Count and symmetry of global and local minimizers of the Cahn-Hilliard energy over some cylindrical domains.

Size: px
Start display at page:

Download "Count and symmetry of global and local minimizers of the Cahn-Hilliard energy over some cylindrical domains."

Transcription

1 Noname manuscript No. (will be inserted by the editor) Count and symmetry of global and local minimizers of the Cahn-Hilliard energy over some cylindrical domains. Global and local minimizers Arnaldo S. do Nascimento João Biesdorf. Received: date / Accepted: date Abstract We consider the problem of minimization of the Cahn-Hilliard energy functional under a mass constraint over two and three-dimensional cylindrical domains. Although existence is presented for a general case the focus is mainly on rectangles, parallelepipeds and circular cylinders. According to the symmetry of the domain the exact number of global and local minimizers are given as well as their geometric profile and interface location; all are one-dimensional increasing/decreasing and odd functions for domains with lateral symmetry in all axes and also for circular cylinders. The selection of global minimizers by the energy functional is made via the smallest interface area criterion. The approach utilizes Γ convergence techniques to prove existence of an oneparameter family of local minimizers of the energy functional for any cylindrical domain. The exact number of global and local minimizers as well as their geometric profiles are accomplished via suitable applications of the unique continuation principle while exploring the domain geometry in each case and also the preservation of global minimizers through the process of Γ convergence. Keywords Global and local minimizers Γ -convergence Isoperimetric inequalities Unique Continuation Principle Mathematics Subject Classification (000) 35J0 35B06 35B07 35B5 49Q0 Arnaldo Simal do Nascimento Universidade Federal de S. Carlos, D.M São Carlos, S.P., Brasil arnaldon@dm.ufscar.br João Biesdorf Universidade Federal de S. Carlos, D.M São Carlos, S.P., Brasil jbiesdorf@dm.ufscar.br

2 Arnaldo S. do Nascimento, João Biesdorf. 1 Introduction From the physical standpoint minimizers of energy functionals, either global or local, are the most interesting critical points as they are the most likely observable configurations. In this work we are concerned with global and local minimizers of the so-called Cahn-Hilliard energy E ε (u) = [ε u + ε 1 (1 u ) ]dx (1.1) Ω under the constraint Ω u = 0 over domains Ω IRn (n =,3) of cylindrical type, i.e., Ω = ( l,l) D where D IR n (n = 1,) is a Lipschitz bounded domain. Our goal is to find, for ε > 0 small, the exact number of solutions to where inf E ε (u) (1.) u M M def = {u H 1 (Ω) : u = 0} Ω as well as local (which are not global) minimizers of E ε in M according to the symmetry and dimension of the domain. Moreover for rectangles, squares, parallelepipeds, cubes and circular cylinders we present a complete picture of its symmetry properties, geometric profile and location of the interfaces {u ε = 0} thus concluding that global and local minimizers are one-dimensional and odd functions, a type of Sturm-Liouville result. As expected the selection of global minimizers by the energy functional is made via the smallest interface area criterion. If the double-well potential F 1 (u) = (1 u ) in (1.1) is replaced with F (u) = (u α) (u β) where 0 < α < β < and the mass constraint replaced with Ω u = L n (Ω) (α + β) (here L n stands for the n-dimensional Lebesgue measure) then the functional (1.1) represents the free energy of a binary alloy and u the relative concentration of one of the two phases where the total mass m = L n (Ω) (α +β) is preserved. The reader interested in the physical background of this model is referred to [4], [3] or [], for instance, and the references therein. We take for simplicity α = 1 and β = 1 since all the results presented here, concerning existence of local and global minimizers (Sections 3 and 4) as well as their inherited symmetry, translate verbatim to the meaningful physical case 0 < α < β <. Let us mention some related works in order to set our own into perspective. The authors in [] showed existence of global and local minimizers of (1.) over the flat torus T n (n =,3), in which case it is called the periodic Cahn-Hilliard problem. In [1], for smooth domains Ω IR and Ω u = m 0 or m L (Ω), the authors showed existence of local minimizers of E ε (for ε > 0 small) whose transition layers are close to circular arcs and intersect the boundary orthogonally. In particular,

3 Global and local minimizers 3 under these hypotheses, they conclude that there are at least two such local minimizers and each of these arcs encloses a point on the boundary where the curvature of Ω attains its local maximum. The problem of minimizing Λ ε = εe ε over the unit square Ω = (0,1) (0,1) has been considered in [1] via a bifurcation approach where either ε or the total mass Ω u = m IR is the bifurcation parameter. Bifurcation of critical points from a particular class of eigenfunctions is considered and in particular for m fixed the author shows that the pointwise limit, as ε 0, of some conditionally critical points are global minimizers of the limiting problem Λ 0. Monotonicity, regularity and other symmetry properties of local minimizers of E ε under the constraint Ω u = m over cylindrical domains has been studied in [6] but existence is not considered. In [] the author gives some accounts on symmetry and monotonicity of local minimizers of some variational problems subjected to a integral constraint in cylinders and annuli with periodic phase separation and also zero Neumann boundary condition. Again existence is not the issue. It follows from [11] that any local minimizer of E ε in M over any smooth strictly convex domain must have a connected phase boundary. Herein we draw the same conclusion for just convex domain of cylindrical type. Let us recall some terminology which is going to be used: 1. A solution u to (1.) is called a global minimizer of E ε in M.. A function u M is called a local minimizer of E ε in M if there is a neighborhood V (u) of u in H 1 (Ω) such that E ε (u) E ε (v), v M V. If the inequality is strict then u is called a strict local minimizer. 3. Any u M satisfying E ε(u ε )ϕ = 0 and E ε (u)ϕ 0, ϕ M is called a conditionally stable critical point of E ε. First we state our main result when the domain is a rectangle. As usual we denote the characteristic function of a set A by χ A. Theorem 1 Let Ω = R = ( l,l) ( r,r) with l r > 0. Then, for ε small, (1.) has a family of solutions {u ε } ε>0 in C (R) C 0 (R) (global minimizers of E ε in M ) satisfying 1. u ε depends just on the first variable,. u ε is odd and increasing in ( l,l) and 3. u ε u 0 L 1 ( l,l) 0, where u 0 = χ ( l,0) χ (0, l). (1.i) If l > r, i.e., R is a rectangle, problem (1.) has, for ε small, only two solutions: u ε (given above) and u ε. There are only two other local (but not global) minimizers v ε and v ε, where v ε depends only on the second variable and is 0, where v 0 = increasing and odd in ( r,r). Moreover v ε v 0 L 1 ( r,r) χ ( r,0) χ (0,r). (1.ii) If R is a square, then problem (1.) has, for ε small, only four solutions u ε (x 1 ), u ε (x 1 ), v ε (x ), v ε (x ) where v ε is obtained from u ε by a rotation of π/ with respect to x 1 -axis. There is no other local minimizer.

4 4 Arnaldo S. do Nascimento, João Biesdorf. A similar result holds also for three-dimensional cylindrical domains. We label the points x = (x 1,x,x 3 ) IR 3 as x = (x 1,x ) where x = (x,x 3 ). Theorem Let Ω = ( l,l) D where D IR is any Lipschitz bounded domain. Then, for ε small, E ε has a family of local minimizers {u ε } ε>0 in C (Ω) C 0 (Ω) such that 1. u ε depends just on x 1,. u ε is odd and increasing in ( l,l) and 3. u ε u 0 L 1 ( l,l) 0, where u 0 = χ ( l,0) χ (0,l). In particular if (.i) D = B R = {x IR : x < R} then (1.) has only two solutions (global minimizers of E ε in M ): u ε (x 1 ) and u ε (x 1 ), as above. Moreover E ε has no other local minimizer in M. (.ii) Ω = ( l,l) ( r,r) ( q,q) with l > r > q then (1.) has only two solutions, u ε (x 1 ) and u ε (x 1 ), and there are only four other local (but not global) minimizers v ε (x ), v ε (x ), w ε (x 3 ) and w ε (x 3 ); v ε is increasing and odd in ( r,r) whereas w ε is increasing and odd in ( q,q). Moreover v ε v 0 L 1 ( r,r) 0 (v 0 as in Theorem1) and w ε w 0 L 1 ( r,r) 0 where w 0 = χ ( q,0) χ (0,q). (.iii) Ω is a cube (l = r = q) then problem (1) has only six solutions u ε (x 1 ), u ε (x 1 ), v ε (x ), v ε (x ), w ε (x 3 ) and w ε (x 3 ) and there is no other local minimizer. The results above regarding existence are sharp and show that, for ε small, the geometry of global and local minimizers of the energy functional restricted to M is quite simple as they present no oscillation, a type of Sturm-Liouville property. Regarding existence the main tool utilized in our approach is a theorem due to De Giorgi [14] based on the theory of Γ convergence. It is well known that the family of functional E ε Γ converges, as ε 0, to the perimeter or area functional { 83 Per Ω {u = 1}, v BV (Ω,{±1}) and Ω v = 0 E 0 (v) =, otherwise. Existence of local minimizers is obtained by proving that E 0 has an isolated L 1 (Ω)-local minimizer, say u 0, and then De Giorgi s theorem assures that close (in L 1 (Ω) norm) to u 0 there is a local minimizer of the original problem. The many cases of uniqueness of local and global minimizers is obtained using a property of Γ -convergence (preservation of global minimizers) and a uniqueness result of a onedimensional monotone solutions of the corresponding elliptic Neumann boundary value problem. In particular we prove that when Ω is a cylindrical domain the admissible function u 0 (x) def = χ {x1 >0}(x) χ {x1 <0}(x), x Ω is indeed a local minimizer of E 0. This gives a positive answer to a question raised in [10], p. 80, for the special case of a rectangle. Note that for competitors with fixed boundary a more general result has recently been established in [5], Theorem. The authors consider a smooth Riemannian

5 Global and local minimizers 5 manifold of dimension n 7 and let S be a smooth oriented constant-mean-curvature hypersurface with boundary with positive second variation for fixed volume and boundary. Then the authors prove that S is uniquely homologically area minimizing for fixed volume among oriented hypersurfaces in a small L 1 neighborhood. Finally it is worthwhile to comment on the procedure used to conclude that global and local minimizers depend only on one variable and is increasing (or decreasing) and odd in this variable. This procedure is not new and has been used in different scenarios and in a systematic way in [8] and [9] (although previously used in [6]) to study radial symmetry or symmetry with respect to a hyperplane [6] of global minimizers of a class of functionals under a class of constraints which, by the way, do not comprise ours. Basically the technique consists of taking a global minimizer U and then defining a function V as the reflection, with respect to a convenient hyperplane, of U restricted to one of the two sides of the partitioned domain. For the class of functionals considered, U and V have the same energy and thus V is also a global minimizer and then the Unique Continuation Principle is used to obtain (in case of an infinite cylinder) that U is symmetric with respect to this hyperplane. If applied to our case such a procedure would produce even functions whereas we are looking for odd ones. Essentially this difficulty is avoided by taking a change of dependent variable w = 1 u and then using the information gathered while obtaining the minimizers via Γ convergence. Another difficulty in our scenario is that, as opposed to the above cases, we cannot reflect the local minimizer around the plane of symmetry of the domain. This will become clear in Section 5. Background and notation In this section we give some notation and basic results of geometric measure theory which are going to be used. For more details on this matter the reader is referred to [16], [17] and [18], for instance. We denote by L n the n dimensional Lebesgue measure and by H n the n dimensional Hausdorff measure. The density of a set E at a point x is defined by L n (E B r (x)) D(E,x) = lim r 0 L n, (B r (x)) with B r (x) denoting the n-dimensional ball centered at x with radius r. Let f be any real-valued function in IR n and let x IR n. The approximate upper limit and lower limit of f at x are defined by f + (x) = inf{t : D({ f > t},x) = 0}and f (x) = sup{t : D({ f < t},x) = 0}, respectively. If f + (x) = f (x) < then the function f is said to be approximately continuous at x. If µ is a Borel measure on a domain Ω with values in [0,+ ) its total variation is denoted by µ and the integral of a µ -integrable function f will be denoted by R f d µ.

