Introduction to nonlinear geometric PDEs
Thomas Marquardt
January 16, 2014
ETH Zurich Department of Mathematics
Contents
I. Introduction and review of useful material 1
1. Introduction 3 1.1. Scope of the lecture ...... 3 1.2. Accompanying books ...... 4 1.3. A historic survey ...... 5
2. Review: Differential geometry 7 2.1. Hypersurfaces in Rn ...... 7 2.2. Isometric immersions ...... 11 2.3. First variation of area ...... 12
3. Review: Linear PDEs of second order 15 3.1. Elliptic PDEs in H¨older spaces ...... 15 3.2. Elliptic PDEs in Sobolev spaces ...... 18 3.3. Parabolic PDEs in H¨older spaces ...... 20
II. Nonlinear elliptic PDEs of second order 25
4. General theory for quasilinear problems 27 4.1. Fixed point theorems: From Brouwer to Leray-Schauder ...... 27 4.2. Reduction to a priori estimates in the C1,β-norm ...... 29 4.3. Reduction to a priori estimates in the C1-norm ...... 30
5. The prescribed mean curvature problem 33 5.1. C0-estimate ...... 33 5.2. Interior gradient estimate ...... 37 5.3. Boundary gradient estimate ...... 40 5.4. Existence and uniqueness theorem ...... 45
6. General theory for fully nonlinear problems 49 6.1. Fully nonlinear Dirichlet problems ...... 50 6.2. Fully nonlinear oblique derivative problems ...... 52
7. The capillary surface problem 57 7.1. C0-estimate ...... 58 7.2. Global gradient estimate ...... 60 7.3. Existence and uniqueness theorem ...... 66 III. Geometric evolution equations 67
8. Classical solutions of MCF and IMCF 69 8.1. Short-time existence ...... 69 8.2. Evolving graphs under mean curvature flow ...... 74 8.3. A Neumann problem for inverse mean curvature flow ...... 79
9. Outlook: Level set flow and weak solutions of (I)MCF 93 9.1. Derivation of the level set problem ...... 93 9.2. Solving the level set problem ...... 94
Bibliography 95 Preface
These notes are the basis for an introductory lecture about geometric PDEs at ETH Zurich in the spring term 2013. The course is based on my diploma thesis about pre- scribed mean curvature problems, my PhD thesis about inverse mean curvature flow and many inspiring lectures by my former thesis advisor Gerhard Huisken.
Further information about the lecture can be found on the course web page:
www.math.ethz.ch/education/bachelor/lectures/hs2013/math/PDEs
If you have questions or comments please feel free to contact me:
I want to thank Malek Alouini, Andreas Leiser and Mario Schulz for very valuable sug- gestions.
i
Part I.
Introduction and review of useful material
1
1. Introduction
1.1. Scope of the lecture
Geometric analysis is a field that has considerably grown over the last decades. Its goal is to answer questions that arise in geometry, topology, physics and many other sciences (e.g. in image processing) with the help of analytic tools. Usually the first task is to model the problem in terms of a (system of) PDE(s)1. Then existence and uniqueness is investigated using tools from PDE theory and/or the calculus of variations as well as functional analytic tools. Most of the time the geometric objects under consideration are not at all smooth. This requires the language of geometric measure theory to treat those problems. The aim of the course is to give an introduction to the field of nonlinear geometric PDEs by discussing two typical classes of PDEs. For the first part of the course we will deal with nonlinear elliptic problems. In particular, we will look at the Dirichlet problem of prescribed mean curvature and the corresponding Neumann problem of capillary surfaces. In the second part we will investigate nonlinear parabolic PDEs. As an example we will discuss the evolution of surfaces under inverse mean curvature flow. We will prove short- time existence as well as convergence results and introduce the notion of weak solutions.
Prescribing the scalar mean curvature H of a hypersurface M n ⊂ Rn+1 which is given as the graph of a scalar function u :Ω ⊂ Rn → R leads to the following equation2:
! Du div = H(·, u, Du) in Ω. (1.1) p1 + |Du|2
Together with a Dirichlet boundary condition u = φ on ∂Ω equation (1.1) is called the prescribed mean curvature problem. If we consider Neumann boundary conditions, i.e. if we prescribe the boundary contact angle the problem goes under the name capillary surface equation. During the first part of the course we will discuss existence and unique- ness of solutions to those problems. Note that if the denominator in (1.1) is replaced by one we obtain the Laplace operator. We will use the knowledge about linear second order elliptic PDEs together with a fixed point argument (or the method of continuity) and a priori estimates to prove existence for the corresponding nonlinear problems.
In the same way as the prescribed mean curvature equation resembles the Poisson equation, the evolution equation for the deformation of a hypersurface M n ⊂ N n+1 in time will resemble the heat equation. In our discussion we will focus on a deformation of hypersurfaces M n along their inverse mean curvature. In terms of the embedding
1This step is not at all unique. People have come up with totally different successful models to answer exactly the same questions. 2We will discuss this in more detail in Section 2.1
3 4 1. Introduction
F : M n × [0,T ] → N n+1 the equation reads
∂F 1 = ν F : M n × [0,T ) → (N n+1, g). (1.2) ∂t H The similarity to the heat equation will become clear during the second part of the course. The investigation of (1.2) will lead us to the topic of nonlinear parabolic PDEs. We will analyze their well-posedness (i.e. short-time existence) as well as their long-time behavior. Finally we will also discuss the construction of weak solutions via the level set method. It turns out this procedure brings us back to a degenerate version of (1.1).
1.2. Accompanying books
The following books contain subjects which are relevant for the topics we will discuss during the semester. We will not follow a particular book. However, for the first part of the course the book by Gilbarg and Trudinger will be closest to the lecture notes. For the second part the book by Gerhardt might be the most relevant one.
Overview about the field of PDEs:
• Evans [17]
Elliptic PDEs of second order:
• Gilbarg, Trudinger [23]
• Ladyˇzenskaja, Ural’ceva [37]
Parabolic PDEs of second order:
• Lieberman [40]
• Ladyˇzenskaja, Solonnikov, Ural’ceva [38]
Maximum principle:
• Protter, Weinberger [53]
Minimal surfaces:
• Giusti [24]
• Dierkes, Hildebrand, Sauvigny [10]
Mean curvature flow and related flows:
• Gerhardt [22]
• Ecker [11]
• Ritor´e,Sinestrari [54]
• Mantegazza, C. [41] 1.3. A historic survey 5
1.3. A historic survey
The field of geometric analysis is becoming more and more active during the last years. To give you an idea about the developments I made a brief survey in form of important results over the last 80 years. Note that the books and articles I cited here are not always written by the people who proved the result. If available I chose review articles which give an easy introduction into the topic.
1930: The Plateau problem Solved independently by Douglas and Rad´oin 1930:
• Dierkes, Hildebrand, Sauvigny [10]
1979: The positive mass theorem Proved by Schoen and Yau:
• Schoen (in Proc. of the Clay summer school 2001) [30]
1984: The Yamabe problem Partial results by Trudinger, Aubin and others, finally solved by Schoen:
• Lee, Parker [39]
• Struwe [61]
• B¨ar[3]
1999: The Penrose inequality Riemannian version proved by Huisken and Ilmanen and in a bit more general version two years later by Bray. The full Penrose inequality is still an open problem:
• Bray [5]
2002: The double bubble conjecture Proved by Hutchings, Morgan, Ritor´eand Ros.
