DOCTEUR DE L UNIVERSITÉ PARIS XII

Size: px
Start display at page:

Download "DOCTEUR DE L UNIVERSITÉ PARIS XII"

Transcription

1 THÈSE présentée par Mohamed Boussaïri JLELI pour obtenir le titre de DOCTEUR DE L UNIVERSITÉ PARIS XII Spécialité: Mathématiques Titre: Hypersurfaces à courbure moyenne constante. Soutenue le 1 janvier 004 devant le jury composé de: M. Ahmad ELSOUFI Président du jury M. Frédéric HELEIN Rapporteur M. Robert KUSNER Rapporteur M. Frank PACARD Directeur de thèse M. Harold ROSENBERG Examinateur M. Etienne SANDIER Examinateur

2

3 A mes parents, A mes frères et mes sœurs, A ma femme et taati,...

4

5 Remerciements Mes premiers remerciements vont à mon directeur, monsieur Frank PACARD, qui m a fait confiance en me proposant ce sujet. Son enthousiasme, son énergie, sa constante disponibilité et surtout ses qualités humaines ont décuplé ma motivation. Je pense avoir appris énormément à son contact, j espère apprendre encore avec lui. Je remercie messieurs Frédéric HELEIN et Robert KUSNER pour le temps qu ils ont consacré à rapporter cette thèse. Je souhaite de plus leur témoigner toute ma gratitude et mon respect. Mes vifs remerciements vont à messieurs Ahmad ELSOUFI, Harold ROSENBERG et Etienne SANDIER qui me font un grand honneur en acceptant d être membre du jury. Je tiens à remercier les deux occupants du bureau P3-437: Hassen Aydi et Aurelia Fraysse qui ont supporté mes humeurs et ma dissipation. Mes remerciements s adressent aussi à tout les membres du centre de Mathématiques de Paris XII, tant les enseignantschercheurs que les thésards. Enfin et dans un registre plus personnel, je remercie mes amis plus particulièrement Abdellatif, Abou Ayyoub, Fethi, Sahbi, Sami, ma famille et surtout mes parents et ma femme qui loin des Mathématiques m ont toujours apporté le soutien nécessaire.

6 6

7 Contents 1 Notation 15 Constant mean curvature hypersurfaces of revolution 17 1 First parameterization Embedded constant mean curvature hypersurfaces Immersed constant mean curvature hypersurfaces Second parameterization The n-unduloids The n-nodoids Asymptotic properties 1 Compactness results The period of σ τ Expansion of the functions σ τ and κ τ as τ tends to The physical period of a n-delaunay hypersurface The Jacobi operator 35 1 The Jacobi operator about a n-delaunay Geometric Jacobi fields Indicial roots associated to a n-delaunay The Jacobi operator about the sphere S n The Jacobi operator about the n-catenoid Maximum principle 43 1 Maximum principle for the hyperplane R n Maximum principle for the Jacobi operator about S n Maximum principle for the Jacobi operator about the n-catenoid Maximum principle for the n-delaunay hypersurface The indicial roots of a n-delaunay hypersurface 50 1 Positivity of the indicial roots The limit of the indicial roots as τ tends to Bifurcation results for n-nodoids 56 1 Preliminaries The spectrum of L τ when τ > The spectrum of L τ when τ < Equivariant eigenspaces The bifurcation result

8 5.1 Bifurcation from simple eigenvalue General bifurcation result Distinguishing the solutions by their symmetry group Moduli space theory 65 1 Preliminaries Weighted spaces A priori estimates Fredholm properties of the Jacobi operator Nondegenerate hypersurfaces From weighted Lebesgue spaces to weighted Hölder spaces Near a nondegenerate constant mean curvature hypersurface Σ Deformations of the ends of Σ Constant mean curvature hypersurface close to Σ(Ξ) Structure of the mean curvature operator Linear analysis of the n-delaunay Jacobi operator 89 1 The Jacobi operator about a half of n-delaunay hypersurface The Poisson operator associated to Constant mean curvature hypersurfaces close to a n-delaunay hypersurface The nonlinear elliptic equation Perturbation of a half n-delaunay hypersurface The Cauchy data mapping Perturbation of constant mean curvature hypersurfaces The mean curvature operator for vertical graphs Linear analysis Mapping properties of Λ A The Poisson operator associated to R n Nonlinear analysis Constant mean curvature hypersurfaces close to Σ Addition of ends with prescribed Delaunay parameters 11 1 Addition of Delaunay ends The nondegeneracy of the hypersurfaces constructed End-to-End connected sum 17 1 Regular ends of constant mean curvature hypersurfaces Connecting two constant mean curvature hypersurfaces together Mapping properties of the Jacobi operator about M m The perturbation argument An end-to-end construction for compact constant mean curvature hypersurfaces The balancing formula for constant mean curvature hypersurfaces Building blocks Type-1 hypersurfaces

9 . Type- hypersurfaces The construction The mean curvature of the approximate solution Extension of the elements of the deficiency spaces Compact constant mean curvature hypersurfaces with topology