6 6 Arnaldo S. do Nascimento, João Biesdorf. For Ω IR N a bounded set, the space BV (Ω) of functions of bounded variation in Ω is defined as the set of all functions v L 1 (Ω) whose distributional gradient Dv is a Radon measure with bounded total variation in Ω, i.e., Dv (Ω) = sup { v(x)divσ(x)dl n <, σ C0(Ω,IR 1 n ), σ 1}. R If v W 1,1 loc then the total variation measure satisfies Dv = L n v, i.e., Dv = v dl n, where denotes the usual gradient. A set E Ω is said to have finite perimeter in Ω if Dχ E (here χ E stands for the characteristic function of E) is a vector valued Radon measure in Ω with finite total variation. The perimeter of E in Ω is defined by Per Ω (E) = Dχ E (Ω). If E has finite perimeter in Ω and L n (E Ω) < then χ E BV(Ω). Denoting by ν E j ( j = 1,...,n) the derivative of the measure D j χ E with respect to Dχ E we have for every x Ω ν E j (x) = lim r 0 D j χ E (B r (x)) Dχ E (B r (x)) ( j = 1,...,n). as long as the limit exists. The reduced boundary of E, denoted by E, is the set of all points x E such that the vector ν E (x) = (ν E 1 (x),...,νe N (x)) exists and νe (x) = 1. If E has finite perimeter then E is a rectifiable set, and consequently and Dχ E = ν E H N 1 E Dχ E = H N 1 E (.1) D j χ E = ν E j H N 1 E ( j = 1,...,N). (.) If E IR n then we say x M E, the measure theoretic boundary of E, if L n (B(x,r) E) lim sup r 0 r n L n (B(x,r) E) > 0 and limsup r 0 r n > 0. It holds that E M E and H n 1 ( M E\ E) = 0. For future references we recall the well-known isoperimetric inequality (see [5], for instance): if E R n (n ) is a set of finite perimeter then H n (E) n 1 n c 1 H n 1 ( E) (.3) where the constant c 1 depends only on n. We denote by BV (Ω;{±1}) the class of all u BV (Ω) which take values 1 and 1 only. If u BV (Ω;{±1}) then we call µ 1,..., µ n the Radon (signed) measures

7 Global and local minimizers 7 defined by the partial derivatives u x j ( j = 1...,n) and µ = Du (in the sense of distribution). Its total variation measures µ j and µ satisfy for any measurable Borel set B µ (B) (Σ n j=1 ( µ j (B)) ) 1/. (.4) Our approach relies strongly on the concept of Γ convergence and therefore we rather give the version which is going to be used. Definition 1 The extended-real family of functionals {Λ ε } ε>0 defined in L 1 (Ω), Γ - converges, as ε 0, to a real-extended functional Λ 0 in a point u L 1 (Ω) if 1. {u ε } L 1 (Ω) : u ε u L 1 (Ω). {w ε } : w ε u L 1 (Ω) 0 = liminf Λ ε (u ε ) Λ 0 (u). 0 and Λ 0 (u) limsup Λ ε (w ε ). As mentioned above the following result which is an application, found in [10], of a general result due to De Giorgi [14] is essential to the solution of our problem. It relates isolated local minimizers of the Γ -limit to local minimizers of the original problem. Theorem 3 Suppose that a family of extended-real functionals {Λ ε } ε>0 defined in L 1 (Ω) Γ -converges, as ε 0, to a extended functional Λ 0 and that the following hypotheses are satisfied (3.1) {v ε } ε>0 : Λ ε (v ε ) constant < is compact in L 1 (Ω) and (3.) There exists an isolated L 1 -local minimizer v 0 of Λ 0. Then there exists ε 0 > 0 and a family {v ε } 0<ε ε0 such that (3.i) v ε is an L 1 -local minimizer of Λ ε and (3.ii) v ε v 0 L 1 (Ω) 0. 3 Existence of local minimizer via Γ convergence In order to work in L 1 (Ω), a space whose topology has the compactness property which we need with respect to BV (Ω), we define E ε : L 1 (Ω) IR { } by { E ε (u) def Ω = [ε u + ε 1 (1 u ) ]dx, if u M (3.1), otherwise. Lemma 1 The family of functionals {E ε } 0<ε<ε0 Γ converges in L 1 (Ω), as ε 0, to { 83 E 0 (v) def Per Ω {v = 1}, v BV (Ω,{±1}) and Ω = v = 0 (3.), otherwise. A proof can be found in [3], for instance; in this case the Γ limit has long been known and it is just the perimeter or area functional given by (3.). Let us define the sets Ω l def = {x Ω : x 1 < 0} and Ω r def = {x Ω : x 1 > 0}

8 8 Arnaldo S. do Nascimento, João Biesdorf. and the function Note that Ω u 0 = 0. When u 0 (x) def = χ Ωr (x) χ Ωl (x), x Ω (3.3) Ω = R def = ( l,l) ( r,r)(l r) then, in addition to u 0, we define the function v 0 as v 0 (x) def = χ Ωr (x) χ Ωl (x), x Ω (3.4) where When Ω l def = {x Ω : x < 0} and Ω r def = {x Ω : x > 0} Ω = Q def = ( l,l) ( r,r) ( q,q)(l r q) then, in addition to u 0 and v 0, we define the function w 0 as w 0 (x) def = χ Ωr (x) χ Ωl (x), x Ω (3.5) where Ω l def = {x Ω : x 3 < 0} and Ω r def = {x Ω : x 3 > 0}. Theorem 4 Let Ω = ( l,l) D where D IR is any Lipschitz bounded domain. There exists ε 0 > 0 and a family {u ε } 0<ε ε0 in C (Ω) C 0 (Ω) satisfying (4.i) u ε is a local minimizer of E ε in M and (4.ii) u ε u 0 L 1 (Ω) 0, with u 0 given by (3.3). When Ω = R (rectangle) then in addition to {u ε } we obtain another family {v ε } 0<ε ε0 of local minimizers satisfying v ε v 0 L 1 (Ω) 0, with v 0 given by (3.4). When Ω = Q (parallelepiped) then, in addition to {u ε } and {v ε }, we obtain another family {w ε } 0<ε ε0 of local minimizers satisfying w ε w 0 L 1 (Ω) 0, with w 0 given by (3.5). This theorem follows straight from Theorem 3 once the hypotheses (3.1) and (3.) are verified. The latter hypothesis, which is by far the most difficult to prove will be the subject of the next section. Therefore the proof of this theorem is postponed until we verify (3.) for the cases.

9 Global and local minimizers 9 4 Proof Theorem 4 Most of this section is devoted to proving that u 0, given by (3.3), is a isolated local minimizer of E 0 in M. But before doing that an important remark is worth being made. For n = this result could have been obtained by generalizing [10], Proposition.4, where the result pursued herein is established for n = in the unconstrained case for non-convex domains which have a narrow region. However the proof provided in [10] is not suited for generalization to dimension n 3. We will explain where and why it fails for n 3. In order to prove Proposition.4 ([10]) the authors split the admissible functions in four classes. In one of these classes the admissible functions BV(R;±1) in the rectangle R = ( l, l) ( r, r), for some a (l/, l) are either 1. not constant along the interval {(x, a), l < x < l} or. not constant along the interval {(x, a), l < x < l}. In either case the admissible function v changes sign at least once in at least one of these intervals enabling the authors to conclude (see [10], first inequality in p. 77) l l ( ) v x 1 (x 1,a) + v x 1 (x 1, a). (4.1) Note that in this case the measure of integration, in view of (.), can be thought of as the Hausdorff counting measure H 0 of points of discontinuities of v(,a). When n = 3, under the same hypotheses, (4.1) depends on the Per D {v = 1}. In such case the lower positive bound given by (4.1) would no longer be available, i.e., when taking the limit of (4.1) over this class of functions the limit inferior could be zero and the procedure utilized in [10] no longer works. The proof rendered here for the three-dimensional domain ( l,l) D is of geometric nature and the basic idea is to find t (l/, l) such that H ((D {t}) {v = u 0 }) Per D [t,l] {v = u 0 } holds for any admissible function v where u 0 is given by (3.3). This is accomplished by a series of technical results which to some extent is the price paid for working in the topology of L 1. This topology, by its turn, is chosen in order check hypothesis (3.1) in Theorem 3 (in order to prove Theorem 4) by exploring the compact imbedding BV (Ω) L 1 (Ω) (see [15] or [3], for instance). We now state the main result of this section. Lemma Let Ω IR n (n =,3) be a cylindrical domain Ω = ( l,l) D. Then the function u 0, given by (3.3), is a local isolated minimizer of E 0 given by (3.), i.e., there is ρ > 0 such that for any v BV (Ω,{±1}), satisfying 0 < v u 0 L 1 (Ω) < ρ and Ω v = 0, we have E 0(u 0 ) < E 0 (v). As preparation for the proof of this lemma we state and prove some technical results.

10 10 Arnaldo S. do Nascimento, João Biesdorf. Lemma 3 Let D IR m (m = 1,) a bounded Lipschitz domain with a finite number of vertices if m = and an interval if m = 1. There exists θ > 0 such that for any A D of finite perimeter satisfying H m (A) < 1 H m (D) it holds that 0 < θ H m 1 ( A\ D) H m 1 (. (4.) A) Proof We may suppose that A is connected and then the general case follows at once. When m = 1, θ = 1/ since H 0 ( A) = and H 0 ( A\ D) 1. Thus suppose m = and let N D denote the finite set of points where D is not C 1. We abbreviate connected component as c.c., for short, and set diag(n D ) = {(P,P) : P N D }. Then we define a function γ : ( D D)\diag(N D ) R + by { { min H γ(p,q) = 1 (C) : C is c.c. of D\{P,Q} } if P Q, 0, se P = Q, Note that D\{P,Q} consists of exactly two connected components when P Q given that now D IR. Hence γ is well defined and continuous. Next we define a function σ : ( D D)\diag{N D } R + by σ(p,q) def = { P Q γ(p,q)+ P Q if P Q, 1/, if P = Q, From the fact that D is a Lipschitz domain one can easily check that lim inf σ(p,q) > 0 (4.3) (P,Q) diag(n D ) On the account that D is compact and σ is continuous there is a constant m σ > 0 satisfying σ m σ > 0 on ( D D)\diag{N D }. (4.4) With c 1 as in (.3) (the isoperimetric inequality) and m σ as in (4.4) we define { } θ def m σ H (D) 1/ = min m σ,. m σ H (D) 1/ + c 1 Per(D) Hence 0 < θ < 1 and we claim that θ satisfies (4.). Indeed given A D satisfying our hypotheses then three cases may occur: (i) H 1 ( A D) = 0 or (ii) 0 < H 1 ( A D) < Per(D). (iii) H 1 ( A D) = Per(D).

11 Global and local minimizers 11 From now on, in order to simplify the notation, we replace the H 1 measure of a set in Ω with its perimeter in Ω. If (i) holds then Per D (A) = Per(A) and as such Per D (A)/Per(A) = 1 > θ. We will say that P and Q, with P Q and P,Q A D, are extremal points of A D if A D\{P,Q} is contained in one of the two c.c. of D\{P,Q}. If (ii) holds then two cases may occur. (ii-a) P and Q, extremal points of A D so that A D is contained in the smaller c.c. (in the H 1 metric) of D\{P,Q}. In this case using that P Q Per D (A) and if a, b and c are strictly positive real numbers with a b, then Per D (A) Per(A) = a c+a b b+c, we have Per D (A) Per D (A) + H 1 ( A D) x y x y + γ(x,y) = σ(x,y) m σ θ. Per D (A) Per D (A) + γ(x,y) (ii-b) For any two points P and Q, extremals of A D, the set A D is contained in the largest c.c. of D\{P,Q}. In this case each c.c. of D\A, say C, satisfies (ii-a) with A replaced with C and therefore Per D (D\A) Per(D\A) m σ. (4.5) Invoking the hypothesis H (A) < 1 H (D) (which has not been used yet) we obtain Now (.3), (4.5) and (4.6) yield H (D\A) > 1 H (D). (4.6) Per D (A) = Per D (D\A) m σ Per(D\A) m σ 1 c 1 H (D\A) 1/ > This implies m σ 1 c 1 1 H (D) 1/. Per D (A) Per(A) = 1 m σ Per D (A) Per D (A) + Per(D) > 1 m σ c1 mσ H (D) 1/ mσ H (D) 1/ + c 1 Per(D) θ. 1 c 1 H (D) 1/ 1 H (D) 1/ + Per(D) In case (iii) each c.c. of D\A fits in (i) and as such the very same argument used in (ii-b) establishes the proof when m =. Remark 1 As it stands the difficulty we have found to generalize our results to dimension n > 3 is to prove Lemma 3 for higher dimensions.