• Morgan [47]
2003: The Poincar´e conjecture Proved by Perelman based on Hamilton’s work on the Ricci flow:
• Ecker [12]
• Morgan, Tian [48]
2007: The differentiable sphere theorem Proved by Brendle and Schoen:
• Brendle [6]
2012: The Lawson conjecture Proved by Brendle:
• Brendle [7] 6 1. Introduction
2012: The Wilmore conjecture Proved by Marques and Neves:
• Marques, Neves [45]
Of course the list can not be complete. But if you have the feeling I missed something very important please let me know. 2. Review: Differential geometry
We will always work with orientable hypersurfaces which are either immersed or embedded in a Riemannian ambient manifold. For most of the things we will use the same notation as in the differential geometry class of Michael Eichmair [14].
2.1. Hypersurfaces in Rn
For the discussion of the prescribed mean curvature problem and the capillary surface problem it will be sufficient to deal with embedded, graphical hypersurfaces in Rn+1: Let us consider a simple submanifold M = φ(U) of Rn+1 where U ⊂ Rn is a chart 1 S domain . The tangent space of M is defined as TM := p∈M TpM where TpM is the tangent space of M at p = φ(x):
n Rno TpM := φ∗(TxU) := φ(x), Dφ xv (x, v) ∈ TxU := U × .
S S ⊥ The normal space is defined as NM := p∈M NpM := p∈M (TpM) . In the case of an orientable hypersurface the normal space can be generated from a single normal vector ν.
The set of smooth tangent fields and normal fields are denoted by
X(M) := Γ(TM → M) := {X : M → TM | X smooth, πTM ◦ X = idM } ,
Γ(NM → M) := {η : M → NM | η smooth, πNM ◦ η = idM } where πTM : TM → M and πNM : NM → M are the base point projections. A basis of Γ(TM) is given by ∂ −1 ∂φ := φ∗ ◦ (x, ei) ◦ φ = p, . ∂xi p ∂xi φ−1(p) The metric (or first fundamental form) of M is the symmetric, positive definite map
g : X(M) × X(M) → C∞(M):(X,Y ) 7→ g(X,Y ) := hX,Y i.
For every p in M the map gp : TpM × TpM → R is the restriction of the Euclidean inner product to TpM. The matrix representation [g] of g at p = φ(x) with respect to the basis D ∂ ∂ E mentioned above has the coefficients gij := ∂xi , ∂xj . The coefficients of the inverse matrix are denoted by gij.
The second fundamental form is the symmetric map
⊥ A : X(M) × X(M) → Γ(NM → M):(X,Y ) 7→ A(X,Y ) := − (DX Y ) .
1Later we will use the word domain instead of chart domain and denote it by Ω instead of U.
7 8 2. Review: Differential geometry where D is the (covariant) derivative on T Rn+1. We obtain2
A(X,Y ) = −hDX Y, νiν =: h(X,Y )ν
The coefficients of the matrix [h] with respect to the basis mentioned above are hij = ∂ −1 −D ∂ ∂xj , ν . The Eigenvalues of [g] [h] are called principle curvatures. Their sum ∂xi is called mean curvature and their product is called Gaussian curvature.
Exercise I.1 (Geometric meaning of the principal curvatures). Let M be a smooth surface in R3 with unit normal ν. Let ε > 0 and γ :(−ε, ε) → M be a smooth curve such that γ(0) = p and γ0(0) = v with |v| = 1. We define the curvature of M at p in direction v as
00 kp : TpM → R : v 7→ kp(v) := hγ (0), νi.
Note that by Meusnier’s theorem kp is well defined. Answer the following questions
(i) Why does it make sense to call kp the curvature of M at p in dir. v?
(ii) What is the relation between kp and h?
(iii) What are the critical values of kp in terms of h and g? The tangential covariant derivative on TM is defined as
> ∇ : X(M) × X(M) → X(M):(X,Y ) 7→ ∇X Y := (DX Y )
> where (∇X Y )(p) := ∇X(p)Y = (DX(p)Y ) . For computations the Leibniz rule and metric compatibility are important tools
∇X fY = (Xf)Y + f∇X Y, Zg(X,Y ) = g(∇Z X,Y ) + g(X, ∇Z Y ) for all X,Y,Z ∈ Γ(TM) and f ∈ C∞(M). Recall that
∂ ∂(f ◦ φ) Xf = Xi f := Xi . ∂xi ∂xi φ−1
∂ where ∂xi is used as a symbol for the basis element as well as for the actual partial deriva- tive in Rn.
k ∂ k ∂ The functions Γij such that ∇ ∂ ∂xj = Γij ∂xk are called Christoffel symbols. It is not ∂xi difficult to verify that 1 ∂g ∂g ∂g Γk = gkl il + lj − ij . ij 2 ∂xj ∂xi ∂xl
i ∂ j ∂ It follows that for X = X ∂xi ,Y = Y ∂xj we have ! ∂Y k ∂ ∇ Y = Xi + Y jΓk . X ∂xi ij ∂xk
2Note that the sign of h depends on the choice of normal ν. 2.1. Hypersurfaces in Rn 9
Based on the rules that taking covariant derivatives commutes with contractions and that ∇X (S ⊗T ) = (∇X S)⊗T +S ⊗(∇X T ) we obtain connections on the associated vector bundles. For example, for differential one forms we get (∇X ω)Y = Xω(Y ) − ω(∇X Y ). i ∂ k In coordinates, i.e. X = X ∂xi , ω = ωkdx this yields ∂ω ∇ ω = Xi k − ω Γj dxk. X ∂xi j ik The tangential gradient of a function f ∈ C∞(M) is defined to be the unique tangent 3 field gradM f such that g(gradM f, X) = df(X) = Xf for all X in Γ(TM). We obtain the formulae m ij ∂f ∂ X grad f = g = grad n+1 f − hgrad n+1 f, νi ν = (D f)τ M ∂xi ∂xj R R τi i i=1 where we used an orthonormal frame (τi) of TM in the last equality. The tangential divergence of a tangent field X can be defined via the contraction of ∇X. We can write n i ! m X ∂X i j X div X := + Γ X = divRn+1 X − hD X, νi = hD X, τ i M ∂xi ij ν τi i i=1 i=1 using once more an orthonormal frame of TM for the last expression. Note that the last two equalities also make sense for vector fields which are not necessarily tangential. ∂f i In the special case ω = df = ∂xi dx we obtain the Hessian of f: ! ∂2f ∂f (Hess f)(X,Y ) = (∇df)(X,Y ) = (∇ df)(Y ) = XiY j − Γk . M X ∂xi∂xj ∂xk ij
Finally, the Laplace Beltrami operator is defined as ∆M f := divM (gradM f). One can compute that ! ∂2f ∂f ∆ f = gij − Γk = gij(Hess f) . M ∂xi∂xj ∂xk ij M ij Thus it is the contraction of the Hessian with respect to the metric. Exercise I.2 (Graphical hypersurfaces in Rn). Suppose that M is a hypersurface in Rn+1 which is given as a graph over a chart domain U ⊂ Rn: M = graph u. Verify the following formulae: DiuDju D uDku g = δ + D uD u, gij = δij − , Γk = ij , ij ij i j 1 + |Du|2 ij 1 + |Du|2
±1 ∓Diju ν = (−Du, 1) , hij = , p1 + |Du|2 p1 + |Du|2
i j ! ! ∓1 ij D uD u ∓Du H = δ − Diju = divRn = divRn+1 (ν), p1 + |Du|2 1 + |Du|2 p1 + |Du|2 ! ! Du 1 −Du H := −Hν = divRn . p1 + |Du|2 p1 + |Du|2 1
3Sometimes people just write ∇f. In the analytical literature such as [23] one also finds the symbol δf. 10 2. Review: Differential geometry
Remark 2.1 (Classical problems). A natural question is, weather for a given function H and given boundary values φ on ∂Ω there exists a graph of a function u :Ω → R which has mean curvature H and attains the boundary values φ, i.e. a solution to the prescribed mean curvature problem ! Du div p = H(·, u, Du) in Ω 1 + |Du|2 (2.1) u = φ on ∂Ω.