10 10

11 Introduction En 1841, C. Delaunay [3] a classifié toutes les surfaces de révolution de R 3 dont la courbure moyenne est constante égale à 1. Ces surfaces sont maintenant connues sous le nom de surfaces de Delaunay. Les génératrices de ces surfaces sont en fait des roulettes de coniques [4]. Une généralisation de cette classification pour des hypersurfaces de R n, n 3 qui ont certaines symétries a été élaborée en 1981 dans [8]. En dimension supérieure, il existe des hypersurfaces de révolution dans R n+1, pour n 3, dont la courbure moyenne est une constante (fixée égale à 1). Ces hypersurfaces forment une famille à un paramètre D τ pour τ (, 0) (0, τ ] Lorsque τ (0, τ ], l hypersurface D τ est plongée et est appelée n-onduloïde, alors que lorsque τ est négatif, l hypersurface D τ est simplement immergée et est appelée n-nodoïde. Dans les chapitres et 3, une analyse de la paramétrisation de ces hypersurfaces est faite. Une attention toute particulière est portée au comportement de ces hypersurfaces lorsque le paramètre τ tend vers une valeur limite (i.e. τ tend vers 0, ou τ ). L analyse du comportement des n-nodoïdes quand τ tend vers permet de montrer qu il existe, le long de la famille des n-nodoïds, une infinité de points de bifurcations. En particulier ceci assure l existence d une infinité de branches d hypersurfaces à courbure moyenne constante qui sont 1-périodiques, incluses dans un cylindre mais qui ne sont pas des hypersurfaces de révolution. Ce résultat généralise en toutes les dimensions le résultat de R. Mazzeo et F. Pacard [19] qui n était valable que pour les surfaces, n =. La démonstration de ce résultat est basée sur une étude précise du spectre de l opérateur de Jacobi (opérateur de courbure moyenne linéarisé) associé à un n-nodoide, lorsque le paramètre τ tend vers. On montre que, quand le paramètre τ tend vers, l indice de Morse des n-nodoïdes tends vers +. Une application du Théorème classique de M. Crandall et P. Rabinowitz [], [9], concernant les bifurcations à partir d une valeur propre simple nous permet alors de démontrer le Théorème 1: On considère G = O(n 1) O(1). Il existe l 0 0 et, pour tout l l 0, il existe τ l (, 0) et une branche régulière d hypersurfaces à courbure moyenne constante égale à 1 qui bifurque de la famille des n-nodoïdes en D τ en τ = τ l. Chaque élément de cette branche est simplement périodique et invariant par l action du groupe {±I} G mais il n est pas invariant par l action de {±I} O(n). Quand l tend vers +, on a une estimation du paramètre de bifurcation l(n + l) τ l := + O(1), n Un résultat semblable est valable si l on considère G = O(p) O(n p). En utilisant un résultat de bifurcation dû à J. Smoller et A. Wasserman [7], résultat basé sur l utilisation de l indice de Conley, le résultat précedent peut être généralisé 11

12 comme suit : Soit k N et G O(n), on note E k,g le sous-espace de E k (l espace propre du Laplacien sur S n 1 correspondant à la valeur propre λ k = k(n + k )) des fonctions invariantes sous l action des éléments de G. Théorème : Soit G O(n) tel que dim E k,g 0 pour une infinité de k N. Alors, il existe une suite (τ l ) l qui tend vers pour laquelle la famille des n-nodoïdes bifurquent. En particulier dans chaque voisinage de l hypersurface D τl, il existe une hypersurface à courbure moyenne constante égale à 1 qui est simplement périodique et invariante sous l action de {±I} G mais qui n est pas invariante sous l action de {±I} O(n). Dans le chapitre 8, nous étudions l ensemble des hypersurfaces complètes non compactes, à courbure moyenne constante, qui ont un nombre fini de bouts asymptotiques à des n-nodoïdes. Cette étude généralise en toutes les dimensions et pour des hypersurfaces dont les bouts peuvent être de type n-nodoïdes, l étude menée par R. Kusner, R. Mazzeo et D. Pollack [16]. Pour tout k N on note M k l espace des hypersurfaces de R n+1 dont la courbure moyenne est constante égale à 1 qui sont complètes non compactes et qui ont k bouts asymptotiques à des n-delaunays dont le paramètre τ (τ, 0) (0, τ ], où τ < 0 est choisi assez petit. On commence par étudier l opérateur de Jacobi L Σ associé à Σ M k, agissant sur des espaces à poids. Nous donnons dans le Chapitre 8 une démonstration détaillée de l analyse fonctionnelle des opérateurs de Jacobi définis sur de telles hypersurfaces car cette analyse n avait jamais été détaillée auparavent. Ensuite, une application (plus ou moins directe) du Théorème des fonctions implicites permet de démontrer le : Théorème 3 : On suppose que Σ est un élément non dégénéré de M k (i.e. si L Σ w = 0 et si w L (Σ) alors w = 0). Supposons de plus que les paramètres de Delaunay τ l des bouts de Σ appartiennent à (τ, 0) (0, τ ] où τ < 0 est choisi assez petit. Alors, il existe un ouvert U M k contenant Σ qui est une variété régulière de dimension k (n + 1). Insistons sur le fait que l on ne quotiente pas par le groupe des isométries de R n+1. De plus ce résultat n est valable que pour les hypersurfaces dont les bouts sont asymptotiques à des n-onduloïdes ou bien des n-nodoïdes dont le paramètre est proche de 0. Dans les années 80, une méthode de construction de surfaces à courbure moyenne constante à été mise au point par N. Kapouleas d une part pour construire des surfaces compactes [11], [1] d autre part pour construire de surfaces noncompactes complètes dans R 3 [10]. Cette méthode est basée sur une méthode analogue utilisée par R. Schoen pour construire des métriques à courbure scalaire constante sur la sphère privée d un nombre fini de points. Ces techniques ont ensuite été développées et systématisées [18], [0], [5]. En particulier, étant donnée Σ une surface à courbure moyenne constante nondégénérée, il est possible de construire une famille de surfaces à courbure moyenne constante, elles aussi nondégénérées, qui sont somme connexe de Σ et d un nombre fini de surfaces de Delaunay. Dans les chapitres 1 et 13, nous étendons ces résultats au cas des hypersurfaces et nous considérons aussi le cas où l hypersurface de départ Σ est dégénérée. Plus précisément, soit Σ une hypersurface à courbure moyenne constante (compacte ou bien élément de M k ). On note L Σ l opérateur de Jacobi associé à Σ et l on note 1