12 1 Arnaldo S. do Nascimento, João Biesdorf. Lemma 4 Let ( l,l) D = Ω R n (n =,3) as in Section 3, c 1 the constant from the isoperimetric inequality (.3) and θ as in Lemma 3. Then there exists a real number ξ ( l,l) and a continuous function η : [ l,ξ ] [0, ) satisfying η(t) = θ ξ c 1 t η(s) n 1 n ds η(ξ ) = 0 η( l ) 1H n 1 (4.7) (D) η > 0 on [ l,ξ ). Proof One can easily check that ( ) θ n 1 η(t) = (ξ t) n 1 (n =,3) (n 1)c 1 is a solution to (4.7) where the point ξ (l/,l) should be chosen, if necessary, close to l/ in order to satisfies the condition η( l ) 1 H n 1 (D). Lemma 5 Let Ω, c 1 and θ be as in Lemma 4. There exists ρ > 0 such that for any L - measurable function ψ : [ l,l] (0,H n 1 (D)] (n =,3) satisfying l l ψ(s)dh 1 < ρ there exists t 0 ( l,l)\ I, where I = {t [ l,l] : ψ(t) 1 H n 1 (D)}, such that ψ(t 0 ) < θ ψ(s) n n 1 ds. c 1 [t 0,l]\I Remark Before rendering the proof let us give a geometric insight of the meaning of the function ψ. Given any function v BV (Ω,±1) in a L 1 (Ω)-neighborhood of u 0 (see statement of Lemma ) consider the set A = {v = u 0 }, say. Then ψ will be taken as ψ(t) = H (n 1) (A (D {t})). In general ψ will not be a continuous function and, in addition, its image must be in [0,H n 1 (D)]. However in the proof of the main result in this section we will see that its domain must be restricted to {t : ψ(t) < 1 H n 1 (D)}. That is why the set I was introduced. Proof Recalling that Ω = ( l,l) D and taking η and ξ as in Lemma 4 we define ξ ρ = min{ η(t)dt, (l ξ ) H n 1 (D)}, n =,3. (4.8) l/ We anticipate that ρ will be the radius of the neighborhood of u 0 in L 1 (Ω) when proving that u 0 is an isolated local minimizer of E 0 in M as in Lemma. Given ψ and I as in our hypotheses we define { [ ] [ ) } l l O = t,l \ I : s,t \ I with L ([0,s]\ I) = L ([0,t]\ I) Ĩ = I O, ϑ = L ([0,l]\ I) and ζ : [ l,l]\ Ĩ [ l,ϑ] given by ζ (t) = L ([0,t]\ Ĩ).

13 Global and local minimizers 13 Note that L (O) = 0 and that O was introduced just to make ζ injective. Hence ζ is a measurable bijective function and we can define a L -measurable function g : [ l,ϑ] (0, 1 H n 1 (D)) by which, from the hypothesis of ψ, satisfies ϑ l/ g = ψ ζ 1 ξ g(s)ds < η(t)dt. (4.9) l/ Claim: t 1 ( l,ϑ) such that g(t 1) < θ c 1 ϑ t 1 g(s) n n 1 ds. Let us first prove that ϑ > ξ ; it suffices to show that L (I) < l ξ. Arguing by contradiction we have ρ (l ξ ) H n 1 (D) (from the definition of ρ) 1 H n 1 (D)L (I) (contradiction hypothesis) < ψ(s)ds (from definition of I) < I l l/ ψ(s)ds (sincei [ l,l] ) ρ (from definition of ρ). Therefore ϑ > ξ and this make it possible to define the set def = {t [ l,ξ ) : g is approximately continuous at t and g(t) η(t)}. Note that g is approximately continuous a.e. in [ l,ξ ] (cf. [4], p.47). Also /0, for if ξ = l/ then by definition of we would have g > η a.e. in [l/,ξ ) and therefore ϑ l/ g > ξ l/ g > ξ l/ η, thus contradicting (4.9). Therefore l/ < ξ. There are two cases to be analyzed. i) sup = ξ. In this case inf{g(t) : t [ l,ξ ] : g aprox. continuous at t} = 0. On the other hand since g n n 1 > 0 we have ϑ ξ g n 1 n ds > 0 and this implies the existence of t 1 [ l,ξ ] such that g is aprox. continuous at t 1 and ξ ξ g(t 1 ) θ ϑ g(s) n n 1 ds θ ϑ g(s) n 1 n ds. c 1 ξ c 1 t 1 ii) sup = ξ ξ. In this case, since ξ < ξ, it follows from the definition of that ξ n g(t) n 1 dt > ξ n η(t) n 1 dt. Also ϑ > ξ, g > 0 and g is approximate continuous a.e. in [l/,ξ ) where l/ < ξ < ξ.

14 14 Arnaldo S. do Nascimento, João Biesdorf. Together all these facts allow us to take t 1 ( l,ξ ] such that g is approximate continuous at t 1, g(t 1 ) η(t 1 ) and ξ t 1 ϑ η(t) n n 1 dt < g(t) n 1 n dt. ξ If g is continuous then we may choose t 1 = ξ. From the above properties of t 1, Lemma 4 and the three aforementioned properties we have g(t 1 ) η(t 1 ) = θ c 1 = θ ξ c 1 < θ c 1 < θ c 1 t 1 ϑ ξ ϑ t 1 ξ t 1 η(t) n n 1 dt η(t) n n 1 dt + θ c 1 g(t) n n 1 dt + θ c 1 g(t) n n 1 dt, ξ ξ ξ ξ η(t) n n 1 (t)dt g(t) n n 1 dt establishing in this way our claim. Now in order to complete the proof of the lemma we take t 0 = ζ 1 (t 1 ). The above sequence of elementary lemmas is utilized to prove the next result which in turn will play an essential role in the proof of Lemma. Lemma 6 Let Ω = ( l, l) D R n (n =,3) as above, c 1 as in (.3), θ as in Lemma 3 and ρ given by (4.8). (6.i) Given A + D [ l,l ], with H n (A + ) < ρ, there exists t 0 ( l,l ) such that H (n 1) (A + (D {t 0 })) 1 H (n 1) ( (A + (D [t 0,l]))\ (D [t 0,l])). (6.ii) Given A D [ l, l ], with H n (A ) < ρ, there exists t 0 ( l, l ) such that H (n 1) (A (D {t 0})) 1 H (n 1) ( (A (D [ l,t 0]))\ (D [ l,t 0])). Proof (6.i) If H (n 1) (A + (D {t})) = 0 for some t ( l,l ), then we may choose t 0 = t. Otherwise let ψ : [ l,l ] (0,H n 1 (D)] defined by ψ(t) def = H (n 1) (A + (D {t})). Then due to our hypotheses, ψ satisfies the hypotheses of Lemma 5 and there follows the existence of t 0 ( l,l ) \I (recall that I = {t [ l,l] : ψ(t) 1 H n 1 (D)})

15 Global and local minimizers 15 satisfying H (n 1) (A + (D {t 0 })) θ (H (n 1) (A + (D {s}))) n 1 n dh 1 c 1 [t 0,l]\I (using isoperimetric inequality (.3)) θ H (n ) ( (A + (D {s})))dh [t 0,l]\I [t 0,l]\I l Per D {s} (A + (D {s})))dh 1 (from Lemma 3) t 0 Per D {s} (A + (D {s})))dh 1 1 H (n 1) ( (A + (D [t 0,l]))\ (D [t 0,l])). This establishes (6.i) and the proof of the other case is similar. Finally, as a consequence of the above sequence of elementary results, we are in position to prove the main result of this section. Proof of Lemma The coarea formula yields E ε (u 0 ) = 8 3 H n 1 (D). (4.10) Next we take ρ as in (4.8) and v BV (Ω,{ 1,1}) satisfying 0 < v u 0 L 1 (Ω) < ρ and Ω v = 0. Note that by fixing t ( l,l) the trace of v(t, ) is well defined in each crosssection D t = {t} D. Set def 1. A + v = {x ( l,l ) D : v(x) = 1}. A def v = {x ( l, ) l D : v(x) = 1}. Since A + v {v = u 0 } and A v {v = u 0 } we have ρ > v u 0 L 1 (Ω) > 1 v u 0 L 1 (Ω) = 1 = 1 dl n = H n ({v = u 0 }) {v= u 0 } {v u 0 } v u 0 dl n H n (A + v ). (4.11) Thus H n (A + v ) < ρ and similarly H n (A v ) < ρ. Therefore A + v e A v satisfy the hypotheses of (6.i) and in (6.ii), respectively, with A + and A replaced with A + v and A v, respectively. In this way by Lemma 6 there are t 0 ( l,l ) and t 0 ( l, l ) such that H n 1 (A + v ({t 0 } D)) 1 H n 1 ( (A + v ([t 0,l] D))\ ([t 0,l] D))

16 16 Arnaldo S. do Nascimento, João Biesdorf. and H n 1 (A v ({t 0} D)) 1 H n 1 ( (A v ([ l,t 0] D)\ ([ l,t 0] D)). By fixing t 0 e t 0 and, for the sake of simplification in notation, setting D v = {x D : (t 0,x ) A + v } {x D : (t 0,x ) A v }, A [t 0,l] v the two inequalities above yield = (A + v ([t 0,l] D))\ ([t 0,l] D) and A [ l,t 0 ] v = (A v ([ l,t 0] D)\ ([ l,t 0] D) H n 1 (D v ) 1 H n 1 (A [t 0,l] v ) + H n 1 (A [ l,t 0 ] v ). (4.1) Our next goal is to use these inequalities to prove By definition of D v H n 1 ( {v = 1} Ω) > H n 1 (D). (4.13) H n 1 (D v ) 1 [ ] H n 1 (A [t 0,l] v ) + H n 1 (A [ l,t 0 ] v ) (from 4.1) = 1 H n 1 ( {v = u 0 } ((( l,t 0) (t 0,l)) D)) = 1 H n 1 ( {v = 1} ((( l,t 0) (t 0,l)) D)). (4.14) On the other hand note that by definition of D v, v(t 0,x ) = u 0 (t 0,x ) = 1 and v(t 0,x ) = u 0 (t 0,x ) = 1, a.e. in D\ D v. (4.15) Denoting the interior of a set A by int(a), we obtain H n 1 ( {v = 1} ((t 0,t 0 ) int(d\d v )) H n 1 (D\D v ) (4.16) If (4.16) is a strict inequality then from (4.14) we would obtain H n 1 ( {v = 1} Ω) H n 1 ( {v =} ((t 0,t 0 ) int(d\d v )) + H n 1 ( {v = 1} ((( l,t 0) (t 0,l)) D)) > H n 1 (D\D v ) + H n 1 (D v ) = H n 1 (D), (4.17) thus establishing (4.13). If (4.16) is an equality then necessarily {v = 1} ((t 0,t 0) int(d\d v ) and D\D v are parallel cross-sections of Ω and therefore two cases have to be analyzed