If at the boundary we prescribed the contact angle instead of the height we obtain the so called capillary surface problem ! Du div = H(·, u, Du) in Ω p1 + |Du|2 (2.2) Dγu = β on ∂Ω. p1 + |Du|2 where β is the cosine of the contact angle and γ is the outward unit normal to ∂Ω. One can regard various modifications of these problems. For example, one can replace the mean curvature by the Gaussian curvature, i.e.
2 det(D u) n+2 = K(·, u, Du) in Ω (1 + |Du|2) 2 (2.3) u = φ on ∂Ω. and similar for the Neumann problem. Furthermore, one can consider different ambient spaces, e.g. hyperbolic space or Minkowski space or general Riemannian manifolds. One way to find solutions to these static problems is to consider the corresponding evolution equation and to investigate what happens in the limit as t → ∞, e.g. ! ∂u Du − div p = H(·, u, Du) in Ω × (0,T ) ∂t 1 + |Du|2 u = φ on ∂Ω × (0,T ) (2.4) u = u0 on Ω × {0}.
The corresponding linear model problem would then be the heat equation. These parabolic problems are also interesting in its own as they might reveal topological information about the evolving surface. A totally different application would be to apply those flows to do a noise reduction in image processing.
Exercise I.3 (Explicit computation). Compute the principal curvatures, the mean curvature and the Gaussian curvature of the graph of the function u : B2(0) → R : (x, y) 7→ x2 − y2 at the origin.
n Exercise I.4 (Counter examples). (i) Let R > 0 and Ω = BR(0) ⊂ R . Is there always a function u :Ω → R with u = 0 on ∂Ω such that the graph of u has constant mean curvature H = c? 2.2. Isometric immersions 11
(ii) Let 0 < R1 < R2 < ∞ and Ω = BR2 (0) \ BR1 (0). Is there always a function u : Ω → R with u = 0 on {|x| = R2} and u = L > 0 on {|x| = R1} such that the graph of u has zero mean curvature? Hint: Try to find an explicit solution.
Exercise I.5 (Flat v.s. harmonic v.s. minimal). Let Ω = B2(0) \ B1(0) and the boundary conditions be u = 0 on {|x| = 2} and u = 1 on {|x| = 1}. Compare the surface area of the graphical minimal surface graph vm to that of the truncated cone, i.e. to graph vc where vc(x) := 2 − |x| as well as to graph vh where vh is the corresponding harmonic function, i.e. the function satisfying ∆vh = 0 in Ω together with the same boundary values.
2.2. Isometric immersions
Let (N, g) be a Riemannian manifold of dimension n. Let M be a smooth manifold of dimension m ≤ n and φ : M → N an immersion, i.e. a smooth map, such that for all p ∈ M the push-forward φ? : TpM → Tφ(p)N is injective (so that in charts we have rank[Dφ] = dim M everywhere). If additionally φ is a homeomorphism onto its image it is called an embedding. We define the pull-back metric on TM via
? gp(v, w) := (φ g)p (v, w) := gφ(p)(φ?v, φ?w) ∀ v, w ∈ TpM. With respect to that metric the map φ :(M, g) → (N, g) is an isometric immersion. We can also pull back the tangent bundle TN → N to obtain the bundle φ?TN → M where
? [ φ TN := {p} × Tφ(p)N p∈M with the tangent space and normal space as subspaces:
[ [ ⊥ TMφ := {p} × φ?(TpM),NMφ := {p} × (φ?(TpM)) . p∈M p∈M
Note that we can identify TM and TMφ via the isometry v 7→ (π(v), φ?v). Using the ⊥ > projections : TN → NMφ and : TN → TM we have the decomposition
> ⊥ V = φ? ◦ (V ) + V ∀ V ∈ TN. On (N, g) there exists a unique covariant derivative (also called connection) which is compatible with g and torsion free, i.e.
Zg(V,W ) = g(∇Z V,W ) + g(V, ∇Z W ), ∇V W − ∇W V = [V,W ]. This covariant derivative ∇ is called Levi-Civita connection of (N, g). It can be shown that the connection > φ ∇X Y := ∇X φ? ◦ Y X,Y ∈ X(M) is the Levi-Civita connection of (M, g). Here φ∇ is the pull-back connection, i.e. the unique connection φ∇ : X(M) × Γ(φ?TN → M) → Γ(φ?TN → M) which satisfies the naturality condition
φ ? ∇vφ η = ∇φ?vη ∀ v ∈ TM, ∀ η ∈ Γ(TN → N). 12 2. Review: Differential geometry
The normal part of the connection is called the second fundamental form:
⊥ φ A(X,Y ) := − ∇X φ? ◦ Y X,Y ∈ X(M).
φ Thus, we can write −A(X,Y ) = ∇X φ? ◦ Y − φ? ◦ ∇X Y . With respect to a basis we have ∂ α ∂ ∂ −A = φ∇ φ ◦ φ? − φ ◦ ∇ ij ∂ ? ∂xj ∂qα ? ∂ ∂xj ∂xi ∂xi α α ∂ ∂ ∂ ∂ φ ? ∂ k ∂ = φ? ◦ + φ? ◦ ∇ φ − Γijφ? ◦ ∂xi ∂xj ∂qα φ ∂xj ∂ ∂qα ∂xk ∂xi α α β ∂ ∂ ∂ ∂ ∂ φ γ ? ∂ k ∂ = φ? ◦ + φ? ◦ φ? ◦ Γαβφ − Γijφ? ◦ ∂xi ∂xj ∂qα φ ∂xj ∂xi ∂qγ ∂xk
2 α β ∂ φ ∂φ ∂φ φ γ ∂ k ∂φ = + Γαβ − Γij . ∂xi∂xj ∂xi ∂xj ∂pγ φ ∂xk
Note that the last equality is just symbolic unless N = Rn. In that case the ∂/∂xr can be read as partial derivatives and Γ ≡ 0.
2.3. First variation of area
Proposition 2.2 (First variation of area). Let M ⊂ Rn be a smooth m-dimensional submanifold. Let O ⊂ Rn be open such that M ∩ O 6= ∅. We consider the deformation of M under a family of diffeomorphisms
Φ:(−ε, ε) × O → O :(t, p) 7→ Φ(t, p) =: Φt(p) satisfying for some compact set K ⊂ O
Φ(0, ·) = Id in O, Φ(t, ·) = Id in O \ K.
Then for the first variation of area we obtain
d d
area(Φt(M)) = divM Φt dµ. dt t=0 ˆM dt t=0 For the prove we will use the following formula.
Exercise I.6 (Derivative of the determinant). Let ε > 0. Suppose that
1 n×n B ∈ C (t0 − ε, t0 + ε), R and that B(t0) is invertible. Show that
d d −1 det B(t) = det B(t0) tr B (t0) B(t) . dt t=t0 dt t=t0 2.3. First variation of area 13
Proof of Proposition 2.2. The area formula tells us that q > area(Φt(M)) = Jac((Φt)?) dµ = det([(Φt)?] [(Φt)?]) dµ. ˆM ˆM Therefore, it remains to compute the derivative of the Jacobian. To simplify our compu- tation we write
2 d Φt(p) = p + tX(p) + O(t ),X := Φt. dt t=0 For the push forward of Φt we obtain Rn 2 (Φt)? : TM → T : τk 7→ (Φt)?τk = τk + tDτk X + O(t ).
