Existence of Minimisers in the Plateau Problem Anthony Salib - - PowerPoint PPT Presentation

existence of minimisers in the plateau problem
SMART_READER_LITE
LIVE PREVIEW

Existence of Minimisers in the Plateau Problem Anthony Salib - - PowerPoint PPT Presentation

Existence of Minimisers in the Plateau Problem Anthony Salib 09/04/2020 The Plateau Problem Given a boundary in R n , does there exist a surface that is bounded by such that the area of it is minimal. The Plateau Problem Given a boundary


slide-1
SLIDE 1

Existence of Minimisers in the Plateau Problem

Anthony Salib 09/04/2020

slide-2
SLIDE 2

The Plateau Problem

Given a boundary Γ in Rn, does there exist a surface that is bounded by Γ such that the area of it is minimal.

slide-3
SLIDE 3

The Plateau Problem

Given a boundary Γ in Rn, does there exist a surface that is bounded by Γ such that the area of it is minimal.

slide-4
SLIDE 4

Sets of finite perimeter

We have the following Plateau type problem in some A ⊂ Rn with boundary data E0, γ = inf{P(E) : E\A = E0\A, P(E) < ∞}.

slide-5
SLIDE 5

Sets of finite perimeter

γ = inf{P(E) : E\A = E0\A, P(E) < ∞}. The existence of minimisers can be established using the direct method: Construct a minimising sequence {Eh}h∈N Extract a convergent subsequence, Ehk → E Show that P(E) = γ

slide-6
SLIDE 6

”Area”

For a k-dimensional set, M ⊂ Rn we settle on the k-dimensional Hausdorff measure as our area functional. That is A(M) =

  • M

dHk.

slide-7
SLIDE 7

”Bounded by”

We will take bounded by Γ to mean that the surface Σ satisfies ∂Σ = Γ.

slide-8
SLIDE 8

Surface

?

slide-9
SLIDE 9

Outline

Submanifolds Currents Varifolds

slide-10
SLIDE 10

Submanifolds

How can we tell if a given submanifold M ⊂ Rn has minimal area?

slide-11
SLIDE 11

Monotonicity and Density

Theorem (Monotonicity)

Suppose M, k-dimensional, is stationary in Rn and fix x ∈ Rn. Then ω−1

k r−kA(M ∩ Br(x))

is an increasing function of r for 0 < r ≤ dist(x, ∂M). We define the density of M at p ∈ M\∂M to be Θ(M, p) = lim

r→0 ω−1 k r−kA(M ∩ Br(x)).

slide-12
SLIDE 12

The disk with Spines

Figure: A minimising sequence may not converge to a Minimal Surface

slide-13
SLIDE 13

Fleming’s Example

Figure: A minimal surface with infinite genus

slide-14
SLIDE 14

Currents

Let U ⊂ Rn and we let Dk(U) denote the space of compactly supported k-forms on U.

Definition (Currents)

A k-current is a linear functional on Dk(U). The space of k-currents, we will denote as Dk(U). Given T ∈ Dk(U), the boundary is defined to be ∂T ∈ Dk−1 such that for any ω ∈ Dk−1(U) ∂T(ω) = T(dω).

slide-15
SLIDE 15

Mass of Currents

Definition (Mass)

Given a current T, we define it’s mass to be M(T) = sup

ω≤1

T(ω).

slide-16
SLIDE 16

Currents

Theorem

The space of currents with finite mass is a Banach space with norm M.

slide-17
SLIDE 17

Plateau Problem

Given a k-current S with ∂S = ∅, is there a k + 1-current T such that ∂T = S and M(T) is minimal.

slide-18
SLIDE 18

Rectifiable Integer Currents

Definition

Let U ⊂ Rn. A rectifiable integer k-current (integral k-current)is a current T such that for ω ∈ Dk(U) T(ω) =