17 Global and local minimizers 17 (4.16.i) H n 1 (D v ) > 0 and (4.16.ii) H n 1 (D v ) = 0 If the former one holds then (4.14) yields H n 1 (D v ) 1 H n 1 ( {v = 1} ((( l,t 0) (t 0,l)) D)) < H n 1 ( {v = 1} ((( l,t 0) (t 0,l)) D)) (4.18) and hence from (4.16) and (4.18) we obtain H n 1 ( {v = 1} Ω) H n 1 ( {v = 1} ((t 0,t 0 ) int(d\d v )) + H n 1 ( {v = 1} ((( l,t 0) (t 0,l)) D)) > H n 1 (D\D v ) + H n 1 (D v ) = H n 1 (D) (4.19) thus proving (4.13). Now if (4.16.ii) holds then taking into account that (4.16) is an equality, {v = 1} ((t 0,t 0) int(d\d v ) and D\D v are parallel cross-sections of Ω, v u 0 L 1 (Ω) > 0 and Ω v dx = 0, we necessarily have Hence H n 1 ( {v = u 0 } ((( l,t 0) (t 0,l)) D)) > 0. H n 1 ( {v = 1} Ω) = H n 1 ( {v = 1} ((( l,t 0) (t 0,l)) D)) + H n 1 ( {v = 1} ((t 0,t 0 ) D)) = H n 1 ( {v = u 0 } ((( l,t 0) (t 0,l)) D)) + H n 1 ( {v = 1} ((t 0,t 0 ) D)) > H n 1 ( {v = 1} ((t 0,t 0 ) D)) = H n 1 (D), (4.0) thus establishing (4.13) in this case too. Altogether (4.0), (4.19) and (4.17) imply (4.13). Finally from (4.13) we obtain E 0 (v) = 8 dh n 1 = 8 3 Ω {v=1} 3 H n 1 (Ω {v = 1}) > 8 3 H n 1 (D) = E 0 (u 0 ), and this completes the proof of Lemma. Proof of Theorem 4 Note that (4.i) and (4.ii) follow from Theorem 3, (3.i) and (3.ii) respectively, once we verify hypotheses (3.1) and (3.). But the proof of (3.1) can be found in [15] and [3], for instance, and (3.) has been verified in Lemma. Regularity of the solutions follows from classic elliptic regularity theory and can be found in [1], for instance.

18 18 Arnaldo S. do Nascimento, João Biesdorf. The proof of existence of a family of local minimizers v ε satisfying v ε v 0 L 1 (Ω) 0 (v 0 as in (3.4)), in case of a rectangle, and also of w ε satisfying w ε w 0 L 1 (Ω) 0 (w 0 as in (3.5)), in case of a parallelepiped, is the same proof given for u ε since each domain has the same symmetry property in each direction. 5 Symmetry inherited by local minimizers The objective in this section is to give a complete geometric picture of the global and local minimizers of E ε, found in Theorem 4, when Ω is any of the cylindrical domains considered in Theorem as well as the exact number of these minimizers in each case. In particular it follows that among all critical points of E ε only the ones whose zero-level set are connected (and intersects the boundary) can minimize the energy functional. In other words, the more a function oscillates the more energy it carries; a kind of Sturm-Liouville result. Since from now on we will deal with the Euler-Lagrange equation for E ε, for simplicity in notation, unless otherwise stated, we re-scale the family of functionals E ε as [ ε E ε (u) = Ω u + 1 ] 4ε (1 u ) dx. In the next lemma we exploit the symmetry of the domains to obtain symmetry properties of the local minimizers via an application of the Unique Continuation Principle under very weak hypotheses. In cylindrical coordinates (x 1,ρ,θ), for any fixed θ [0,π), we set Ω r ( θ) def = {(x 1,ρ,θ) : l < x 1 < l, 0 < ρ < R, θ < θ < θ + π} and denote by Ω l ( θ) the interior of the set Ω \ Ω r ( θ). Also, as usual, B R (0,0) = {(ρ,θ), 0 < ρ R, 0 θ < π}. Lemma 7 Consider the family {u ε } of local minimizers of E ε in M, found in Theorem 4, satisfying u ε u 0 L 1 (Ω) 0 (u 0 as in (3.3)), for the domain Ω = ( l,l) D where D IR n (n =,3) is any bounded Lipschitz domain. Then (7.i) u ε is increasing in the first variable and u ε (x 1,x ) = u ε ( x 1,x ), x D. In particular, u ε < 0 on Ω l and u ε > 0 on Ω r. Proof (7.i) A regularity result presented in [6] for cylindrical domains assures that u ε W,3 (Ω) C 3+δ (Ω), 0 < δ < 1. Any critical point of E ε in M satisfies the Euler-Lagrange equation ε u ε + ε 1 (u ε u 3 ε) = λ where the constant λ is the Lagrange multiplier corresponding to the constraint.

19 Global and local minimizers 19 Since u ε u 0 in L 1 (Ω) it holds that 1 u ε / 0, for ε small enough. Since 1 u ε / 0 we may evoke [], Theorem 3.3, to conclude 1 u ε > 0 on Ω (5.1) This fact is going to be essential to our analysis. From Ω u ε = 0, u ε changes sign in Ω and we set Z(u ε ) def = {x Ω : u ε (x) = 0}. Next we make the change of dependent variables w = 1 u ε (the subscript ε has been dropped in the notation for w since it is kept fixed throughout the proof) which transforms the above equation into ε w + ε w 1 w + ε 1 w(w 1) = λ w 1 w. (5.) Hence 0 < w 1 and the Lagrange multiplier λ w = Ω w 1 wdx. Note that and also w(x) = 1 x Z(u ε ) w 1 w = 4 u ε. (5.3) One sees from (5.3) and the regularity of u ε that, on the set of H -measure zero {u ε = 0}, the second term in (5.) is finite. Claim 1: w(x) = 1 x = (0,x ), x D. Indeed suppose by contradiction that x = ( x 1, x ) such that w( x) = 1 and x 1 0. We may suppose x 1 > 0. Since 1 u ε > 0 (from (5.1)), by the Implicit Function Theorem, there exists a neighborhood of x D, and a function g C (V δ ) satisfying V δ def = {x : x x < δ} w = 1 on Z δ (u ε ) def = {(g(x ),x ), x V δ } Z(u ε ). From x 1 > 0, g can be taken so that g > 0 and therefore l < g(x ) l, x D. Next we define w(x 1,x ), if g(x ) < x 1 < l, x V δ w(x 1,x ) def = w(g(x ) x 1,x ), if g(x ) l < x 1 < g(x ), x V δ and let D(w) def = {(x 1,x ) : g(x ) l < x 1 < l, x V δ }, D (w) def = {(x 1,x ) : g(x ) l < x 1 < g(x ), x V δ }, D + (w) def = {(x 1,x ) : g(x ) < x 1 < l, x V δ }.

20 0 Arnaldo S. do Nascimento, João Biesdorf. Note that w(x) = 1, x D(w) x = (g(x ),x ), x V δ. Also 1 u ε > 0 on Ω implies that u ε (x 1,x ) = 0, for x V δ, if and only if x 1 = g(x ), i.e., u ε 0 on {(g(x ) l,x ), x V δ }. We deduce Next we define the function 1 w = u ε 1 u ε 0 on {(g(x ) l,x ), x V δ }. (5.4) θ def = w w. Since w w on D + (w) one uses the constraint to verify that λ w = λ w. Bearing in mind that 1 w = 1 w = 0 on Z δ (u ε ) (this is crucial here) one verifies that θ W, (D(w)) and, in addition, is a weak solution to ε θ + ε V (x) θ + ε 1 [U(x) +W(x) + λ w Λ(x)]θ = 0 in D(w) (5.5) where and 1 (t w( ) + (1 t) w( )) V ( ) = 0 1 (tw( ) + (1 t)w( )) dt, U( ) = 1 W( ) = Λ( ) = 1 (t w( ) + (1 t) w( )) [1 (tw( ) + (1 t)w( ))] dt 1 dt [1 (tw( ) + (1 t)w( ))] 1/ tw( ) + (1 t)w( )dt, with U, W, Λ and each component of V being in Lloc (D(w)) (recall that u ε C 3+δ (Ω)). Therefore, since θ W, loc, the Unique Continuation Principle applied to (5.5) yields θ 0 on the account that θ = 0 on D + (w). Therefore w = w on D(w) and given that 1 w(l,x ) = 0, x V δ (recall that 1 u ε (l,x ) = 0, u ε (l,x ) 0 for x V δ and w is C 1 in Ω up to the edges and corners of Ω) we conclude 1 w(g(x ) l,x ) = 0, x V δ. (5.6) But (5.6) contradicts (5.4) and therefore Claim 1 is proved. Claim : It holds that w(x 1,x ) = w( x 1,x ), x Ω. Recalling that Ω l = {x Ω : x 1 < 0} and Ω r = {x Ω : x 1 > 0} we now define w as the reflection with respect to {x 1 = 0} of w restricted to Ω r, i.e., w(x 1,x ) def = w(x 1,x ) on Ω r w( x 1,x ) on Ω l.

21 Global and local minimizers 1 Now, thanks to Claim 1 (which is crucial here), one verifies that w W, (Ω). Going through the same procedure as above we now define θ def = w w in Ω. One easily verifies, using the symmetry of the domain, that θ is a weak solution to (5.5). Again since θ 0 on Ω r and θ W, (Ω) the Unique Continuation Principle implies w = w on Ω. This establishes our claim on the evenness of w with respect to the first variable. The next step is to note that being w = 1 u ε then u ε(x 1,x ) = uε( x 1,x ), which amounts to saying that u ε (x 1,x ) = ±u ε ( x 1,x ), (x 1,x ) Ω. (5.7) Finally in view of (5.1) and (5.7) we conclude (7.i). Lemma 8 Consider the family {u ε } of local minimizers of E ε in M, as in Lemma 7, when (8.i) Ω = ( l,l) ( r,r)(l r) or (8.ii) Ω = ( l,l) ( r,r) ( q,q)(l r q) or (8.iii) Ω = ( l,l) B R (0,0). Then u ε (x 1,x ) = u ε (x 1 ), x Ω, i.e., u ε depends just on the first variable and is odd and increasing in ( l,l). Proof The proofs of the first two cases are analogous therefore just rectangular domains are considered. So suppose by contradiction that u ε / 0. Then, by symmetry, the very same argument used in the proof of Lemma 7 ([], Theorem 3.3) yields u ε < 0 on Ω {x < 0} and u ε > 0 on Ω {x 1 > 0}. But this contradicts the fact, proved in Lemma 7, that u ε < 0 on Ω l = Ω {x 1 < 0} and u ε > 0 on Ω r = Ω {x 1 > 0}. Hence we must have u ε 0. As for the case Ω = ( l,l) B R (0,0) it suffices to note that the property u ε Ωr ( θ) = ±u ε Ωl ( θ) (from Lemma 7) for any 0 < θ < π implies that u ε (x 1,ρ,θ) cannot depend on the angular variable. In other words, u ε (x 1,x,x 3 ) depends on x = (x,x 3 ) only through its absolute value x and then, using that u ε is a conditionally stable critical point of E ε, we can resort to [6], p. 309, to conclude that u ε depends only on its first variable x 1. The oddness and monotonicity have been proved in Lemma 7. In the next two lemmas the question of number as well as characterization of local and global minimizers is addressed. Since the argument for all the three types of domains considered in Lemma 8 is analogous we focus on the simplest case, the rectangular domain, in the hope that the main ideas can be easier enhanced in a simpler setting. In particular, the one-dimension character of minimizers will be used to conclude a uniqueness result. Lemma 9 Let Ω = ( l,l) ( r,r)(l r). Then, for ε small, E ε has four local minimizers in M which depend on just one variable: u ε (x 1 ) and v ε (x ) (as in Theorem 4) and also u ε (x 1 ) and v ε (x ).