13 φ 1,..., φ N une base du noyau de L Σ dans L (Σ). Si p =: (p 1,..., p m ) sont m points de Σ, on note Φ(p) := (φ i (p j )) ij. Alors on a le : Théorème 4 : On suppose que les conditions suivantes sont vérifiées : et il existe a := (a 1,..., a m ) (R {0}) m tel que Rang (Φ(p)) = N, (1) Φ(p) t a = 0. () Alors pour tout τ > 0 assez petit, il existe une hypersurface à courbure moyenne constante à k + m bouts qui est obtenue en effectuant la somme connexe de m demi-hypersurfaces de type n-delaunay dont le paramètre de Delaunay est égal à τ i := a i a i (1 n)/n τ à Σ aux points p i Σ. Ces hypersurfaces sont (pour τ assez petit) nondégénérées. Dans le cas particulier où l hypersurface Σ est nondégénérée, il est toujours possible d ajouter un nombre fini de bouts de type n-delaunay et la position des points où le recollement a lieu est arbitraire. Un Corollaire immédiat de ce résultat est le : Théorème 5 : Pour tout k, l ensemble M k n est pas vide et contient des éléments nondégénérés. Une nouvelle technique de construction dite recollement de surfaces bout-à-bout, a été récemment introduite par J. Ratzkin [4]. Dans le Chapitre 14, nous donnons une généralisation de cette construction en dimension quelconque. Dans cette construction, on considère deux hypersurfaces Σ 1 M k1 et Σ M k dont on suppose qu elles ont des bouts E 1 Σ 1 et E Σ qui sont asymptotiques à la même n-delaunay de paramètre τ [τ, 0) (0, τ ). On peut alors aligner Σ 1 et Σ de telle sorte que les axes de E 1 et de E soient confondus (avec des directions opposées). On peut alors translater l une des surfaces le long de cet axe et construire un nombre dénombrable de familles d hypersurfaces dont la courbure moyenne est proche de 1. Ces hypersurfaces peuvent alors être perturbées en des hypersurfaces à courbure moyenne constante éléments de M k1 +k. Le principal avantage de cette construction (outre qu elle permet de construire de nouveau exemples) réside dans la simplicité de sa démonstration. Cette procédure de recollement permet de donner une méthode relativement élémentaire de construction d hypersurfaces compactes à courbure moyenne constante. Dans un premier temps, on utilise le résultat du chapitre 13 pour construire des hypersurfaces à courbure moyenne constante avec un nombre fini de bouts. Ensuite, on montre que la méthode de recollement bout-à-bout peut être mise en œuvre pour construire des hypersurfaces compacte dont la courbure moyenne est constante. L analyse nécessaire est essentiellement basée sur l analyse développée dans le chapitre 8 afin de démontrer le Théorème 4. En particulier nous obtenons le : Théorème 6 : Il existe un ensemble dénombrable d hypersurfaces compactes à courbure moyenne constante (géométriquement distinctes) qui ont la topologie de S n k H 1 (i.e. une sphère S n avec k anses attachées), k. En fait les hypersurfaces que nous construisons ont de nombreuses symétries et l on montre qu il existe un intervalle non vide I et une fonction f régulière définie sur I telle 13

14 que toute solution (x, m) I N de l équation f(x) + m x N, permet de construire une hypersurface compacte à courbure moyenne constante égale 1 dont la topologie est celle d une somme connexe de k copies de S n 1 S 1. 14

15 Chapter 1 Notation In this brief section, we record some notation which will be used in this work. To begin with, let us denote by θ e j (θ), for j N the eigenfunctions of the Laplace-Beltrami operator on S n 1 with corresponding eigenvalue λ j. That is S n 1 e j = λ j e j. We further assume that these eigenvalues are counted with multiplicity, namely λ 0 = 0, λ 1 =... = λ n = n 1, λ n+1 = n,... and λ j λ j+1, and that the eigenfunctions are normalized by for all j N. S n 1 e j(θ) dθ = 1, It will be convenient to define some orthogonal projections π I and π II acting on functions of L (S n 1 ) as follows : If the eigenfunction decomposition of the function φ L (S n 1 ) is given by φ = j N a j e j L (S n 1 ) (1.1) then, we set and We define π I (φ) := π II (φ) := n a j e j, j=0 j n+1 a j e j. This being understood, we define the linear operator ( ( ) ) 1 n δ j := λ j +. (1.) D θ : H 1 (S n 1 ) L (S n 1 ), by D θ (φ) := j N δ j a j e j, 15

16 if the eigendecomposition of the function φ is given by (1.1). In other words, we have ( ( ) ) 1 n D θ = S n 1 +. It is an easy exercise to check that D θ is well defined and bounded. We will agree that, given two functions f and g, the notation g = O C (f) means that, for all k N, there exists a constant c k > 0 such that k s f c k k s g on the domain where the functions f and g are defined. 16

17 Chapter Constant mean curvature hypersurfaces of revolution We give two different parameterizations of a one parameter family of hypersurfaces of revolution in R n+1, which are immersed or embedded and have constant mean curvature normalized to be equal to 1. These hypersurfaces, which were originally studied in [13], generalize the classical constant mean curvature surfaces in R 3 which were discovered by Delaunay in [3] in the middle of the 19-th century. 1 First parameterization We are looking for constant mean curvature hypersurfaces of revolution (say around the x n+1 axis). Such an hypersurface can be locally parameterized by X : (t 1, t ) S n 1 R n+1 (t, θ) (ρ(t) θ, t) where the function t ρ(t) is a smooth positive function which is defined over some interval (t 1, t ). The first fundamental form g of the hypersurface parameterized by X is given by g = (1 + ( t ρ) ) dt dt + ρ dθ i dθ j where dθ i dθ j denotes the first fundamental form of S n 1. Let assume that the orientation of this hypersurface is chosen so that the unit inward normal vector field is given by N := ( t ρ) ( θ, tρ). With this chosen orientation, the second fundamental form b of the hypersurface parameterized by X is given by b = ( t ρ) ( t ρ dt dt + ρ dθ i dθ j ). 17