The matrix representation of (Φt)? with respect to an orthonormal basis (τi)1≤i≤m of TM n+1 and the standard basis (ej)1≤j≤n+1 of T R is given by j j 2 [(Φt)?]ij = τi + tDτi X + O(t ). Using the formula for the derivative of the determinant we can compute that q d > det([(Φt)?] [(Φt)?]) dt t=0 d q 2 = det(δij + thτi,Dτj Xi + thDτi X, τji + O(t )) dt t=0 1 = tr hτ ,D Xi + hD X, τ i 2 i τj τi j
= divM X which proves the result. Remark 2.3. To derive a formula for the second variation of area one can proceed in a similar way. For the details see [56], Chapter 2. d Corollary 2.4. Let us define X := Φt then dt t=0 d
area(Φt(M)) = divM X dµ =+ Hhν, Xi dµ + hµ, Xi dσ dt t=0 ˆM ˆM ˆ∂M where ν is the unit normal of M and µ is the outward unit conormal of ∂M, i.e. normal to ∂M, tangent to M and pointing away from M. Proof. Proposition 2.2 implies the first equality. To verify the second equality we choose an orthonormal frame (τi)i∈N of TM and compute
> ⊥ divM X dµ = divM X dµ + divM X dµ ˆM ˆM ˆM
D >E D ⊥ E = µ, X dσ + Dτi X , τi dµ ˆ∂M ˆM
D ⊥ E D ⊥ E = hµ, Xi dσ + τi X , τi dµ − (Dτi τi) ,X dµ ˆ∂M ˆM ˆM D E = hµ, Xi dσ − hDτi τi, νiν, X dµ ˆ∂M ˆM 14 2. Review: Differential geometry which is the desired result.
Remark 2.5 (Minimal surfaces). Observe that if X is normal to M then the first derivative of area vanishes exactly for surfaces of zero mean curvature. Even though H = 0 surfaces are just critical points of the area functional they are often called minimal surfaces. Try to picture what happens with M if X has a non-vanishing tangential component.
Remark 2.6 (Variational approach). Note that in the special case of graphical hy- persurfaces in Rn minimizing area means
q I(u) := 1 + |Du(x)|2dx → min . ˆΩ The corresponding Euler-Lagrange equation is exactly the minimal surface equation. Fur- thermore, the analogous linear problem is the minimization of the Dirichlet energy, i.e.
I(u) := |Du(x)|2dx → min ˆΩ whose Euler-Lagrange equations is ∆u = 0.
Remark 2.7 (Capillary surfaces in gravitational fields). If in addition to area (which up to a constant equals surface tension) we also take gravity into account as well as the adhesion forces at the boundary, physical considerations lead to minimizing
q I(u) := 1 + |Du(x)|2dx + κ u(x)2dx − βu(x)ds → min . ˆΩ ˆΩ ˆ∂Ω The corresponding elliptic problem is the capillary surface problem, where H(·, u, Du) = κu. 3. Review: Linear PDEs of second order
3.1. Elliptic PDEs in H¨olderspaces
Definition 3.1 (Linear elliptic operators). Let Ω ⊂ Rn be a bounded domain and u ∈ C2(Ω). Suppose that aij, bk, c ∈ C0(Ω) and that the matrix [aij] is symmetric. The differential operator L defined by
ij k Lu := a (x)Diju + b (x)Dku + c(x)u is called elliptic in Ω if the matrix [aij(x)] is positive definite for all x in Ω. In this case
ij n 0 < λ(x) ≤ a (x)ξiξj ≤ Λ(x) ∀ ξ ∈ S , ∀ x ∈ Ω (3.1) where λ(x) and Λ(x) are the smallest and largest Eigenvalues of [aij(x)]. Furthermore, L is called uniformly elliptic in Ω if there exist λmin and Λmax such that
ij 0 < λmin ≤ a (x)ξiξj ≤ Λmax < ∞ ∀ ξ ∈ S, ∀ x ∈ Ω. (3.2)
Theorem 3.2 (Maximum principle). Let L be uniformly elliptic in the bounded domain Ω and c ≤ 0. Let u ∈ C2(Ω) ∩ C0(Ω). If Lu ≥ f then
! |f −| sup u ≤ sup u+ + c sup . Ω ∂Ω Ω λ
If Lu = f then
|f| sup |u| ≤ sup |u| + c sup . Ω ∂Ω Ω λ
In both cases c = c(diam Ω, sup |b|/λ).
Proof. See [23], Theorem 3.7.
The following comparison principle is a useful consequence.
Corollary 3.3. Let Ω and L be as above. Suppose that u, v ∈ C0(Ω) ∩ C2(Ω) satisfy Lu ≥ Lv in Ω and u ≤ v on ∂Ω. Then u ≤ v in Ω.
Proof. Apply Theorem 3.2 to w := u − v.
Before we can state the existence theorem for linear elliptic equations of second order we want to recall the definition of H¨olderspaces.
15 16 3. Review: Linear PDEs of second order
Definition 3.4 (H¨older spaces). Let Ω ⊂ Rn be a bounded domain and 0 < α < 1. A function f :Ω → R is called α-H¨older continuous in x0 if
f(x) − f(x0) [f] := sup < ∞. α,{x0} α x∈Ω |x − x0| We say that f is uniformly α-H¨older continuous in Ω if
f(x) − f(y) [f]α,Ω := sup α < ∞. x,y∈Ω |x − y| x6=y The spaces k,α k β n C (Ω) := f ∈ C (Ω) D f unif. α-H¨oldercont. in Ω, ∀ β ∈ N , |β| = k equipped with the norm
X β kfkCk,α(Ω) := kfkCk(Ω) + D f α,Ω |β|=k are Banach spaces. If α = 1 we say Lipschitz instead of 1-H¨older.
n Lemma 3.5. Let Ω ⊂ R be an open, bounded Lipschitz domain. Let k1, k2 ≥ 0 and k ,α k ,α 0 ≤ α1, α2 ≤ 1. If k1 + α1 > k2 + α2 then the inclusion of C 1 1 (Ω) into C 2 2 (Ω) is compact. Proof. See [1], Section 8.6.
Now we are ready to quote a classical existence theorem for linear Dirichlet problems. Theorem 3.6 (Existence for the linear Dirichlet problem). Let Ω ⊂ Rn be a bounded C2,α-domain. Let L be an uniformly elliptic operator with coefficients in C0,α(Ω) and c ≤ 0. Furhtermore, assume that f ∈ C0,α(Ω) and φ ∈ C2,α(Ω). Then the Dirichlet problem
( Lu = f in Ω u = φ on ∂Ω has a unique solution u ∈ C2,α(Ω) satisfying kukC2,α(Ω) ≤ C kukC0(Ω) + kφkC2,α(Ω) + kfkC0,α(Ω)
ij i where C = C n, α, Ω, λmin, ka kC0,α(Ω), kb kC0,α(Ω), kckC0,α(Ω) . Proof. See [23], Theorem 6.6 and Theorem 6.14.