22 Arnaldo S. do Nascimento, João Biesdorf. Proof Existence of u ε and of v ε follows as in Theorem 4 by replacing u 0 with u 0 and v 0 with v 0, respectively. In order to obtain the dependence of just one variable in each case we follow the proofs of Lemmas 7 and 8. Lemma 10 Let Ω = R = ( l,l) ( r,r) with l > r. Then (10.i) E ε has only two global minimizers in M : u ε and u ε (as in Lemma 9) (10.ii) E ε has only two other local minimizers v ε and v ε (as in Lemma 9 ) which are increasing and decreasing, respectively, and odd in ( r,r). Proof (10.i) Let U ε be a solution to (1.). Existence of such a global minimizer is obtained as usual via the direct method of Calculus of Variations. Since U ε M, it is not a constant function in R. Claim 1: U ε depends only on one variable. Suppose by contradiction that 1 U ε / 0 and also U ε / 0. Since U ε is also a local minimizer of E ε in M we conclude, as in (5.1), that 1 U ε > 0 on Ω and U ε > 0 on Ω. Then we follow the proofs of Lemma 7 and Lemma 8 (eventually (5.1) may be reversed) to conclude from 1 U ε / 0 that U ε depends only on the first variable and then, from U ε / 0 that U ε depends only on the second variable: a contradiction. Claim : Either U ε (x 1 ) = u ε (x 1 ) or U ε (x 1 ) = u ε (x 1 ). Since, by Claim 1, U ε depends only on one variable let us suppose by contradiction that it depends on the second one, i.e., U ε (x 1,x ) = U ε (x ). Hence, from symmetry, we can follow the proof of Lemma 7 (note that U ε / 0 allows us to use [], Theorem 3.3, so that (5.1) holds) to conclude that U ε is monotone and odd in x. For definiteness we may suppose that U ε (x ) is increasing and compare it with v ε (x ) (given in Lemma 9), as follows. Actually since U ε, v ε are odd functions in ( r,r) and f (u) = u u 3, the corresponding Lagrange multipliers satisfy λ Uε = r r f (U ε )dx = 0 and λ vε = r r f (v ε )dx = 0. Therefore we have { εu ε + ε 1 f (U ε ) = 0, U ε( r) = U ε(r) = 0 in( r,r) (5.8) and { εv ε + ε 1 f (v ε ) = 0, v ε( r) = v ε(r) = 0 in( r,r) (5.9) Thus U ε and v ε are both monotone functions and satisfy, for ε small, the equation ε u + u u 3 = 0, subjected to Neumann boundary condition u ( r) = u (r) = 0. Hence by uniqueness of such solutions (see [7] or [8], p. 3130, for instance) we must have U ε v ε. Therefore v ε would be a global minimizer of E ε in M. Recall from Lemma 9 that v ε is obtained through the same procedure used to obtain u ε and as such v ε v 0 L 1 (R) 0, (5.10)

23 Global and local minimizers 3 where v 0 is given by (3.4). It is also a basic fact in the theory of Γ convergence that since v ε is a global minimizer of E ε then, due to (5.10), v 0 is also a global minimizer of E 0. But this is impossible due to the fact that u 0, given by (3.3), is a admissible function for E 0 and, since l > r, satisfies E 0 (v 0 ) = 8 3 l > 8 3 r = E 0(u 0 ), (5.11) against the fact that v 0 is a global minimizer of E 0. Hence U ε depends only on the first variable x 1 and therefore comparing it with u ε (x 1 ) (both satisfy equations (5.8) and (5.9), respectively, and are monotone) the same uniqueness argument used above yields U ε u ε, for ε small enough. We conclude the proof of Claim after remarking that since E ε (u ε ) = E ε ( u ε ) it follows that u ε is also a solution to (1.). (10.ii) Suppose now that V ε is local minimizer of E ε in M but not a solution to (1.), i.e., not a global minimizer of E ε in M. As in Claim 1 we conclude that V ε depends on just one variable and is monotone. Suppose that it depends on the first one and, for definiteness, suppose that V ε is increasing. Hence comparing it with u ε we would conclude as in Claim (by uniqueness) that V ε u ε and therefore, since u ε is a global minimizer, so would be V ε : a contradiction. Therefore V ε depends just on the second variable. Comparing it with v ε we go through the same procedure used in Claim to finally conclude that V ε (x ) v ε (x ). To complete the proof it suffices to show that necessarily v ε is a also a local minimizer. By hypothesis, δ > 0 and so that V δ (v ε ) def = {v M : v ε v H 1 (Ω) < δ} E ε (v ε ) E ε (v), v V δ (v ε ). (5.1) Claim 3: E ε ( v ε ) E ε (w), w B δ ( v ε ). Indeed, for any w B δ ( v ε ) it holds δ w ( v ε ) H 1 (Ω) = ( w) v ε H 1 (Ω), thus implying that w B δ (v ε ). Since E ε is quadratic functional and w B δ (v ε ), we use (5.1) to conclude E ε ( v ε ) = E ε (v ε ) E ε ( w) = E ε (w). This completes the proof. We are now in condition to complete the proofs of our main results. Proof of Theorem 1 In view of the above lemmas it remains only the case in which the domain is a square to be analyzed. Let u ε and u ε be the global minimizers found in Lemma 10. Since in this case l = r we see that v ε (x 1,x ) = u ε (x,x 1 ), i.e., v ε is obtained from u ε by a rotation of π/ with respect to x 1 -axis. Since E ε (u ε ) = E ε ( u ε ) = E ε (v ε ) = E ε ( v ε ) it follows that all four functions u ε, u ε, v ε and v ε are solutions to (1.) and the only ones (either local or global) by Lemma 10. Proof of Theorem

24 4 Arnaldo S. do Nascimento, João Biesdorf. When Ω = ( l,l) D,with D IR any Lipschitz bounded domain, the proof of existence of a family of local minimizers {u ε } has been established in Theorem 4. From Lemma 7 it is increasing and odd on the first variable. (.i) When D = B R (0,0) the existence of the two local minimizers u ε (x) and u ε (x) follows from Theorem 4. From Lemma 8 we have that u ε (x 1,x ) = u ε (x 1 ) is increasing and odd in ( l,l). It is our goal now to prove that these are the only two solutions to (1.) in the circular cylinder. Claim 1: Any local minimizer U ε of E ε in M satisfies either U ε = u ε or U ε = u ε. Unfortunately the argument used in the previous cases for a rectangle and a parallelepiped does not work for a circular cylinder. So for ϕ M, ϕ / 0 and F(ξ ) = (1/4)(1 ξ ) we define I(U ε ;ϕ) def = ( ϕ + ε F (U ε )ϕ )dx (5.13) Ω Therefore, from our hypothesis, I(U ε ;ϕ) 0, ϕ M. (5.14) Also, letting for simplicity in notation U ε = U, we have, in cylindrical coordinates, U ρρ +U ρ /ρ +U θθ /ρ +U x1 x 1 + ε f (U ε ) + λ U = 0, in( l,l) D U ρ (x 1,R,θ) = 0, for l < x 1 < l, 0 < θ π (5.15) U x1 (±l,ρ,θ) = 0 for 0 < ρ < R, 0 < θ π where λ U is a Lagrange multiplier. By defining V def = U θ one easily verifies that V M, V C (Ω), and V ρρ +V ρ /ρ +V θθ /ρ +V x1 x 1 + ε f (U ε )V = 0, in( l,l) D V ρ (x 1,R,θ) = 0, for l < x 1 < l, 0 < θ π V x1 (±l,ρ,θ) = 0 for 0 < ρ < R, 0 < θ π (5.16) Hence from (5.14) and (5.16) we deduce that I(U;V ) = 0 and, given that V M, I(U;V ) = 0 = min{i(u;ϕ) : ϕ M,ϕ / 0}. It then follows from a variant of Krein-Rutman s Theorem (see, for instance, [], p. 4, for a proof) that V does not change sign in Ω. Therefore, due to the periodicity of U in the angular variable θ, we must have U(x 1,ρ,θ) = U(x 1,ρ). We now evoke a result in [6], p. 308, to finally conclude that being a local minimizer in a domain with rotational symmetry we actually have that U depends just on the first variable. Recovering the sub-script ε, (5.15) becomes { εu ε + ε 1 f (U ε ) + λ Uε = 0, U ε( l) = U ε(l) = 0 in( l,l) (5.17)

25 Global and local minimizers 5 Lemma 7 (for one-dimensional domains) yields U ε (x 1 ) = U ε ( x 1 ) and therefore l λ Uε = f (U ε (x 1 ))dx 1 = 0. l This allows us to utilize the very same argument used in Claim of Lemma 10, for the equations (5.8) and (5.9), to finally conclude that either U ε = u ε or U ε = u ε. Therefore there are only two local minimizers, u ε and u ε, and on the account that E ε (u ε ) = E ε ( u ε ), they must be global, i.e., the only solutions to (1.). (.ii) When Ω = ( l,l) ( r,r) ( q,q) with l > r > q, the proof that (1.) has only two solutions, u ε and u ε, depending only on the first variable with u ε increasing and odd, follows from the same argument used in Lemma 10 since dimension of the domain is not an issue here. We briefly describe the steps. Existence of the other three local minimizers, u ε (x 1 ), v ε (x ) and w ε (x 3 ), follows as in Theorem 4 by replacing u 0, v 0 and w 0 with u 0, v 0 and w 0, respectively. To assure that any solution to (1.), say U ε, depends only on one variable we argue as in Claim 1 of Lemma 10 except that now we must consider the possibilities 1. 1 U ε / 0, U ε / 0 and 3 U ε / 0, and. 1 U ε / 0 and U ε / 0 or 3. 1 U ε / 0 and 3 U ε / 0 or 4. U ε / 0 and 3 U ε / 0. Next we prove that either U ε (x 1 ) = u ε (x 1 ) or U ε (x 1 ) = u ε (x 1 ). Since U ε depends only on one variable, suppose first that it depends on the second one. Since U ε is a global minimizer, as in the rectangular domain we conclude that U ε is monotone and, for instance, increasing ( r,r). Following the same argument utilized in Lemma 10, Claim, we obtain, by uniqueness, that U ε v ε. Hence v ε would be a global minimizer of E ε over M and, by the Γ convergence property, v 0 would be a global minimizer of E 0. Hence since u 0 is a admissible function and l > r > q it would hold that E 0 (u 0 ) = 8 3 rq < 8 3 ql = E 0(v 0 ). (5.18) But then (5.18) would contradict the fact that v 0 is a global minimizer. Therefore U ε cannot depend on the second variable. Likewise we conclude that U ε cannot depend on the third variable. Therefore, since U ε is not a constant function, U ε must depend only on its first variable. Had we supposed that U ε depended on its third variable then the procedure would have been the same except that now U ε would have been compared to w ε (x 3 ), if U ε is supposed increasing, and to w ε (x 3 ) if it were decreasing. Having established that U ε depends only on its first variable we again utilize an argument as in Lemma 10, Claim, to obtain, by uniqueness, that U ε u ε. Therefore any solution to (1.) must be either u ε or u ε.

THE STEINER REARRANGEMENT IN ANY CODIMENSION

THE STEINER REARRANGEMENT IN ANY CODIMENSION THE STEINER REARRANGEMENT IN ANY CODIMENSION GIUSEPPE MARIA CAPRIANI Abstract. We analyze the Steiner rearrangement in any codimension of Sobolev and BV functions. In particular, we prove a Pólya-Szegő

More information

Slow motion for the nonlocal Allen Cahn equation in n-dimensions

Slow motion for the nonlocal Allen Cahn equation in n-dimensions Slow motion for the nonlocal Allen Cahn equation in n-dimensions Ryan Murray Carnegie Mellon University Pittsburgh, PA, USA Matteo Rinaldi Carnegie Mellon University Pittsburgh, PA, USA December 4, 215

More information

P(E t, Ω)dt, (2) 4t has an advantage with respect. to the compactly supported mollifiers, i.e., the function W (t)f satisfies a semigroup law:

P(E t, Ω)dt, (2) 4t has an advantage with respect. to the compactly supported mollifiers, i.e., the function W (t)f satisfies a semigroup law: Introduction Functions of bounded variation, usually denoted by BV, have had and have an important role in several problems of calculus of variations. The main features that make BV functions suitable

More information

MINIMAL GRAPHS PART I: EXISTENCE OF LIPSCHITZ WEAK SOLUTIONS TO THE DIRICHLET PROBLEM WITH C 2 BOUNDARY DATA

MINIMAL GRAPHS PART I: EXISTENCE OF LIPSCHITZ WEAK SOLUTIONS TO THE DIRICHLET PROBLEM WITH C 2 BOUNDARY DATA MINIMAL GRAPHS PART I: EXISTENCE OF LIPSCHITZ WEAK SOLUTIONS TO THE DIRICHLET PROBLEM WITH C 2 BOUNDARY DATA SPENCER HUGHES In these notes we prove that for any given smooth function on the boundary of

More information

+ 2x sin x. f(b i ) f(a i ) < ɛ. i=1. i=1

+ 2x sin x. f(b i ) f(a i ) < ɛ. i=1. i=1 Appendix To understand weak derivatives and distributional derivatives in the simplest context of functions of a single variable, we describe without proof some results from real analysis (see [7] and

More information

Minimal Surface equations non-solvability strongly convex functional further regularity Consider minimal surface equation.