  • M

< ω(x), ξ(x) > θ(x)dHk(x), where M ⊂ U is countably k-rectifiable, θ is a Hk integrable function that takes values positive integers and ξ : M → (Λk(Rn))∗ is Hk measurable which can be expressed at almost every point x, ξ(x) = τ1 ∧ · · · ∧ τn where τ1, . . . , τn is an orthonormal basis for the approximate tangent plane. θ is called the multiplicity and ξ is called the orientation.

slide-19
SLIDE 19

Federer and Fleming Compactness Theorem

Theorem (Federer and Flemming Compactness Theorem)

If {Tj} ⊂ Dk(U) is a sequence of integral currents with sup(MW (Tj) + MW (∂Tj)) < ∞ for all W ⊂⊂ U, then {Tj} is sequentially compact.

slide-20
SLIDE 20

Plateau Problem

Given an integral k-current S and ∂S = ∅, is there an integral k + 1-current T such that ∂T = S and M(T) is minimal.

slide-21
SLIDE 21

Mobius band

Figure: Not every minimal surface is orientable

slide-22
SLIDE 22

Rectifiable Varifolds

Let M be a countably k-rectifiable subset of Rn and θ be a locally Hk-integrable function on M. The rectifiable n-varifold v(M, θ) is the set of all equivalence classes of (M, θ), where (M, θ) ≡ (N, φ) if Hk((M\N) ∪ (N\M)) = 0 and φ = θ Hk-a.e on M ∩ N. θ is called the multiplicity function. If it is integer valued we will call V an integral varifold.

slide-23
SLIDE 23

Rectifiable Varifolds

We associate to a varifold V = v(M, θ) the Radon measure µV such that for any Hk measurable set A, µV (A) =

  • A∩M

θdHk. The mass of a varifold V is M(V ) = µV (Rn). We say that Vk → V if µVk → µV .

slide-24
SLIDE 24

General Varifolds

We define the Grassmannian, G(k, n) to be the collection of all k-dimensional subsets of Rn. Given some A ⊂ Rn, we define Gk(A) = A × G(n, k), with the product metric.

Definition (General Varifold)

An k-varifold is a Radon measure on Gk(Rn).

slide-25
SLIDE 25

General Varifolds

Given a k-varifold V on GkU, there is an associated Radon measure µV on U (weight of V) defined for A ⊂ U as µV (A) = V (π−1(A)). The mass of the varifold is M(V ) = µV (U) = V (Gk(U)).

slide-26
SLIDE 26

First Variation

δV (X) = d dt M(φt#(V Gk(K)))|t=0 =

  • Gk(U)

divSXdV (x, S). (1) V is stationary if δV (X) = 0 for all X : U → Rn continuous and compactly supported.

slide-27
SLIDE 27

Monotonicity

Theorem (Monotonicity)

Suppose V is stationary in U, then r−nµV (Br(x)) is increasing for 0 < r < dist(x, ∂U).

slide-28
SLIDE 28

Density

Definition (Density)

Let V be stationary in U. The density for x ∈ U is defined to be Θk(µV , x) = ω−1

k r−kµV (Br(x))

− ω−1

k

  • Gk(Br(x))

r−n−2|pS⊥(y − x)|2dV (y, S), for 0 < r < dist(x, ∂U).

slide-29
SLIDE 29

Rectifiability

Theorem (Rectifiability)

Suppose V has locally bounded first variation in U and that Θk(µV , x) > 0 for µV -a.e. x ∈ U. Then V is a k-rectifiable varifold.

slide-30
SLIDE 30

Compactness

Theorem (Compactness for Rectifiable Varifolds)

Suppose {Vj}j∈N is a sequence of rectifiable varifolds that have bounded first variation in U, Θk(µVj, x) ≥ 1 in U and that sup(µVj(U) + δVj(U)) < ∞. Then {Vj}j∈N is sequentially compact and Θk(µV , x) ≥ 1.