18 It follows at once from the above expressions that the mean curvature H of the hypersurface parameterized by X (which is the average of the trace of the shape form) is given by H = n n ρ 1 + ( t ρ) 1 1 n (1 + ( t ρ) ) 3 t ρ. (.1) Hence, the condition that the mean curvature of the hypersurface parameterized by X is equal to some given function H, is given by the equation t ρ n 1 ( 1 + ( t ρ) ) + n H ( 1 + ( t ρ) ) ) 3 = 0. (.) ρ We introduce H(ρ, t ρ) := ρ n ( t ρ) H ρn. (.3) In the case where the function H is constant, it follows from a simple computation that H(ρ, t ρ) is constant along solutions of (.). This property will be extensively used to derive a priori estimates for solutions of (.). 1.1 Embedded constant mean curvature hypersurfaces Using the above computation, it is easy to show that there exists a one parameter family of embedded constant mean curvature hypersurfaces of revolution. Indeed, in the case where H = 1, there are two special solutions of (.) which can be immediately determined. The first is the constant solution ρ n 1 n S n 1 ( n 1 n which corresponds to the cylinder ) R R n+1. The other explicit solution is ρ 0 (t) := 1 t, which is defined for t ( 1, 1) and which corresponds to a sphere of radius 1 centered at the origin. These two solutions are the two end-points of a one parameter family of solutions of (.) which produce embedded constant mean curvature hypersurfaces. Let us now describe this family. It will be convenient to define τ := 1 n 1 (n 1) n. n Given τ (0, τ ], we define ε (0, n 1 n ] by the identity ε n 1 (1 ε) = τ n and we define the function ρ τ to be the unique solution of (.), with H = 1, such that ρ τ (0) = ε and t ρ τ (0) = 0. Using the fact that H(ρ τ, t ρ τ ) = τ n, is constant along solutions of (.), it is a simple exercise to show that the function ρ τ is periodic. We will denote by D τ the hypersurface of revolution which is parameterized by X τ (t, θ) = (ρ τ (t) θ, t). By construction, D τ is an embedded constant mean curvature hypersurface of revolution will be called an n-unduloid 1. 1 This has not to be confused with the terminology used in [7] where this terminology refers to surfaces with k ends 18

19 Figure.1: Graphs of ρ τ 1. Immersed constant mean curvature hypersurfaces This time, we construct a one parameter family of immersed constant mean curvature hypersurfaces of revolution. Given τ (, 0), we define ε (0, + ) by the identity ε n 1 (1 + ε) = ( τ) n and we define the function ρ τ to be the unique solution of (.), with H = 1 such that ρ τ (0) = ε and t ρ τ (0) = 0. Using the fact that H( ρ τ, t ρ τ ) = ( τ) n, is constant along solutions of (.), it is a simple exercise to show that, ρ τ is defined over some maximal interval ( T τ, T τ ) where T τ > 0. Furthermore, we have lim ρ τ = τ and lim t ρ τ = ± t ± T τ t ± T τ Now, for the same value of τ, we define ε (1, + ) by the identity ε n 1 (1 ε) = ( τ) n. This being done, we define the function ρ τ to be the unique solution of (.), with H = 1 such that ρ τ (0) = ε and t ρ τ (0) = 0. Using the fact that H( ρ τ, t ρ τ ) = ( τ) n, is constant along solutions of (.), it is a simple exercise to show that, ρ τ is defined in ( T τ, T τ ) where T τ > 0. Furthermore, we have lim ρ τ = τ and lim t ρ τ =. t ± T τ t ± T τ Finally, the graph of the function ρ τ and the graph of the function ρ τ (once translated along the x n+1 axis by T τ T τ ) can be glued together to produce a piece of constant mean curvature hypersurface of revolution. In turn, we can extend this piece of hypersurface by periodicity along the x n+1 axis to produce a complete immersed constant mean curvature surface. These surfaces will be referred to as n-nodoids. Figure.: Graphs of ρ τ and ρ τ Second parameterization The previous parameterization is probably interesting in order to investigate the geometric properties of the hypersurfaces. However, in our analysis, it will be more interesting to consider an isothermal type parameterization for which the geometric properties of the hypersurfaces are hidden but which will be more convenient for analytical purposes. 19

20 Hence, we are now looking for hypersurfaces of revolution which can be parameterized by X(s, θ) = (τ e σ(s) θ, κ(s)), for (s, θ) R S n 1. The constant τ being fixed, the functions σ and κ are determined by asking that the hypersurface parameterized by X has constant mean curvature equal to H and also by asking that the metric associated to the parameterization is conformal to the product metric on R S n 1, namely ( s κ) = τ e σ ( 1 ( s σ) ). (.4) We choose the orientation of the hypersurface parameterized by X so that, the unit normal vector field is given by ( N := ) sκ τe σ θ, sσ. This time, using (.4) the first fundamental form g of the hypersurface parameterized by X is given by g = τ e σ (ds ds + dθ i dθ j ), and its second fundamental form b is given by b = ( s κ s σ s κ ( s σ + ( s σ) ) ) ds ds + s κ dθ i dθ j. Therefore, the mean curvature H of the hypersurface parameterized by X is given by H = 1 nτ e σ ( (n 1) s κ s κ ( s σ + ( s σ) ) + s κ s σ ). This is a rather intricate second order ordinary differential equation in the functions σ and τ which has to be complimented by the equation (.4). In order to simplify our analysis, we use of (.4) to get rid of the factor τ e σ in the above equation. This yields ( s σ s κ = s κ 1 n + s σ + ( s σ) + n H s κ ( 1 ( s σ) ) ) 1. Now, we can differentiate (.4) with respect to s, and we obtain s κ s κ = τ e σ s σ ( 1 s σ ( s σ) ). The difference between the last equation, multiplied by s σ, and the former equation, multiplied by s κ, yields s σ + (1 n)(1 ( s σ) ) + n H s κ = 0. (.5) Hence, in order to find constant mean curvature hypersurfaces of revolution, we have to solve (.4) together with (.5). This is the content of the next two paragraphs..1 The n-unduloids Recall that we have defined τ := 1 n 1 (n 1) n. n 0