Minimal Surface equations non-solvability strongly convex functional further regularity Consider minimal surface equation. Lecture 7 Minimal Surface equations non-solvability strongly convex functional further regularity Consider minimal surface equation div + u = ϕ on ) = 0 in The solution is a critical point or the minimizer

More information

Laplace s Equation. Chapter Mean Value Formulas

Laplace s Equation. Chapter Mean Value Formulas Chapter 1 Laplace s Equation Let be an open set in R n. A function u C 2 () is called harmonic in if it satisfies Laplace s equation n (1.1) u := D ii u = 0 in. i=1 A function u C 2 () is called subharmonic

More information

Pseudo-Poincaré Inequalities and Applications to Sobolev Inequalities

Pseudo-Poincaré Inequalities and Applications to Sobolev Inequalities Pseudo-Poincaré Inequalities and Applications to Sobolev Inequalities Laurent Saloff-Coste Abstract Most smoothing procedures are via averaging. Pseudo-Poincaré inequalities give a basic L p -norm control

More information

Calculus of Variations. Final Examination

Calculus of Variations. Final Examination Université Paris-Saclay M AMS and Optimization January 18th, 018 Calculus of Variations Final Examination Duration : 3h ; all kind of paper documents (notes, books...) are authorized. The total score of

More information

Analysis in weighted spaces : preliminary version

Analysis in weighted spaces : preliminary version Analysis in weighted spaces : preliminary version Frank Pacard To cite this version: Frank Pacard. Analysis in weighted spaces : preliminary version. 3rd cycle. Téhéran (Iran, 2006, pp.75.

More information

SYMMETRY RESULTS FOR PERTURBED PROBLEMS AND RELATED QUESTIONS. Massimo Grosi Filomena Pacella S. L. Yadava. 1. Introduction

SYMMETRY RESULTS FOR PERTURBED PROBLEMS AND RELATED QUESTIONS. Massimo Grosi Filomena Pacella S. L. Yadava. 1. Introduction Topological Methods in Nonlinear Analysis Journal of the Juliusz Schauder Center Volume 21, 2003, 211 226 SYMMETRY RESULTS FOR PERTURBED PROBLEMS AND RELATED QUESTIONS Massimo Grosi Filomena Pacella S.

More information

(NON)UNIQUENESS OF MINIMIZERS IN THE LEAST GRADIENT PROBLEM

(NON)UNIQUENESS OF MINIMIZERS IN THE LEAST GRADIENT PROBLEM (NON)UNIQUENESS OF MINIMIZERS IN THE LEAST GRADIENT PROBLEM WOJCIECH GÓRNY arxiv:1709.02185v1 [math.ap] 7 Sep 2017 Abstract. Minimizers in the least gradient problem with discontinuous boundary data need

More information

Domains in metric measure spaces with boundary of positive mean curvature, and the Dirichlet problem for functions of least gradient

Domains in metric measure spaces with boundary of positive mean curvature, and the Dirichlet problem for functions of least gradient Domains in metric measure spaces with boundary of positive mean curvature, and the Dirichlet problem for functions of least gradient Panu Lahti Lukáš Malý Nageswari Shanmugalingam Gareth Speight June 22,

More information

Problem List MATH 5143 Fall, 2013

Problem List MATH 5143 Fall, 2013 Problem List MATH 5143 Fall, 2013 On any problem you may use the result of any previous problem (even if you were not able to do it) and any information given in class up to the moment the problem was

More information

3 (Due ). Let A X consist of points (x, y) such that either x or y is a rational number. Is A measurable? What is its Lebesgue measure?

3 (Due ). Let A X consist of points (x, y) such that either x or y is a rational number. Is A measurable? What is its Lebesgue measure? MA 645-4A (Real Analysis), Dr. Chernov Homework assignment 1 (Due ). Show that the open disk x 2 + y 2 < 1 is a countable union of planar elementary sets. Show that the closed disk x 2 + y 2 1 is a countable

More information

2 A Model, Harmonic Map, Problem

2 A Model, Harmonic Map, Problem ELLIPTIC SYSTEMS JOHN E. HUTCHINSON Department of Mathematics School of Mathematical Sciences, A.N.U. 1 Introduction Elliptic equations model the behaviour of scalar quantities u, such as temperature or

More information

Coupled second order singular perturbations for phase transitions

Coupled second order singular perturbations for phase transitions Coupled second order singular perturbations for phase transitions CMU 06/09/11 Ana Cristina Barroso, Margarida Baía, Milena Chermisi, JM Introduction Let Ω R d with Lipschitz boundary ( container ) and

More information

Partial Differential Equations

Partial Differential Equations Part II Partial Differential Equations Year 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2015 Paper 4, Section II 29E Partial Differential Equations 72 (a) Show that the Cauchy problem for u(x,

More information

Mid Term-1 : Practice problems

Mid Term-1 : Practice problems Mid Term-1 : Practice problems These problems are meant only to provide practice; they do not necessarily reflect the difficulty level of the problems in the exam. The actual exam problems are likely to

More information

Some Background Material

Some Background Material Chapter 1 Some Background Material In the first chapter, we present a quick review of elementary - but important - material as a way of dipping our toes in the water. This chapter also introduces important

More information

OPTIMAL POTENTIALS FOR SCHRÖDINGER OPERATORS. 1. Introduction In this paper we consider optimization problems of the form. min F (V ) : V V, (1.

OPTIMAL POTENTIALS FOR SCHRÖDINGER OPERATORS. 1. Introduction In this paper we consider optimization problems of the form. min F (V ) : V V, (1. OPTIMAL POTENTIALS FOR SCHRÖDINGER OPERATORS G. BUTTAZZO, A. GEROLIN, B. RUFFINI, AND B. VELICHKOV Abstract. We consider the Schrödinger operator + V (x) on H 0 (), where is a given domain of R d. Our

More information

Fractional Perimeter and Nonlocal Minimal Surfaces

Fractional Perimeter and Nonlocal Minimal Surfaces arxiv:58.64v [math.ap] 5 Aug 5 Corso di Laurea Magistrale in Matematica Fractional Perimeter and Nonlocal Minimal Surfaces Relatore: Prof. Enrico VALDINOCI TESI DI LAUREA DI Luca LOMBARDINI Matr. 896 ANNO

More information

2 (Bonus). Let A X consist of points (x, y) such that either x or y is a rational number. Is A measurable? What is its Lebesgue measure?

2 (Bonus). Let A X consist of points (x, y) such that either x or y is a rational number. Is A measurable? What is its Lebesgue measure? MA 645-4A (Real Analysis), Dr. Chernov Homework assignment 1 (Due 9/5). Prove that every countable set A is measurable and µ(a) = 0. 2 (Bonus). Let A consist of points (x, y) such that either x or y is

More information

The optimal partial transport problem

The optimal partial transport problem The optimal partial transport problem Alessio Figalli Abstract Given two densities f and g, we consider the problem of transporting a fraction m [0, min{ f L 1, g L 1}] of the mass of f onto g minimizing

More information

Remarks on the Gauss-Green Theorem. Michael Taylor

Remarks on the Gauss-Green Theorem. Michael Taylor Remarks on the Gauss-Green Theorem Michael Taylor Abstract. These notes cover material related to the Gauss-Green theorem that was developed for work with S. Hofmann and M. Mitrea, which appeared in [HMT].

More information

Analysis Finite and Infinite Sets The Real Numbers The Cantor Set

Analysis Finite and Infinite Sets The Real Numbers The Cantor Set Analysis Finite and Infinite Sets Definition. An initial segment is {n N n n 0 }. Definition. A finite set can be put into one-to-one correspondence with an initial segment. The empty set is also considered

More information

SHARP BOUNDARY TRACE INEQUALITIES. 1. Introduction

SHARP BOUNDARY TRACE INEQUALITIES. 1. Introduction SHARP BOUNDARY TRACE INEQUALITIES GILES AUCHMUTY Abstract. This paper describes sharp inequalities for the trace of Sobolev functions on the boundary of a bounded region R N. The inequalities bound (semi-)norms

More information

for all x,y [a,b]. The Lipschitz constant of f is the infimum of constants C with this property.

for all x,y [a,b]. The Lipschitz constant of f is the infimum of constants C with this property. viii 3.A. FUNCTIONS 77 Appendix In this appendix, we describe without proof some results from real analysis which help to understand weak and distributional derivatives in the simplest context of functions

More information

The Skorokhod problem in a time-dependent interval

The Skorokhod problem in a time-dependent interval The Skorokhod problem in a time-dependent interval Krzysztof Burdzy, Weining Kang and Kavita Ramanan University of Washington and Carnegie Mellon University Abstract: We consider the Skorokhod problem

More information

PHASE TRANSITIONS: REGULARITY OF FLAT LEVEL SETS

PHASE TRANSITIONS: REGULARITY OF FLAT LEVEL SETS PHASE TRANSITIONS: REGULARITY OF FLAT LEVEL SETS OVIDIU SAVIN Abstract. We consider local minimizers of the Ginzburg-Landau energy functional 2 u 2 + 4 ( u2 ) 2 dx and prove that, if the level set is included

More information

FRANK MORGAN AND ANTONIO ROS

FRANK MORGAN AND ANTONIO ROS STABLE CONSTANT MEAN CURVATURE HYPERSURFACES ARE AREA MINIMIZING IN SMALL L 1 NEIGHBORHOODS arxiv:0811.3126v1 [math.dg] 19 Nov 2008 FRANK MORGAN AND ANTONIO ROS Abstract.- We prove that a strictly stable

More information

Minimization problems on the Hardy-Sobolev inequality

Minimization problems on the Hardy-Sobolev inequality manuscript No. (will be inserted by the editor) Minimization problems on the Hardy-Sobolev inequality Masato Hashizume Received: date / Accepted: date Abstract We study minimization problems on Hardy-Sobolev

More information

The heat equation in time dependent domains with Neumann boundary conditions

The heat equation in time dependent domains with Neumann boundary conditions The heat equation in time dependent domains with Neumann boundary conditions Chris Burdzy Zhen-Qing Chen John Sylvester Abstract We study the heat equation in domains in R n with insulated fast moving

More information

Contents: 1. Minimization. 2. The theorem of Lions-Stampacchia for variational inequalities. 3. Γ -Convergence. 4. Duality mapping.

Contents: 1. Minimization. 2. The theorem of Lions-Stampacchia for variational inequalities. 3. Γ -Convergence. 4. Duality mapping. Minimization Contents: 1. Minimization. 2. The theorem of Lions-Stampacchia for variational inequalities. 3. Γ -Convergence. 4. Duality mapping. 1 Minimization A Topological Result. Let S be a topological

More information

An introduction to Mathematical Theory of Control

An introduction to Mathematical Theory of Control An introduction to Mathematical Theory of Control Vasile Staicu University of Aveiro UNICA, May 2018 Vasile Staicu (University of Aveiro) An introduction to Mathematical Theory of Control UNICA, May 2018

More information

Global minimization. Chapter Upper and lower bounds

Global minimization. Chapter Upper and lower bounds Chapter 1 Global minimization The issues related to the behavior of global minimization problems along a sequence of functionals F are by now well understood, and mainly rely on the concept of -limit.