Remark 3.7. The condition c ≤ 0 is only needed for the existence and uniqueness statement but not for the estimate. In the case that c ≤ 0 is not satisfied existence and uniqueness still hold as long as the homogeneous problem Lu = 0 in Ω, u = 0 on ∂Ω has only the zero solution. That is (one part of) the Fredholm alternative (see [23], Theorem 6.15). 3.1. Elliptic PDEs in H¨olderspaces 17
If the data of the problem are more regular then also the solution possesses more regularity. Theorem 3.8 (Interior regularity). Let k ∈ N and Ω ⊂ Rn a bounded domain. Let L be a linear, uniformly elliptic operator with coefficients aij, bi, c ∈ Ck,α(Ω). Furthermore, assume that f ∈ Ck,α(Ω) and that u ∈ C2(Ω) satisfies Lu = f. Then u ∈ Ck+2,α(Ω). Proof. See [23], Theorem 6.17.
Theorem 3.9 (Global regularity). Let k ∈ N and Ω ⊂ Rn a bounded Ck+2,α-domain. Let L be a linear, uniformly elliptic operator with coefficients aij, bi, c ∈ Ck,α(Ω). Fur- thermore, assume that f ∈ Ck,α(Ω) and φ ∈ Ck+2,α(Ω). If u ∈ C0(Ω) ∩ C2(Ω) satisfies Lu = f in Ω and u = φ on ∂Ω. Then u ∈ Ck+2,α(Ω). Proof. See [23], Theorem 6.19.
Finally, we also want to mention the corresponding result for the oblique derivative problem. It includes the Neumann problem as a particular case. Theorem 3.10 (Existence for the linear oblique Derivative problem). Let Ω ⊂ Rn be a bounded C2,α-domain. Let L be a linear, uniformly elliptic operator with coefficients in C0,α(Ω) and c ≤ 0. Furhtermore, assume that f ∈ C0,α(Ω) and γ, β, φ ∈ C1,α(Ω). If γhβ, νi > 0 (ν exterior unit normal of ∂Ω) or hβ, νi > 0, γ ≥ 0 and either c 6= 0 or γ 6= 0, then the oblique derivative problem
( Lu = f in Ω
γu + Dβu = φ on ∂Ω has a unique solution u ∈ C2,α(Ω) satisfying kukC2,α(Ω) ≤ C kukC0(Ω) + kφkC1,α(Ω) + kfkC0,α(Ω) (3.3) where
ij i C = C n, α, Ω, λmin, ka kC0,α(Ω), kb kC0,α(Ω), kckC0,α(Ω), kγkC1,α(Ω), kβikC1,α(Ω), hβ, νi . Proof. See [23], Theorem 6.30, Theorem 6.31 and the following remarks about the Fred- holm alternative.
Exercise I.7 (Understanding H¨older spaces). (i) Can you find a function which is in C0,α(Ω) but not in C0,α+ε(Ω) for ε > 0? Can you find a function that is in C1,1(Ω) but not in C2(Ω)? (ii) Can you find a domain Ω ⊂ R2 and a function u ∈ C1(Ω) which is not in C0,3/4(Ω)? (iii) Why do we need to work with H¨olderspaces Ck,α(Ω) instead of the easier Ck(Ω) spaces? 18 3. Review: Linear PDEs of second order
3.2. Elliptic PDEs in Sobolev spaces
Definition 3.11 (Sobolev spaces). Let Ω ⊂ Rn be a bounded domain and α ∈ Nn a 1 1 multi index. A function v ∈ Lloc(Ω) is called weak α-derivative of u ∈ Lloc(Ω) if
|α| α |α| ψv dx = (−1) uD ψ dx ∀ ψ ∈ C0 (Ω). ˆΩ ˆΩ We write v = Dαu weakly. For k ∈ N and 1 ≤ p ≤ ∞ the sets
k,p p α p W (Ω) := u ∈ L (Ω) D u ∈ L (Ω) for |α| ≤ k are called Sobolev spaces. Equipped with the norms
X α kukW k,p(Ω) := kD ukLp(Ω), 1 ≤ p < ∞ |α|≤k
X α kukW k,∞(Ω) := kD ukL∞(Ω), p = ∞ |α|≤k they are Banach spaces. The spaces Hk(Ω) := W k,2(Ω) are even Hilbert spaces. Further- more, we extend the notion of functions having zero boundary values by defining
k·k k,p k W k,p(Ω) W0 (Ω) := C0 (Ω) . To get a better understanding of these spaces let us have a look at the following theorem. Theorem 3.12 (Trace operator). Let 1 ≤ p < ∞ and Ω ⊂ Rn be a bound domain with C1-boundary. There exists a bounded linear operator T : W 1,p(Ω) → Lp(∂Ω) called the trace operator defined via
0 1,p T u = u|∂Ω for u ∈ C (Ω) ∩ W (Ω).
1,p 1,p If u ∈ W (Ω) then u ∈ W0 (Ω) if and only if T u|∂Ω = 0. Proof. See [17], Section 5.5, Theorem 1 and Theorem 2.
Lemma 3.13. Let Ω ⊂ Rn be a bounded domain and u ∈ W 1,p(Ω). Then
u+ := max{u, 0}, u− := min{u, 0}, |u| ∈ W 1,p(Ω).
Furthermore, Du = 0 on any set where u is constant. Proof. See [23], Lemma 7.6. and Lemma 7.7.
Theorem 3.14 (Poincar´e type inequalities). Let Ω ⊂ Rn be a bounded domain. Let 1,p 1 ≤ p < n and u ∈ W0 (Ω). Then np kuk q ≤ C(p, q, n, Ω)kDuk p ∀ q ∈ 1, . L (Ω) L (Ω) n − p
If p = 1 we can choose the constant to be c = (nω1/n)−1. Another special case is
|Ω|1/n kukLp(Ω) ≤ kDukLp(Ω). ωn 3.2. Elliptic PDEs in Sobolev spaces 19
Proof. See [17], Section 5.6, Theorem 3 and [23], inequality (7.44).
An important relation between the H¨olderspaces and the Sobolev spaces is given by (one of) the Sobolev embedding theorem(s). Theorem 3.15 (Sobolev embedding theorem). Let Ω ⊂ Rn be a bounded domain with Lipschitz boundary. Let p ∈ [1, ∞) and u ∈ W k,p(Ω). If n 0 ≤ m < k − < m + 1 p
m,k− n −m then u ∈ C p (Ω). For smaller H¨oldercoefficients the embedding is even compact. k,p Note that the regularity assumption for ∂Ω can be dropped if we consider the space W0 (Ω) instead. Proof. See for example [23], Chapter 7, Theorem 7.26 or [17], Chapter 5, Section 6, Theorem 6.
Theorem 3.16 (Morrey’s estimate). Let u ∈ W 1,1(Ω). If there exists some α ∈ (0, 1) and K > 0 such that
n−1+α |Du|dx ≤ KR , ∀ BR ⊂ Ω ˆBR 0,α Then u ∈ C (Ω) with [Du]α,Ω ≤ c(n, α, K). If Ω = Ω˜ ∩ {xn > 0} for some domain n 0,α Ω˜ ⊂ R and the above inequality holds for all BR ⊂ Ω˜ then u ∈ C (Ω ∩ Ω)˜ . Proof. See [23], Theorem 7.19.