21 For all τ (0, τ ]. We define σ τ to be the unique smooth nonconstant solution of ( s σ) + τ ( e σ + e (1 n)σ) = 1, (.6) with initial condition s σ(0) = 0 and σ(0) < 0. Next we define the function κ τ to be the unique solution of s κ = τ ( e σ + e ( n)σ), with κ(0) = 0. (.7) It is easy to check that σ τ and κ τ solve (.4) and (.5), when H = 1. Moreover, the function κ τ is monotone increasing since s κ τ > 0. In particular, the hypersurface parameterized by X τ (s, θ) := (τe στ (s) θ, κ τ (s)), for (s, θ) R S n 1, is an embedded constant mean curvature hypersurface of revolution. This hypersurface coincides with the hypersurface defined in 1.1 and, as already mentioned, this hypersurface will be referred to as the n-unduloid of parameter τ. Observe that the extreme element in this family which corresponds to τ = τ, is the cylinder S n 1 ( n 1 n ) R. While, as τ tends to 0 the family of n-unduloids converges to a sequence of spheres arranged along the x n+1 axis.. The n-nodoids For all τ < 0, we define σ τ to be the unique smooth nonconstant solution of ( s σ) + τ ( e σ e (1 n)σ) = 1, (.8) with initial condition s σ(0) = 0 and σ(0) < 0. Next, we define the function κ τ to be the unique solution of s κ = τ ( e σ e ( n)σ), with κ(0) = 0. (.9) Again, it is easy to check that σ τ and κ τ solve (.4) and (.5), when H = 1. However, observe that this time the function κ τ is not monotone anymore since κ τ changes sign. Hence, the hypersurface parameterized by X τ (s, θ) := ( τe στ (s) θ, κ τ (s)) for (s, θ) R S n 1 is an immersed constant mean curvature hypersurface of revolution. This hypersurface coincides with the hypersurface defined in 1. and, will be referred to as the n-nodoid of parameter τ. 1

22 Chapter 3 Asymptotic properties In this chapter, we are concerned about the behavior of the hypersurfaces which have been defined in the previous chapter, as the parameter τ tends to the limit values, namely, as τ tends to 0 or to. 1 Compactness results We begin with the study of the behavior of σ τ as τ tends to 0. We define ϕ τ := τ e στ, and η τ := τ e (1 n)στ, Using (.4) and (.5), one can check that, according to the sign of τ, the functions ϕ τ and η τ are nonconstant solutions of and ( s η) = (n 1) η (1 (ϕ ± η) ), (3.1) ( s ϕ) = ϕ (1 (ϕ ± η) ), (3.) with a + when τ > 0 and a when τ < 0. In addition, we have ϕ n 1 τ η τ = τ n. (3.3) Our first Lemma states that the functions ϕ τ and η τ and their derivatives, are uniformly bounded with respect to τ, provided that τ remains bounded. Lemma 3.1. Assume that τ 0 < 0 is fixed. Then, for all k N, there exists a constant c k > 0 which only depends on τ 0 and k, such that for all τ [τ 0, 0) (0, τ ]. ϕ τ C k + η τ C k c k Proof: When τ > 0, observe that (3.) already implies that the functions ϕ τ and η τ are uniformly bounded by 1. When τ < 0 is bounded from below by τ 0, (3.) together with (3.3) imply that the functions ϕ τ and η τ are uniformly bounded by a constant only depending on τ 0. Now that we know that the functions ϕ τ and η τ are uniformly bounded. We use (3.1) and (3.) inductively to show that the same property is also true for the derivatives of the functions ϕ τ and η τ.

23 Assume that s l is a sequence of real numbers, and that τ l is a sequence which tends to 0. For all l N, we define ϕ l := ϕ τl ( s l ) and η l := η τl ( s l ). The previous result together with Ascoli s theorem allows one to extract from the sequence (ϕ l, η l ) l, a subsequence which converges, as l tends to, to (ϕ, η ) in C k topology on any compact sets. The following Lemma classifies the possible limits (ϕ, η ). Lemma 3.. Under the above hypothesis, the following holds : Either ϕ = η 0, or ϕ 0 and there exists s such that η = or η 0 and there exists s such that 1 cosh((n 1)( s )), ϕ = 1 cosh( s ). Proof: Passing the limit in (3.3) we get ϕ n 1 η 0. This implies that, at least one of the functions η and ϕ has to be identically equal to 0. It only remains to identify the possible nontrivial limits. If ϕ 0, we can pass to the limit in (3.1) and in the derivative of (3.1) with respect to s to get the equation satisfied by η Furthermore, we have s η = (n 1) η ( 1 η ). ( s η) = (n 1) η ( 1 η ). The nontrivial solutions of these equations are all of the form for some s 0 R. s 1 cosh((n 1)( s 0 )) Now, if η 0, we can pass to the limit in (3.) and in the derivative of (3.) to get equation satisfied by η s ϕ = ϕ ( 1 ϕ ). Furthermore, we have ( s ϕ) = ϕ (1 ϕ ). This time the only nontrivial solutions of this equation are of the form s 1 cosh( s 0 ), for some s 0 R. This completes the proof of the result. 3

24 The period of σ τ When τ τ, the function σ τ is periodic and not constant. Hence, given τ (, 0) (0, τ ), we can define s τ > 0 to be the least period of the function σ τ. When τ = τ, the function σ τ is constant and we extend the function τ s τ by continuity as τ tends to τ. The value of s τ can be computed explicitly. This is the content of the : Lemma 3.3. As τ tends to τ, we have lim s τ = τ τ π n 1. Proof: Assume that τ (0, τ ) and that σ τ is a solution of (.6). For the sake of simplicity in the notations, we drop the τ indices. We can write where we have defined s σ = 1 τ f(σ), (3.4) f(σ) = e σ + e (1 n)σ. (3.5) It is easy to see that f is convex and achieves its minimum at σ := 1 n log(n 1). Using Taylor s expansion of f about σ, we can rephrase (3.4) as ( s σ) = τ τ τ ( ) τ (n 1) f(σ) (σ σ ) τ where f is a smooth function which satisfies f(σ ) = 1. Now, we define the function ω by the identity τ σ = σ + τ ω, By construction, the function ω solves ( ) τ ( s ω) = 1 (n 1) ω ˆf(ω) (3.6) where the smooth function ˆf is defined by ( ˆf(ω) := f σ + τ τ τ τ In particular, ˆf 1 and this implies that ( ) τ (n 1) ω 1 (3.7) By definition, the period of σ (or ω) is given by s τ = ω+ (τ) ω (τ) τ 1 (n 1) τ dx ( τ τ ω ) ) x ˆf(x). 4