More information

NONLINEAR FREDHOLM ALTERNATIVE FOR THE p-laplacian UNDER NONHOMOGENEOUS NEUMANN BOUNDARY CONDITION

NONLINEAR FREDHOLM ALTERNATIVE FOR THE p-laplacian UNDER NONHOMOGENEOUS NEUMANN BOUNDARY CONDITION Electronic Journal of Differential Equations, Vol. 2016 (2016), No. 210, pp. 1 7. ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu NONLINEAR FREDHOLM ALTERNATIVE FOR THE p-laplacian

More information

VISCOSITY SOLUTIONS. We follow Han and Lin, Elliptic Partial Differential Equations, 5.

VISCOSITY SOLUTIONS. We follow Han and Lin, Elliptic Partial Differential Equations, 5. VISCOSITY SOLUTIONS PETER HINTZ We follow Han and Lin, Elliptic Partial Differential Equations, 5. 1. Motivation Throughout, we will assume that Ω R n is a bounded and connected domain and that a ij C(Ω)

More information

Robustness for a Liouville type theorem in exterior domains

Robustness for a Liouville type theorem in exterior domains Robustness for a Liouville type theorem in exterior domains Juliette Bouhours 1 arxiv:1207.0329v3 [math.ap] 24 Oct 2014 1 UPMC Univ Paris 06, UMR 7598, Laboratoire Jacques-Louis Lions, F-75005, Paris,

More information

Maths 212: Homework Solutions

Maths 212: Homework Solutions Maths 212: Homework Solutions 1. The definition of A ensures that x π for all x A, so π is an upper bound of A. To show it is the least upper bound, suppose x < π and consider two cases. If x < 1, then

More information

Hausdorff Measure. Jimmy Briggs and Tim Tyree. December 3, 2016

Hausdorff Measure. Jimmy Briggs and Tim Tyree. December 3, 2016 Hausdorff Measure Jimmy Briggs and Tim Tyree December 3, 2016 1 1 Introduction In this report, we explore the the measurement of arbitrary subsets of the metric space (X, ρ), a topological space X along

More information

The Caratheodory Construction of Measures

The Caratheodory Construction of Measures Chapter 5 The Caratheodory Construction of Measures Recall how our construction of Lebesgue measure in Chapter 2 proceeded from an initial notion of the size of a very restricted class of subsets of R,

More information

Taylor and Laurent Series

Taylor and Laurent Series Chapter 4 Taylor and Laurent Series 4.. Taylor Series 4... Taylor Series for Holomorphic Functions. In Real Analysis, the Taylor series of a given function f : R R is given by: f (x + f (x (x x + f (x

More information

SCALE INVARIANT FOURIER RESTRICTION TO A HYPERBOLIC SURFACE

SCALE INVARIANT FOURIER RESTRICTION TO A HYPERBOLIC SURFACE SCALE INVARIANT FOURIER RESTRICTION TO A HYPERBOLIC SURFACE BETSY STOVALL Abstract. This result sharpens the bilinear to linear deduction of Lee and Vargas for extension estimates on the hyperbolic paraboloid

More information

REGULARITY FOR INFINITY HARMONIC FUNCTIONS IN TWO DIMENSIONS

REGULARITY FOR INFINITY HARMONIC FUNCTIONS IN TWO DIMENSIONS C,α REGULARITY FOR INFINITY HARMONIC FUNCTIONS IN TWO DIMENSIONS LAWRENCE C. EVANS AND OVIDIU SAVIN Abstract. We propose a new method for showing C,α regularity for solutions of the infinity Laplacian

More information

Singular Perturbations of Stochastic Control Problems with Unbounded Fast Variables

Singular Perturbations of Stochastic Control Problems with Unbounded Fast Variables Singular Perturbations of Stochastic Control Problems with Unbounded Fast Variables Joao Meireles joint work with Martino Bardi and Guy Barles University of Padua, Italy Workshop "New Perspectives in Optimal

More information

Γ -convergence of the Allen Cahn energy with an oscillating forcing term

Γ -convergence of the Allen Cahn energy with an oscillating forcing term Interfaces and Free Boundaries 8 (2006), 47 78 Γ -convergence of the Allen Cahn energy with an oscillating forcing term N. DIRR Max-Planck Institute for Mathematics in the Sciences, Inselstr. 22, D-04103

More information

Some lecture notes for Math 6050E: PDEs, Fall 2016

Some lecture notes for Math 6050E: PDEs, Fall 2016 Some lecture notes for Math 65E: PDEs, Fall 216 Tianling Jin December 1, 216 1 Variational methods We discuss an example of the use of variational methods in obtaining existence of solutions. Theorem 1.1.

More information

Integration on Measure Spaces

Integration on Measure Spaces Chapter 3 Integration on Measure Spaces In this chapter we introduce the general notion of a measure on a space X, define the class of measurable functions, and define the integral, first on a class of

More information

The Isoperimetric Problem in the Heisenberg group H n

The Isoperimetric Problem in the Heisenberg group H n The Isoperimetric Problem in the Heisenberg group H n First Taiwan Geometry Symposium, NCTS South November 20, 2010 First Taiwan Geometry Symposium, NCTS South The () Isoperimetric Problem in the Heisenberg

More information

ON WEAKLY NONLINEAR BACKWARD PARABOLIC PROBLEM

ON WEAKLY NONLINEAR BACKWARD PARABOLIC PROBLEM ON WEAKLY NONLINEAR BACKWARD PARABOLIC PROBLEM OLEG ZUBELEVICH DEPARTMENT OF MATHEMATICS THE BUDGET AND TREASURY ACADEMY OF THE MINISTRY OF FINANCE OF THE RUSSIAN FEDERATION 7, ZLATOUSTINSKY MALIY PER.,

More information

Math The Laplacian. 1 Green s Identities, Fundamental Solution

Math The Laplacian. 1 Green s Identities, Fundamental Solution Math. 209 The Laplacian Green s Identities, Fundamental Solution Let be a bounded open set in R n, n 2, with smooth boundary. The fact that the boundary is smooth means that at each point x the external

More information

POINTWISE BOUNDS ON QUASIMODES OF SEMICLASSICAL SCHRÖDINGER OPERATORS IN DIMENSION TWO

POINTWISE BOUNDS ON QUASIMODES OF SEMICLASSICAL SCHRÖDINGER OPERATORS IN DIMENSION TWO POINTWISE BOUNDS ON QUASIMODES OF SEMICLASSICAL SCHRÖDINGER OPERATORS IN DIMENSION TWO HART F. SMITH AND MACIEJ ZWORSKI Abstract. We prove optimal pointwise bounds on quasimodes of semiclassical Schrödinger

More information

Multi-Phase Optimal Partitions

Multi-Phase Optimal Partitions Multi-Phase Optimal Partitions Functions of Bounded Variation Candidate: Supervisors: Beniamin Bogoşel Prof. Dorin Bucur Prof. Édouard Oudet Chambéry 2012 Contents 1 Introduction 4 2 Existence results

More information

NECESSARY CONDITIONS FOR WEIGHTED POINTWISE HARDY INEQUALITIES

NECESSARY CONDITIONS FOR WEIGHTED POINTWISE HARDY INEQUALITIES NECESSARY CONDITIONS FOR WEIGHTED POINTWISE HARDY INEQUALITIES JUHA LEHRBÄCK Abstract. We establish necessary conditions for domains Ω R n which admit the pointwise (p, β)-hardy inequality u(x) Cd Ω(x)

More information

n [ F (b j ) F (a j ) ], n j=1(a j, b j ] E (4.1)

n [ F (b j ) F (a j ) ], n j=1(a j, b j ] E (4.1) 1.4. CONSTRUCTION OF LEBESGUE-STIELTJES MEASURES In this section we shall put to use the Carathéodory-Hahn theory, in order to construct measures with certain desirable properties first on the real line

More information

Equilibria with a nontrivial nodal set and the dynamics of parabolic equations on symmetric domains

Equilibria with a nontrivial nodal set and the dynamics of parabolic equations on symmetric domains Equilibria with a nontrivial nodal set and the dynamics of parabolic equations on symmetric domains J. Földes Department of Mathematics, Univerité Libre de Bruxelles 1050 Brussels, Belgium P. Poláčik School

More information

1 Lyapunov theory of stability

1 Lyapunov theory of stability M.Kawski, APM 581 Diff Equns Intro to Lyapunov theory. November 15, 29 1 1 Lyapunov theory of stability Introduction. Lyapunov s second (or direct) method provides tools for studying (asymptotic) stability

More information

Recall that if X is a compact metric space, C(X), the space of continuous (real-valued) functions on X, is a Banach space with the norm

Recall that if X is a compact metric space, C(X), the space of continuous (real-valued) functions on X, is a Banach space with the norm Chapter 13 Radon Measures Recall that if X is a compact metric space, C(X), the space of continuous (real-valued) functions on X, is a Banach space with the norm (13.1) f = sup x X f(x). We want to identify

More information

Estimates for probabilities of independent events and infinite series

Estimates for probabilities of independent events and infinite series Estimates for probabilities of independent events and infinite series Jürgen Grahl and Shahar evo September 9, 06 arxiv:609.0894v [math.pr] 8 Sep 06 Abstract This paper deals with finite or infinite sequences

More information

8 Nonlinear change of variables

8 Nonlinear change of variables Tel Aviv University, 206 Analysis-III 98 8 Nonlinear change of variables 8a Introduction....................... 98 8b Examples......................... 00 8c Measure 0 is preserved................. 02

More information

Introduction to Algebraic and Geometric Topology Week 3

Introduction to Algebraic and Geometric Topology Week 3 Introduction to Algebraic and Geometric Topology Week 3 Domingo Toledo University of Utah Fall 2017 Lipschitz Maps I Recall f :(X, d)! (X 0, d 0 ) is Lipschitz iff 9C > 0 such that d 0 (f (x), f (y)) apple

More information

Regularity of flat level sets in phase transitions

Regularity of flat level sets in phase transitions Annals of Mathematics, 69 (2009), 4 78 Regularity of flat level sets in phase transitions By Ovidiu Savin Abstract We consider local minimizers of the Ginzburg-Landau energy functional 2 u 2 + 4 ( u2 )

More information

Radial Symmetry of Minimizers for Some Translation and Rotation Invariant Functionals

Radial Symmetry of Minimizers for Some Translation and Rotation Invariant Functionals journal of differential equations 124, 378388 (1996) article no. 0015 Radial Symmetry of Minimizers for Some Translation and Rotation Invariant Functionals Orlando Lopes IMECCUNICAMPCaixa Postal 1170 13081-970,

More information

On the definitions of Sobolev and BV spaces into singular spaces and the trace problem

On the definitions of Sobolev and BV spaces into singular spaces and the trace problem On the definitions of Sobolev and BV spaces into singular spaces and the trace problem David CHIRON Laboratoire J.A. DIEUDONNE, Université de Nice - Sophia Antipolis, Parc Valrose, 68 Nice Cedex 2, France

More information

at time t, in dimension d. The index i varies in a countable set I. We call configuration the family, denoted generically by Φ: U (x i (t) x j (t))

at time t, in dimension d. The index i varies in a countable set I. We call configuration the family, denoted generically by Φ: U (x i (t) x j (t)) Notations In this chapter we investigate infinite systems of interacting particles subject to Newtonian dynamics Each particle is characterized by its position an velocity x i t, v i t R d R d at time

More information

02. Measure and integral. 1. Borel-measurable functions and pointwise limits

02. Measure and integral. 1. Borel-measurable functions and pointwise limits (October 3, 2017) 02. Measure and integral Paul Garrett garrett@math.umn.edu http://www.math.umn.edu/ garrett/ [This document is http://www.math.umn.edu/ garrett/m/real/notes 2017-18/02 measure and integral.pdf]

More information

NOTES ON THE REGULARITY OF QUASICONFORMAL HOMEOMORPHISMS

NOTES ON THE REGULARITY OF QUASICONFORMAL HOMEOMORPHISMS NOTES ON THE REGULARITY OF QUASICONFORMAL HOMEOMORPHISMS CLARK BUTLER. Introduction The purpose of these notes is to give a self-contained proof of the following theorem, Theorem.. Let f : S n S n be a

More information

Lebesgue Measure on R n

Lebesgue Measure on R n CHAPTER 2 Lebesgue Measure on R n Our goal is to construct a notion of the volume, or Lebesgue measure, of rather general subsets of R n that reduces to the usual volume of elementary geometrical sets

More information

Perron method for the Dirichlet problem.