Remark 3.17. The last part of the theorem will be particularly useful for local estimates near a flattened boundary. Next we want to recall the H¨oldercontinuity of weak solutions. Definition 3.18 (Weak solutions). Let Ω ⊂ Rn be a bounded domain. We consider the linear operator
ij k Lu := Di(a Dju) + b Dku + cu with coefficients aij, bk, c ∈ C0(Ω) and such that [aij] satisfies (3.2). Let g, f i ∈ L1(Ω). If u ∈ W 1,2(Ω) satisfies
h ij i k i 1 a Dju − f Djξ − b Dku + cu − g ξ = 0 ∀ ξ ∈ C0 (Ω) ˆΩ i we say that u is a weak solution of Lu = g + Dif . Remark 3.19. Note that weak solutions can be defined for more general operators and under weaker conditions on the coefficients. Note also that the integral equality remains 1,2 true for test functions ξ ∈ W0 (Ω). Theorem 3.20 (Maximum principle). Let Ω ⊂ Rn be a bounded domain and L as in Definition 3.18. Let u ∈ C0(Ω) ∩ W 1,2(Ω) satisfy Lu ≥ 0 in a weak sense. If c ≤ 0 then
sup u ≤ sup u+. Ω ∂Ω 20 3. Review: Linear PDEs of second order
Proof. See [23], Theorem 8.1.
Theorem 3.21 (Interior H¨older continuity). Let Ω ⊂ Rn be a bounded domain. Let i q 1,2 ij i f ∈ L (Ω) for some q > n. Let u ∈ W (Ω) be a weak solutionf of Di(a Dju) = Dif in Ω. Then
−1 kukC0,α(Ω0) ≤ c(n, q, d, Λmax/λmin) kukL2(Ω) + λminkfkLq(Ω)
0 0 where Ω ⊂⊂ Ω, d = dist(Ω , ∂Ω) and α = α(n, Λmax/λmin).
Proof. See [23], Theorem 8.24.
Theorem 3.22 (Boundary H¨older continuity). Let Ω ⊂ Rn be a domain satisfying a uniform exterior cone condition with cones V on a boundary portion T . Let f i ∈ Lq(Ω) 1,2 ij i for some q > n. Let u ∈ W (Ω) be a weak solution of Di(a Dju) = Dif in Ω. If there exits K, α0 > 0 such that
α0 osc u ≤ KR ∀ x0 ∈ T, R > 0. ∂Ω∩BR(x0)
Then
−1 kukC0,α(Ω0) ≤ c(n, q, d, α0,V, Λmax/λmin) kukC0(Ω) + K + λminkfkLq(Ω) .
0 0 Here Ω ⊂⊂ Ω ∪ T , d = dist(Ω , ∂Ω \ T ) and α = α(n, q, α0,V, Λmax/λmin).
Proof. See [23], Theorem 8.29.
For the sake of completeness let us also mention the existence result in the weak setting.
Theorem 3.23 (Weak existence). Suppose that Ω ⊂ Rn satisfies an exterior cone condition on all points of ∂Ω. Let φ ∈ C0(∂Ω) and f i ∈ Lq(Ω) for some q > n.Then there 1,2 ij i exists a unique weak solution u ∈ Wloc (Ω) of Di(a Dju) = Dif satisfying u = φ on ∂Ω. Proof. See [23], Theorem 8.30.
3.3. Parabolic PDEs in H¨olderspaces
Definition 3.24 (Parabolic H¨older spaces). Let Ω ⊂ Rn be a bounded domain and T > 0. We set QT := Ω × (0,T ), ST := ∂Ω × (0,T ) and start with a definition of the parabolic analogue of the spaces Ck(Ω):
k k X X γ k;b 2 c R t γx C (QT ) := f : QT → kuk k;b k c := sup |Dt Dx f| < ∞ C 2 (Q ) T QT i=0 |γx|+2|γt|=i
Let us denote the H¨oldercoefficients of a function f : QT → R by |f(y, t) − f(x, t)| [f]x,α,QT := sup α (x,t),(y,t)∈QT |y − x| x6=y 3.3. Parabolic PDEs in H¨olderspaces 21 and |f(x, t) − f(x, s)| [f]t,α,QT := sup α . (x,s),(x,t)∈QT |t − s| s6=t The parabolic H¨olderspaces are defined as
k α k k,α;b 2 c, 2 k,b 2 c C (QT ) := u ∈ C (QT ) kuk k,α;b k c, α < ∞ C 2 2 (QT ) with
kuk k,α;b k c, α := kuk k;b k c C 2 2 (QT ) C 2 (QT )
X γ X γ + [D t Dγx u] + [D t Dγx u] t x x,α,QT t x t,β,QT 2|γt|+|γx|=k 0 Exercise I.8. What are the components of the k · k 2,α;1, α -norm? C 2 (QT ) Definition 3.25 (Linear uniformly parabolic differential operators). Let aij, bk, c ∈ 0 ij 2,1 C (QT ) and suppose that [a ] is symmetric. Let u ∈ C (QT ). The operator ∂/∂t − L defined by ∂u ∂u − Lu := − aijD u + bkD u + cu ∂t ∂t ij k is called parabolic. It is called uniformly parabolic in QT if additionally ij 0 < λmin ≤ a (x, t)ξiξj ≤ Λmax < ∞ holds for all (x, t) ∈ QT and all ξ ∈ S. Theorem 3.26 (Existence for the parabolic Dirichlet boundary value problem). Let Ω ⊂ Rn be a bounded domain with C2,α-boundary and T > 0. Let ∂/∂t − L be ij k 0,α,0, α 2,α,1, α uniformly parabolic in QT . Suppose that a , b , c, f ∈ C 2 (QT ), φ ∈ C 2 (ST ) and 2,α u0 ∈ C (Ω). If the compatibility conditions ∂ ij k φ(·, 0) = u0, φ = a (·, 0)Diju0 − b (·, 0)Dku0 − c(·, 0)u0 + f(·, 0) ∂t t=0 are satisfied on ∂Ω. Then the problem ∂u − Lu = f in Ω × (0,T ) ∂t u = φ on ∂Ω × (0,T ) u(·, 0) = u0 on Ω × {0} 2,α,1, α has a unique solution u ∈ C 2 (QT ) which satisfies kuk 2,α;1, α ≤ C kfk 0,α;0, α + kφk 2,α;1, α + ku0k 2,α . C 2 (QT ) C 2 (QT ) C 2 (ST ) C (Ω) If the coefficients and right hand sides are more regular the solution will be more regular too. However, to obtain more regular solutions up to t = 0 one also has to impose higher order compatibility conditions. 22 3. Review: Linear PDEs of second order Proof. See [38], Chapter 5, Theorem 5.2. Theorem 3.27 (Existence for the parabolic Neumann boundary value prob- lem). Let Ω ⊂ Rn be a bounded domain with C2,α-boundary and T > 0. Let ∂/∂t − ij k 0,α,0, α k L be uniformly parabolic in QT . Suppose that a , b , c, f ∈ C 2 (QT ), β , γ, φ ∈ 1,α;0, α 2,α C 2 (ST ) and u0 ∈ C (Ω). If the compatibility condition k φ(·, 0) = β (·, 0)Dku0 + γ(·, 0)u0 on ∂Ω is satisfied and the transversality condition holds, i.e. for the outward unit normal µ of ∂Ω × (0,T ) we have hβ, µi > 0 on ∂Ω × (0,T ). Then the problem ∂u − Lu = f in Ω × (0,T ) ∂t k β Dku + γu = φ on ∂Ω × (0,T ) (3.4) u(·, 0) = u0 on Ω × {0} 2,α,1, α has a unique solution u ∈ C 2 (QT ) which satisfies kuk 2,α;1, α ≤ C kfk 0,α;0, α + kφk 1,α;0, α + ku0k 2,α . C 2 (QT ) C 2 (QT ) C 2 (ST ) C (Ω) If the coefficients and right hand sides are more regular the solution will be more regular too. However, to obtain more regular solutions up to t = 0 one also has to impose higher order compatibility conditions. Proof. See [38], Chapter 5, Theorem 5.3. Remark 3.28 (Differentiable functions defined on hypersurfaces/ boundaries). As in the differential geometry section we say that a function is in Ck(S) for some Ck- hypersurface S ⊂ Rn if the composition with a chart has that regularity as a map from an open subset of Rn into R. Let k ≥ 0 and a domain Ω ⊂ Rn with Ck-boundary. If k k k φ ∈ C (Ω) then φ ∂Ω