mass :=
Measure.add_apply _ _ _
/-- Pairing with a continuous, compactly supported test function on positions and planes. -/
def testIntegral (V : Varifold E k) (φ : C_c(E × Grassmannian E k, ℝ)) : ℝ :=
∫ z, φ z ∂V.measure
theorem integrable_testFunction (V : Varifold E k) (φ : C_c(E × Grassmannian E k, ℝ)) :
Integrable φ V.measure := φ.integrable
@[simp] theorem testIntegral_add (V W : Varifold E k)
(φ : C_c(E × Grassmannian E k, ℝ)) :
(V + W).testIntegral φ = V.testIntegral φ + W.testIntegral φ :=
integral_add_measure φ.integrable φ.integrable
/-- Integrating a spatial observable against the weight equals integrating its lift. -/
theorem integral_weightMeasure (V : Varifold E k) {f : E → ℝ}
(hf : AEStronglyMeasurable f V.weightMeasure) :
∫ x, f x ∂V.weightMeasure = ∫ z, f z.1 ∂V.measure :=
integral_map measurable_fst.aemeasurable hf
end Varifold
end OpenGA
Source