Perron method for the Dirichlet problem. Introduzione alle equazioni alle derivate parziali, Laurea Magistrale in Matematica Perron method for the Dirichlet problem. We approach the question of existence of solution u C (Ω) C(Ω) of the Dirichlet

More information

REACTION-DIFFUSION EQUATIONS FOR POPULATION DYNAMICS WITH FORCED SPEED II - CYLINDRICAL-TYPE DOMAINS. Henri Berestycki and Luca Rossi

REACTION-DIFFUSION EQUATIONS FOR POPULATION DYNAMICS WITH FORCED SPEED II - CYLINDRICAL-TYPE DOMAINS. Henri Berestycki and Luca Rossi Manuscript submitted to Website: http://aimsciences.org AIMS Journals Volume 00, Number 0, Xxxx XXXX pp. 000 000 REACTION-DIFFUSION EQUATIONS FOR POPULATION DYNAMICS WITH FORCED SPEED II - CYLINDRICAL-TYPE

More information

THE INVERSE FUNCTION THEOREM FOR LIPSCHITZ MAPS

THE INVERSE FUNCTION THEOREM FOR LIPSCHITZ MAPS THE INVERSE FUNCTION THEOREM FOR LIPSCHITZ MAPS RALPH HOWARD DEPARTMENT OF MATHEMATICS UNIVERSITY OF SOUTH CAROLINA COLUMBIA, S.C. 29208, USA HOWARD@MATH.SC.EDU Abstract. This is an edited version of a

More information

Foliations of hyperbolic space by constant mean curvature surfaces sharing ideal boundary

Foliations of hyperbolic space by constant mean curvature surfaces sharing ideal boundary Foliations of hyperbolic space by constant mean curvature surfaces sharing ideal boundary David Chopp and John A. Velling December 1, 2003 Abstract Let γ be a Jordan curve in S 2, considered as the ideal

More information

PARTIAL REGULARITY OF BRENIER SOLUTIONS OF THE MONGE-AMPÈRE EQUATION

PARTIAL REGULARITY OF BRENIER SOLUTIONS OF THE MONGE-AMPÈRE EQUATION PARTIAL REGULARITY OF BRENIER SOLUTIONS OF THE MONGE-AMPÈRE EQUATION ALESSIO FIGALLI AND YOUNG-HEON KIM Abstract. Given Ω, Λ R n two bounded open sets, and f and g two probability densities concentrated

More information

and BV loc R N ; R d)

and BV loc R N ; R d) Necessary and sufficient conditions for the chain rule in W 1,1 loc R N ; R d) and BV loc R N ; R d) Giovanni Leoni Massimiliano Morini July 25, 2005 Abstract In this paper we prove necessary and sufficient

More information

A note on W 1,p estimates for quasilinear parabolic equations

A note on W 1,p estimates for quasilinear parabolic equations 200-Luminy conference on Quasilinear Elliptic and Parabolic Equations and Systems, Electronic Journal of Differential Equations, Conference 08, 2002, pp 2 3. http://ejde.math.swt.edu or http://ejde.math.unt.edu

More information

A FABER-KRAHN INEQUALITY FOR THE CHEEGER CONSTANT OF N-GONS

A FABER-KRAHN INEQUALITY FOR THE CHEEGER CONSTANT OF N-GONS A FABER-KRAHN INEQUALITY FOR THE CHEEGER CONSTANT OF N-GONS DORIN BUCUR, ILARIA FRAGALÀ Abstract. We prove that the regular N-gon minimizes the Cheeger constant among polygons with a given area and N sides.

More information

A REMARK ON MINIMAL NODAL SOLUTIONS OF AN ELLIPTIC PROBLEM IN A BALL. Olaf Torné. 1. Introduction

A REMARK ON MINIMAL NODAL SOLUTIONS OF AN ELLIPTIC PROBLEM IN A BALL. Olaf Torné. 1. Introduction Topological Methods in Nonlinear Analysis Journal of the Juliusz Schauder Center Volume 24, 2004, 199 207 A REMARK ON MINIMAL NODAL SOLUTIONS OF AN ELLIPTIC PROBLEM IN A BALL Olaf Torné (Submitted by Michel

More information

Metric Spaces and Topology

Metric Spaces and Topology Chapter 2 Metric Spaces and Topology From an engineering perspective, the most important way to construct a topology on a set is to define the topology in terms of a metric on the set. This approach underlies

More information

2. Function spaces and approximation

2. Function spaces and approximation 2.1 2. Function spaces and approximation 2.1. The space of test functions. Notation and prerequisites are collected in Appendix A. Let Ω be an open subset of R n. The space C0 (Ω), consisting of the C

More information

Topology. Xiaolong Han. Department of Mathematics, California State University, Northridge, CA 91330, USA address:

Topology. Xiaolong Han. Department of Mathematics, California State University, Northridge, CA 91330, USA  address: Topology Xiaolong Han Department of Mathematics, California State University, Northridge, CA 91330, USA E-mail address: Xiaolong.Han@csun.edu Remark. You are entitled to a reward of 1 point toward a homework

More information

Lecture Notes in Advanced Calculus 1 (80315) Raz Kupferman Institute of Mathematics The Hebrew University

Lecture Notes in Advanced Calculus 1 (80315) Raz Kupferman Institute of Mathematics The Hebrew University Lecture Notes in Advanced Calculus 1 (80315) Raz Kupferman Institute of Mathematics The Hebrew University February 7, 2007 2 Contents 1 Metric Spaces 1 1.1 Basic definitions...........................

More information

LECTURE 15: COMPLETENESS AND CONVEXITY

LECTURE 15: COMPLETENESS AND CONVEXITY LECTURE 15: COMPLETENESS AND CONVEXITY 1. The Hopf-Rinow Theorem Recall that a Riemannian manifold (M, g) is called geodesically complete if the maximal defining interval of any geodesic is R. On the other

More information

A LOCALIZATION PROPERTY AT THE BOUNDARY FOR MONGE-AMPERE EQUATION

A LOCALIZATION PROPERTY AT THE BOUNDARY FOR MONGE-AMPERE EQUATION A LOCALIZATION PROPERTY AT THE BOUNDARY FOR MONGE-AMPERE EQUATION O. SAVIN. Introduction In this paper we study the geometry of the sections for solutions to the Monge- Ampere equation det D 2 u = f, u

More information

Construction of a general measure structure

Construction of a general measure structure Chapter 4 Construction of a general measure structure We turn to the development of general measure theory. The ingredients are a set describing the universe of points, a class of measurable subsets along

More information

Course 212: Academic Year Section 1: Metric Spaces

Course 212: Academic Year Section 1: Metric Spaces Course 212: Academic Year 1991-2 Section 1: Metric Spaces D. R. Wilkins Contents 1 Metric Spaces 3 1.1 Distance Functions and Metric Spaces............. 3 1.2 Convergence and Continuity in Metric Spaces.........

More information

Numerical Approximation of Phase Field Models

Numerical Approximation of Phase Field Models Numerical Approximation of Phase Field Models Lecture 2: Allen Cahn and Cahn Hilliard Equations with Smooth Potentials Robert Nürnberg Department of Mathematics Imperial College London TUM Summer School

More information

Rose-Hulman Undergraduate Mathematics Journal

Rose-Hulman Undergraduate Mathematics Journal Rose-Hulman Undergraduate Mathematics Journal Volume 17 Issue 1 Article 5 Reversing A Doodle Bryan A. Curtis Metropolitan State University of Denver Follow this and additional works at: http://scholar.rose-hulman.edu/rhumj

More information

1.2 Fundamental Theorems of Functional Analysis

1.2 Fundamental Theorems of Functional Analysis 1.2 Fundamental Theorems of Functional Analysis 15 Indeed, h = ω ψ ωdx is continuous compactly supported with R hdx = 0 R and thus it has a unique compactly supported primitive. Hence fφ dx = f(ω ψ ωdy)dx

More information

Introduction to Topology

Introduction to Topology Introduction to Topology Randall R. Holmes Auburn University Typeset by AMS-TEX Chapter 1. Metric Spaces 1. Definition and Examples. As the course progresses we will need to review some basic notions about

More information

Partial Differential Equations, 2nd Edition, L.C.Evans The Calculus of Variations

Partial Differential Equations, 2nd Edition, L.C.Evans The Calculus of Variations Partial Differential Equations, 2nd Edition, L.C.Evans Chapter 8 The Calculus of Variations Yung-Hsiang Huang 2018.03.25 Notation: denotes a bounded smooth, open subset of R n. All given functions are

More information

arxiv: v2 [math.mg] 29 Sep 2015

arxiv: v2 [math.mg] 29 Sep 2015 ON BODIES WITH DIRECTLY CONGRUENT PROJECTIONS AND SECTIONS arxiv:1412.2727v2 [math.mg] 29 Sep 2015 M. ANGELES ALFONSECA, MICHELLE CORDIER, AND DMITRY RYABOGIN Abstract. Let K and L be two convex bodies

More information

Topological vectorspaces

Topological vectorspaces (July 25, 2011) Topological vectorspaces Paul Garrett garrett@math.umn.edu http://www.math.umn.edu/ garrett/ Natural non-fréchet spaces Topological vector spaces Quotients and linear maps More topological

More information

GIOVANNI COMI AND MONICA TORRES

GIOVANNI COMI AND MONICA TORRES ONE-SIDED APPROXIMATION OF SETS OF FINITE PERIMETER GIOVANNI COMI AND MONICA TORRES Abstract. In this note we present a new proof of a one-sided approximation of sets of finite perimeter introduced in

More information

Local semiconvexity of Kantorovich potentials on non-compact manifolds

Local semiconvexity of Kantorovich potentials on non-compact manifolds Local semiconvexity of Kantorovich potentials on non-compact manifolds Alessio Figalli, Nicola Gigli Abstract We prove that any Kantorovich potential for the cost function c = d / on a Riemannian manifold

More information

Review of Multi-Calculus (Study Guide for Spivak s CHAPTER ONE TO THREE)

Review of Multi-Calculus (Study Guide for Spivak s CHAPTER ONE TO THREE) Review of Multi-Calculus (Study Guide for Spivak s CHPTER ONE TO THREE) This material is for June 9 to 16 (Monday to Monday) Chapter I: Functions on R n Dot product and norm for vectors in R n : Let X

More information

NOTES ON CALCULUS OF VARIATIONS. September 13, 2012

NOTES ON CALCULUS OF VARIATIONS. September 13, 2012 NOTES ON CALCULUS OF VARIATIONS JON JOHNSEN September 13, 212 1. The basic problem In Calculus of Variations one is given a fixed C 2 -function F (t, x, u), where F is defined for t [, t 1 ] and x, u R,

More information

Sobolev Spaces. Chapter Hölder spaces

Sobolev Spaces. Chapter Hölder spaces Chapter 2 Sobolev Spaces Sobolev spaces turn out often to be the proper setting in which to apply ideas of functional analysis to get information concerning partial differential equations. Here, we collect

More information

MINIMAL SURFACES AND MINIMIZERS OF THE GINZBURG-LANDAU ENERGY

MINIMAL SURFACES AND MINIMIZERS OF THE GINZBURG-LANDAU ENERGY MINIMAL SURFACES AND MINIMIZERS OF THE GINZBURG-LANDAU ENERGY O. SAVIN 1. Introduction In this expository article we describe various properties in parallel for minimal surfaces and minimizers of the Ginzburg-Landau

More information

A NOTE ON RANDOM HOLOMORPHIC ITERATION IN CONVEX DOMAINS

A NOTE ON RANDOM HOLOMORPHIC ITERATION IN CONVEX DOMAINS Proceedings of the Edinburgh Mathematical Society (2008) 51, 297 304 c DOI:10.1017/S0013091506001131 Printed in the United Kingdom A NOTE ON RANDOM HOLOMORPHIC ITERATION IN CONVEX DOMAINS FILIPPO BRACCI

More information