defines a function in C (∂Ω). Conversely, if φ ∈ C (∂Ω) then there exists φ ∈ Ck(Ω) such that both functions agree on the boundary and have equivalent norms. The same result carries over to the parabolic setting and to H¨oldernorms where the norms are compute in charts and the global norm is defined via a partition of unity. That is how we understand terms like kφk 2,α;1, α . C 2 (ST ) Remark 3.29 (PDEs on manifolds). In the situation where Ω ⊂ Rn is replaced by a smooth Riemannian manifold (M n, g) we have to modify the definition of the norms. They can defined locally via charts and globally via a partition of unity. Note that in this context the norms depend on the choice of atlas. Once this is done, one obtains the same existence results for the linear Dirichlet and Neumann problem as above. Furthermore, one could allow a time dependent metric g(·, t). The most important tool for second order parabolic equations is the maximum principle. Before we mention it we define sub- and supersolutions. 3.3. Parabolic PDEs in H¨olderspaces 23 Definition 3.30 (Super- and subsolutions). Let v+, v− ∈ C2,1(Ω × (0,T )) ∩ C0(Ω × [0,T ]). We say that v+ is a supersolutions of (3.4) if it satisfies + ∂v + − Lv ≥ f1 in Ω × (0,T ) ∂t + k + + Nv := β D v + γv ≥ f2 on ∂Ω × (0,T ) k + v ≥ u0 on Ω × {0}. The function v− is called subsolution if the opposite inequalities hold. Now we can state the version of the maximum principles which we use in this work. Theorem 3.31 (Parabolic maximum principle). Let u ∈ C0(Ω × [0,T ]) ∩ C2,1(Ω × (0,T )) be a solution of (3.4). Assume that L and N have bounded coefficients, that ∂/∂t−L is uniformly parabolic and that the Neumann condition is oblique. If v+ and v− are super- − + and subsolutions of (3.4) then v ≤ u ≤ v in QT . Proof. Note that for w := v+ − u and w := u − v− we have ∂w/∂t − Lw ≥ 0, Nw ≥ 0 and w( ., 0) ≥ 0. So we can reduce the proof to the case of the upper bound for f1 = 0, f2 = 0 and u0 = 0. This proof is contained in [53] Chapter 3, Section 3, Theorem 5,6 and 7. Furthermore Stahl proved in [57] the generalization which in particular allows for the more general operator N which occurs here. + Corollary 3.32. If f1 ≡ 0 and f2 ≡ 0, then v := maxΩ u0 is a supersolution if c max u0 ≥ 0 and γ max u0 ≥ 0. Ω Ω − Furthermore v := minΩ u0 is a subsolution if c min u0 ≤ 0 and γ min u0 ≤ 0. Ω Ω Obviously, these inequalities are all satisfied for c ≡ 0 and γ ≡ 0. Corollary 3.33 (Comparison to the ODE). Assume that f1 ≡ 0, f2 ≡ 0, γ = 0 and c(x, t) = c(t). Then v+ given as a solution of + ∂v + + cv ≥ 0 on Ω × (0,T ) (ODE) ∂t + v (0) = max u0 Ω is a supersolution. Furthermore, the function v− satisfying the same ODE with the reverse inequality and the initial value minΩ u0 is a subsolution. Part II. Nonlinear elliptic PDEs of second order 25 4. General theory for quasilinear problems 4.1. Fixed point theorems: From Brouwer to Leray-Schauder We start by recalling Brouwer’s fixed point theorem. n Theorem 4.1 (Brouwer). Let us denote by B the closed ball Br(x0) ⊂ R of radius r centered at x0. If f : B → B is continuous then f has a fixed point. Idea of proof. Suppose that f : Br(x0) → Br(x0) has no fixed point. Then x and f(x) always span a line. Therefore, we can define a map R : Br(x0) → ∂Br(x0): x 7→ Rx where Rx is given as the point of intersection between the line segment starting from f(x) in direction x and ∂Br(x0). Note that R is a retraction, i.e. R is continuous and satisfies Rx = x for all x in ∂Br(x0). If we regard the set of points in Br(x0) as a membrane. Then the map R describes a continuous deformation of such a membrane which moves all points to the boundary. Intuitively, the membrane will be torn apart. For a proof of the non-existence of such a retraction we refer to literature, e.g. [67], Section 1.14 or [55], Section 1.2. The proof also occurred as an exercise on homework sheet four in Michael Eichmair’s class Differential Geometry II from last semester. This result can be extended to compact, convex subsets of a Banach space. Theorem 4.2 (Schauder). Let (X, k · k) be a Banach space and A ⊂ X compact and convex. If T : A → A is continuous then T has a fixed point. Proof. Since A is compact for every k ∈ N we find a finite number N ∈ N of points xi ∈ A such that N [ A ⊂ B1/k(xi),N = N(k), xi = xi(k). i=1 co co Let us denote the convex hull of {xi | 1 ≤ k ≤ N} by Ak . Note that Ak ⊂ A. We define the continuous map PN dist x, A \ B (xi) xi J : A → Aco : x 7→ J (x) := i=1 1/k . k k k PN i=1 dist x, A \ B1/k(xi) co Since Ak is convex and generated by a finite number of elements it is homeomorphic via a map h to a closed ball B in some Euclidean space. From Brouwer’s Theorem 4.1 we −1 see that h ◦ (Jk ◦ T ) ◦ h : B → B has a fixed point and thus the same holds for Jk ◦ T co restricted to Ak . 27 28 4. General theory for quasilinear problems (k) We denote the sequence of fixed points of Jk ◦ T by (x )k∈N. Note that A is compact. (k ) Therefore, there exists a subsequence (x l )l∈N which converges to some x ∈ A as kl tends to infinity. We observe that 1 T x − x = lim T x(kl) − x(kl) = lim T x(kl) − (T ◦ J )(x(kl)) ≤ lim = 0. kl l→∞ l→∞ l→∞ kl Thus, x is a fixed point of T . Corollary 4.3. Let (X, k · k) be a Banach space and A ⊂ X closed and convex. If T : A → A is continuous and TA is relativly compact then T has a fixed point. Exercise II.1. Try to prove Corolloary 4.3. Hint: Ist the set (TA)co compact? Based on Corollary 4.3 we obtain the version which is important for our application. Theorem 4.4 (Schaefer). Let (X, k · k) be a Banach space and T : X → X continuous and compact. If M = x ∈ X ∃σ ∈ [0, 1] : x = σT x is bounded then T has a fixed point. ∗ Proof. Let M be strictly bounded by M > 0. Define the map T : BM (0) → BM (0) by T x for kT xkX ≤ M, T ∗x := M T x for kT xkX > M. kT xkX ∗ Note that T is continuous and BM (0) is closed and convex. Since T is compact T BM (0) ∗ ∗ is relatively compact. Thus, also T BM (0) is relatively compact and by Corollary 4.3 T ∗ ∗ has a fixed point x . Now, suppose that kT x kX > M. On the one hand ∗ ∗ ∗ M ∗ x = T x = ∗ T x kT x kX ∗ ∗ ∗ ∗ ∗ yields kx kX = M. On the other hand x = σT x (σ = M/kT x kX ) implies kx kX < M. ∗ ∗ ∗ ∗ ∗ Therefore, kT x kX ≤ M and x = T x = T x . One can even allow a more general dependence on the paramenter σ. Theorem 4.5 (Leray-Schauder). Let (X, k·k) be a Banach space and T : X×[0, 1] → X continuous and compact. If T (·, 0) = 0 and M = x ∈ X ∃σ ∈ [0, 1] : x = T (x, σ) is bounded then T (·, 1) has a fixed point. Proof. For our purpose the theorem of Schaefer will be sufficient. Therefore, we skip the proof and refer for it to [23], Theorem 11.6. 