25 where ω (τ) < 0 < ω + (τ) are the roots of the equation ( ) τ (n 1) ω ˆf(ω) = 1 τ In view of (3.7), the function ω remains bounded as τ tends to τ. Hence, we obtain lim ω ± (τ) = ± 1. τ τ n 1 Finally, Lebesgue s dominated convergence theorem implies that lim s τ = τ τ 1 n 1 1 n 1 dx 1 (n 1)x The result then follows at once. We define the variable e (0, 1) by e := τ τ τ for all τ (0, τ ). It is easy to see that the solution σ τ depends smoothly on e at e = 0. Using the proof of Lemma 3.3, we show that, as τ tends to τ, the function σ τ can be expanded as e σ τ = σ cos( n 1 ) + O(e ). n 1 where This is the content of the : σ := 1 n log(n 1). (3.8) Lemma 3.4. The solution of (3.6) satisfies e ω e=0 = 1 n 1 cos( n 1 s). Proof: With our notations, we have σ = σ + e ω Hence, is remains to understand ω, the limit of ω as τ tends to τ. By (3.7), the function ω is bounded uniformly as τ tends to τ and, passing to the limit in (3.6), we conclude that ( s ω ) + (n 1) ω = 1. The result then follows at once from the fact that σ τ achieves its minimum at s = 0 and hence this is also the case for the function ω. In the next result, we study the asymptotic behavior of the function τ s τ, as the parameter τ tends to 0. 5

26 Proposition 3.1. As τ tends to 0, we have s τ = n n 1 log τ + O(1). (3.9) Proof: In the case where τ < 0, the function σ τ is the solution of (.8). We have the expression of the period s τ as s τ = σ where we have defined the function g by and σ < 0 < σ + are the solutions of σ + dσ 1 τ g(σ) g(σ) = e σ e (1 n)σ 1 τ g(σ ± ) = 1 τ. The function g being clearly a diffeomorphism from R into R, the expression of s τ can be written as 1 τ 1 dx s τ = g g 1 (x) 1 τ x We perform the change of variables y = τx, this yields s τ = τ g g 1 (y/τ) dy 1 y. We define t τ := τ +1 1 Expanding g at ±, we get Finally, the function ( ( ) ( )) dy 1 1 g g 1 (y/τ) = g 1 g 1 τ τ t τ = n n 1 log τ + O( τ n z z g g 1 (z) n 1 ). 1 converges to 1 as z tends to + and to n 1 as z tends to. This together with Lebesgue s dominated convergence theorem, implies that the difference s τ t τ satisfies lim (s τ t τ ) = τ 0 n n y dy 1 y y. In the case where τ < 0, the result follows from the estimate of t τ and the boundedness of the last limit. Let us now assume that τ > 0. This time, the function σ τ is the solution of (.6) and we can write σ+ dσ σ s τ = 1 τ f(σ) + dσ 1 τ f(σ) σ 6 σ

27 where f has been defined in (3.5), σ is defined in (3.8) and where σ < 0 < σ + are defined to be the solutions of f(σ ± ) = 1 τ, Using arguments which are similar to the one developed in the case where τ < 0, we obtain σ+ t + dσ τ := = f 1 (1/τ) + O(1) 1 τ f(σ) t τ := σ σ σ dσ = f 1 (1/τ) + O(1) 1 τ f(σ) where f+ 1 1 (resp. f ) is the inverse of the restriction of f to (σ, + ) (resp. to (, σ )). The proof of the result when τ > 0 follows from a similar argument as in the proof of the result when τ < 0, we leave the details to the reader. + 3 Expansion of the functions σ τ and κ τ as τ tends to 0 In the remaining of this section, we obtain some precise expansions of the functions σ τ and κ τ as the parameter τ tends to 0. It will be convenient to observe that the function satisfies the equation (n 1) στ ξ τ := e ( s ξ) = (n 1) ( ξ τ (1 ± ξ n n 1 ) ). (3.10) with a + when τ > 0 and a when τ < 0. Furthermore, this function achieves its minimum at s = 0, namely ξ(s) ξ(0) := ξ 0 for all s R, since s = 0 corresponds to the point where the function σ τ is minimal. Therefore, it is possible to define a function ω by the identity ξ(s) := ξ 0 cosh(ω(s)). This being understood, our first result gives an expansion of the function ξ τ over the interval [ s τ /, s τ /] as τ tends to 0. The functions σ τ and ξ τ being s τ periodic, this gives us an expansion of the function ξ τ in all R. Lemma 3.5. For all s [ sτ, sτ ] we have ξ τ τ cosh((n 1)s) = 1 + τ n n 1 OC (1) + τ n n 1 OC ((cosh s) ), uniformly in τ as τ tends to 0. Proof: Using (3.10), we see that the function ω satisfies ( )) ( s ω) = (n 1) 1 τ (ξ n 1 0 A(cosh ω) ± τ n n 1 ξ0 B(cosh ω) (3.11) where we have defined for all z > 1 A(z) := z n n 1 1 z 1 7

28 and B(z) := z n n 1 1 z 1. Now, using (.6) or (.8), together with the fact that s σ(0) = 0, we see that ) ξ 0 = τ (1 + O( τ n n 1 ), as τ tends to 0. Plugging this information into (3.11), and assuming that we obtain ω(s) (n 1) s 1 (3.1) s ω(s) = n 1 + τ n n 1 OC ((cosh s) n ) + τ n n 1 OC ((cosh s) ), in the range where (3.1) holds. Hence, as long as (3.1) is fulfilled, we can integrate the previous expansion and conclude that ω(s) = (n 1) s + τ n n 1 OC (1) + τ n n 1 OC ((cosh s) ). (3.13) Finally, remember that the result of Proposition 3.1 implies that s τ = n n 1 log τ + O(1). This expansion, together with (3.13) prove that (3.1) is fulfilled in some interval of the form [ sτ + s, sτ s], for some fixed s > 0, independent of τ (0, τ ]. Finally, we use (3.11) to extend this estimate to [ sτ, sτ ]. The result then follows from the expansions of ω and ξ 0. The previous expansion being understood, we make use of (.7) or (.9) to obtain the corresponding estimates for the function κ τ. For further use, it will be more convenient to use the variable r := τ e σ(s). instead of the variable s. In the next Lemma, we agree that the notation f = O C 0 (r l ) means that, for all k N, there exists a constant c k > 0 such that r k k r f c k r l on the domain where the function f is defined. This definition being understood, we can now state the : Lemma 3.6. As τ tends to zero the function κ admits this expansion κ τ (s) = c τ ± τ n n r n ( 1 + τ n O C 0 (r n ) + τ n O C 0 (r n ) ), (3.14) for all r ( τ n n 1, 1), with a + when τ > 0 and a when τ < 0. Here the constant c τ R satisfies c τ c τ n n 1, for some constant c > 0 which does not depend on τ. 8