4.2. Reduction to a priori estimates in the C1,β-norm 29 4.2. Reduction to a priori estimates in the C1,β-norm We consider the following family of quasilinear Dirichlet problems of second order ij Qσu := a (·, u, Du)Diju + σb(·, u, Du) = 0 in Ω, (DP)σ u = σφ on ∂Ω with σ ∈ [0, 1], continuous coefficients aij, b ∈ C0(Ω × R × Rn) and symmetric matrix ij [a ]. Furthermore, we set Q := Q1 and (DP) := (DP)1. Definition 4.6. Let A ⊂ Rn. The operator Q is called elliptic in A if [aij(x, z, p)] is positive definite for all (x, z, p) ∈ A × R × Rn. Furthermore, Q is called elliptic w.r.t. v ∈ C1(A) if [aij(x, v(x), Dv(x))] is positive definite for all x ∈ A. Let T be the operator which assignes to v the solution u of the linear problem ij a (·, v, Dv)Diju + b(·, v, Dv) = 0 in Ω, u = φ on ∂Ω. (4.1) We see that the existence of a fixed point of T guarantees the existence of a solution of (DP). Based on this observation Theorem 4.4 yields a first criterion for existence. Theorem 4.7 (Existence criterion: C1,β-version). Let Ω ⊂ Rn be a bounded domain with C2,α-boundary. Let Q be elliptic in Ω with coefficients aij, b ∈ C0,α(Ω × R × Rn) and φ ∈ C2,α(Ω) for some α ∈ (0, 1). If there exists some β ∈ (0, 1) such that the set 2,α u ∈ C (Ω) ∃ σ ∈ [0, 1] : u solves (DP)σ is bounded in C1,β(Ω) independently of σ. Then (DP) has a solution in C2,α(Ω). Proof. Let v ∈ C1,α(Ω). Then (4.1) is a linear, uniformly elliptic problem for u with coefficients in C0,αβ(Ω). By Theorem 3.6 there exists a unique solution u ∈ C2,αβ(Ω) ⊂ C1,β(Ω). Thus, the operator T : C1,β(Ω) → C1,β(Ω) : v 7→ T v := u, is well defined. We need to show that T has a fixed point. By Theorem 4.4 this follows if T is a continuous, compact operator and the set M = {u ∈ C1,β(Ω) | ∃σ ∈ [0, 1] : u = σT u} is bounded. The latter is true by assumption1 so it remains to verify continuity and compactness. Compactness: By Theorem 3.6 we have kT vkC2,αβ (Ω) ≤ c kT vkC0(Ω) + kφkC2,αβ (Ω) + k|b(·, v, Dv)kC0,αβ (Ω) ij with c = c n, α, Ω, ka (·, v, Dv)kC0,αβ (Ω), minΩ λ(·, v, Dv) and from Theorem 3.2 we obtain the C0 estimate b(·, v, Dv) kT vkC0(Ω) ≤ kφkC0(Ω) + c(diam Ω) . λ(·, v, Dv) C0(Ω) 1Note that v ∈ M ⇒ v = σT v ∈ C2,αβ (Ω) ⇒ v = σT v ∈ C2,α(Ω). 30 4. General theory for quasilinear problems Therefore, we see that T maps bounded subsets of C1,β(Ω) into bounded subsets of C2,αβ(Ω) which by Arzel`a-Ascoliare relatively compact in C1,β(Ω), i.e. T is compact. 1,β 1,β Continuity: Suppose that {vm}m∈N ⊂ C (Ω) converges in the C -norm to some v. In particular this sequence is bounded in the C1,β-norm. As above this implies that 2,αβ {T vm}m∈N is bounded in the C -norm. Thus by Arzel`a-Ascoli {T vm}m∈N is relatively 2 compact in C (Ω). This is equivalent to the existence of a subsequence {T vmk }k∈N which converges in the C2(Ω)-norm to some u ∈ C2(Ω). Since aij and b are continuous this yields ij 0 = lim a (x, vm (x), Dvm (x))Dij(T vm (x)) + b(x, vm (x), Dvm (x)) k→∞ k k k k k ij = a (x, v(x), Dv(x))Diju(x) + b(x, v(x), Dv(x)) ∀ x ∈ Ω. This shows that T v = u. Finally, all subsequences have to converge to the same limit which shows that T is continuous. 4.3. Reduction to a priori estimates in the C1-norm The previous result shows that the existence proof is reduced to a priori estimates in C1,β(Ω). It turns out that the H¨olderestimate for Du can be carried out under very mild assumptions on the operator Q. This will help us to then formulate an existence criterion based on C1 a priori estimates. Theorem 4.8. Let Ω ⊂ Rn be a bounded domain with C2-boundary. Let Q be elliptic in Ω with coefficients aij ∈ C1(Ω × R × Rn), b ∈ C0(Ω × R × Rn) and let φ ∈ C2(Ω). If u ∈ C2(Ω) solves (DP) then there exists β ∈ (0, 1) such that ij Du ≤ C n, Ω,K, min λ, ka k 1 , kbk 0 , kφk < ∞ β,Ω UK C (UK ) C (UK ) C2(Ω) n with K := kukC1(Ω) and UK := Ω × [−K,K] × [−K,K] . Remark 4.9. The result was first obtained independently by De Giorgi and Nash for ij linear operators in divergence form: Lu = Di(a (x)Dju). It was a major break through in the study of nonlinear elliptic PDEs. Later Morrey and Stampachia extended the work to linear elliptic operators of general form. The theorem as it is stated above goes back to Ladyˇzhenskaya and Ural’ceva. Proof. The proof can be found in [23], Theorem 13.7. For our later application to the prescribed mean curvature equation it will be enough to consider operators in divergence form, i.e. a ∈ C1(Ω × R × Rn, Rn), b ∈ C0(Ω × R × Rn, R): div a(·, u, Du) + b(·, u, Du) = 0. (4.2) In that case the proof is contained in [23], Theorem 13.1 (interior estimate) and Theorem 13.2 (boundary estimate). We will only give a sketch of these arguments here. 4.3. Reduction to a priori estimates in the C1-norm 31 1 Interior estimate: Multiplying (4.2) by a test function ξ ∈ C0 (Ω), integrating over Ω and doing an integration by parts on the first term yields h i i a (·, u, Du)Diξ − b(·, u, Du)ξ dx = 0. ˆΩ We put ξ := Dkη and perform an integration by parts on the first term w.r.t. Dk: " ! # ∂ai ∂ai ∂ai k + Dku + j Dkju Diη + bDkη dx = 0 ˆΩ ∂x ∂z ∂p where all partial derivatives of ai and b are evaluated at (x, u(x), Du(x)). Thus we get