29 Proof: We recall that s κ = ±τ e ( n)σ (1 ± e nσ ). with a + when τ > 0 and a when τ < 0. Then, using the result of the last Lemma we get ) s κ = ± τ n n n 1 (cosh(n 1)s) n 1 (1 + τ n n 1 OC ((cosh s) n )), and for s [0, sτ ] this equation becomes ( ) s κ = ± n n n 1 τ n 1 e ( n)s 1 + τ n n 1 OC (e ns ) + O C (e ( n)s ). We integrate this quantity over [0, s] to conclude that κ satisfy κ = c τ ± n n 1 τ n n 1 e( n)s n ) (1 + τ n n 1 OC (e ns ) + O C (e ( n)s ), (3.15) for some constant c τ R which is bounded by a constant, independent of τ times τ n n 1. Now, using the expansion of e σ which is given in the last Lemma together with the definition of r = τ e σ(s) we deduce that ( ) r = 1 n n 1 τ n 1 e s 1 + O C (e ( n)s ) + τ n n 1 OC (1) + τ n n 1 OC (e s ). In particular, this implies that c 0 τ n n 1 e s r C 0 τ n n 1 e s, for some positive constants c 0 < 1 < C 0 which do not depend on τ. Inserting this in (3.15) to obtain the result. The estimate of κ τ when s [ sτ, 0] can be established similarly. We end this chapter with the study of the behavior of s τ and σ τ when the parameter τ tends to. We assume that τ < 0 and we introduce an auxiliary function ( sτ ) γ τ (t) := τσ τ π t. With this normalization, the function γ τ is π periodic. We prove in the next result that γ τ converges to a multiple of the cosine function, as τ tends to. We also give the asymptotic of the period s τ as τ tends to. Lemma 3.7. The following expansions hold as τ tends to s τ = π ( 1 + O( τ 1 ) ) and τ s τ = π nτ nτ (1 + O( τ 1 )). (3.16) Furthermore, as τ, the function γ τ converges uniformly to the function cos t. More precisely γ τ (t) = 1 n cos t + O C ( τ 1 ) and τ γ τ = O C ( τ ). (3.17) 9

30 Proof: Step 1 We define The proof of the Lemma is decomposed into three steps. σ M := σ τ (s τ /), the maximal value of the function σ τ and s 0 (0, sτ ) denotes the value of s for which σ τ vanishes. The function σ τ being maximal when s = s τ /, we can define the function v by the identity σ τ (s) = σ M cos(τv(s)), for all s [s 0, sτ ] and, to ensure uniqueness of v, we require that v(s τ /) = 0 and s v(s τ /) > 0. The parameter τ being negative, the function σ τ satisfies (.8), hence the function v satisfies ( s v) = g(σ τ ) g(σ M ) στ σm (3.18) where we recall that the function g is defined by g(σ) := e σ e (1 n)σ. The function σ τ being maximal at s τ /, we have from (.8) e σ M e (1 n)σ M = 1 τ and this implies that σ M can be expanded as σ M = 1 nτ + O(τ ). Hence σ τ = O C ( τ 1 ). Using the first term of the Taylor expansion of the function g we conclude that g(σ τ ) g(σ M ) σ τ σ M = n + O( τ 1 ) and g(σ τ ) + g(σ M ) σ τ + σ M = n + O( τ 1 ). These expansions together with the integration of (3.18) yield ( v = n s s ) τ (1 + O( τ 1 ) ) (3.19) for s [s 0, sτ ]. Step Now, we define σ m = σ τ (0) and as above we write for s [0, s 0 ], with σ τ = σ m cos(τw(s)), τ w(0) = 0 and s w (0) > 0. Following the strategy developed in the first part of this proof, we show that and also that σ m = 1 nτ + O(τ ). w = n s ( 1 + O C ( τ 1 ) ). 30

31 for all s [0, s 0 ]. Step 3 We use the fact that σ M cos (τ v(s 0 )) = σ m cos (τ w(s 0 )) = 0. The expansions derived in the first two steps imply that ( τ n s O( τ 1 ) ) ( = τ n s 0 s ) τ (1 + O( τ 1 ) ) = π. Hence, we conclude that s τ = π ( 1 + O( τ 1 ) ). nτ Inserting this information into the expansions obtained in the first two steps, we obtain the expansion γ τ = 1 n cos t + O C ( τ 1 ) for all t [0, s τ /]. The result then follows from the fact that σ τ is even and periodic of period s τ. Finally, to obtain the estimate for the derivative of s τ, we remember that s τ = τ g g 1 (y/τ) dy 1 y. Then τ s τ = 1 τ s τ τ = 1 τ s τ τ ( 1 ) τ g g 1 (y/τ) dy 1 y g g 1 (y/τ) ydy (g g 1 (y/τ)) 3. 1 y Since, the function z g g 1 (z) (g g 1 (z)) 3 is bounded, we use the expansion of s τ established below to deduce that τ s τ = π nτ + O C ( τ 3 ). 4 The physical period of a n-delaunay hypersurface Recall that a n-delaunay hypersurface can be parameterized as X τ : (s, θ) (τ e σ(s) θ, κ(s)) where the function σ τ is s τ periodic. The physical period of this hypersurface is defined by T τ = κ τ (s τ ), 31

32 so that X τ (s + s τ, θ) = X τ (s, θ) + T τ (0, 1). We would like to study the behavior of T τ as τ varies. When τ τ, it follows from the analysis given in the previous Chapter that the function σ τ is strictly increasing for s (0, s τ /). Hence it can be used as a change of parameter and we can define It follows from (.6) and (.8) that K := κ τ σ 1 τ. σ K = τ e σ H(σ) 1 H (σ) where H(σ) := τ (e σ (1 n) + ι e σ) and ι is equal to + when τ > 0 and equal to when τ < 0. Then, we have the expression of T τ which is given by σ+ T τ = τ e σ H(σ) dσ σ 1 H (σ) (3.0) where σ < 0 < σ + are the roots of H(σ) = 1. Now, we study the asymptotic behavior of T τ as the parameter τ tends to 0. Proposition 3.. As τ tends to 0, we have T τ = + ι c n τ n n 1 + O(τ ), with c n := n x 1 n 1 1 x dx. Proof: We will distinguish two cases according to the sign of τ Case 1: Assume that τ < 0 and let us define the function h(y) : y y 1 n, which is increasing and can be used as a change of variable. Writing h(e σ ) = x the physical period will be write as τ in (3.0), with T τ := 1 1 G(x) := x G( x ) dx, (3.1) 1 x τ 1 h h 1 (x). 3

33 First, Lebesgue s dominated convergence Theorem implies that lim T τ =. τ 0 Then, there exists c > 0 such that for all x positive the function G satisfies the two estimates G(x) 1 c (1 + x) n, and x G (x) c (1 + x) n. Using these, satisfies and T 1 := 1 0 x G( x 1 x τ ) dx, T 1 c τ, τ T 1 c τ. Now, it is easy to verify that there exists a constant c > 0 such that for all x < 0, we have G(x) 1 n n n 1 x n 1 c (1 x) n 1, and x G (x) c (1 x) n n 1. Hence satisfies T + n 1 τ n n 1 T := x 1 x G( x τ ) dx, x 1 n 1 1 x dx c τ 1 τ 0 x 1 n 1 (1 + x) n n 1 Finally, we see that the last integral converges. Which finishes the proof in this case. Case : Assume that τ (0, τ ) and let us define the function Then, changing e σ by y, (3.0) becomes f(y) := y + y 1 n. dx. with T τ := τ y+ y f(y) 1 τ f (y) dy, f(y ) = f(y + ) = 1 τ. Now, it is clear that f 1 the restriction of f in (y, (n 1) 1 n ) is strictly decreasing and f the other restriction of f in ((n 1) 1 n, y + ) is strictly increasing. Hence, the physical period can be written as 33

34 where ( τ τ T τ = 1 As in the first case, we prove that x G 1( x 1 1 x τ ) dx + τ τ x 1 x G ( x τ ) dx ), G 1 (x) := f 1 f 1 1 (x) and G (x) := f f 1 (x). τ τ 1 x G 1( x 1 x τ ) dx = τ n n 1 n x 1 n 1 1 x dx + O(τ ), and 1 This completes the proof. τ τ x 1 x G ( x τ ) dx = 1 + O(τ ). Remark 3.1. In the case where τ < 0, T τ is monotone and τ T τ > 0. Indeed, thanks to (3.1) we have τ T τ = 1 1 x 1 x n(n 1) z n 1 dx ( ) (n 1) z n τ > 0 where z := h 1 ( x τ ). 34

35 Chapter 4 The Jacobi operator In this chapter, we define and study the Jacobi operator about a n-delaunay hypersurface, about the n-sphere and also about the n-catenoid, a minimal hypersurface of revolution which generalizes the -dimensional catenoid. Next, we give the expression of the geometric Jacobi fields. Finally, we define the indicial roots associated with the Jacobi operator about a n-delaunay hypersurface. 1 The Jacobi operator about a n-delaunay Recall that the n-delaunay hypersurface D τ can be parameterized as X τ = (ι τ e στ θ, κ τ ). (4.1) Assume that the orientation of D τ is chosen so that the unit normal vector field is given by ( N τ := ι ) sκ τ e θ, sσ τ. (4.) στ Any hypersurface, close enough to D τ, can be parameterized (at last locally) as a normal graph over D τ. Namely, by X ω = X τ + ω N τ, for some (small) smooth function ω. The hypersurface parameterized by X ω will be denoted by D τ (ω) and we define the mean curvature operator H(ω) to be the mean curvature of D τ (ω). It is well known [5] that the linearized mean curvature operator about D τ, which is usually referred to as the Jacobi operator, is given by L τ := τ + A τ where τ is the Laplace-Beltrami operator and A τ is the square of the norm of the shape operator A τ on D τ. Recall that we have defined in Chapter 3 the function ϕ τ := τ e στ and, in the above parameterization, the metric on D τ is given by g = ϕ τ (ds ds + dθ i dθ j ), 35

36 and the second fundamental form is given by b = ϕ τ ( (1 ± (1 n) τ n ϕ n τ ) ds ds + (1 ± τ n ϕ n ) τ ) dθ i dθ j, with a + when τ > 0 and a when τ < 0. Using this, we find the expression of the Jacobi operator in term of the function ϕ τ L τ := ϕ n τ s (ϕ n τ It will be convenient to define the conjugate operator s ) + ϕ τ S n 1 + n + n (n 1) τ n ϕ n τ. (4.3) L τ := ϕ n+ τ which is explicitly given in terms of the function ϕ τ by ( ) n L τ = s + S n 1 + L τ ϕ n τ, (4.4) n(n + ) 4 ϕ τ + n(3n ) 4 τ n ϕ n τ. (4.5) Since the operators L τ and L τ are conjugate, the mapping properties of one of them will easily translate for the other one. With slight abuse of terminology, we shall refer to any of them as the Jacobi operator about D τ. We finally define the operator ( ) n 0 := s + S n 1, (4.6) which appears in the expression of L τ. This is conjugate to the Jacobi operator about the hyperplane R n {0} R n+1 in polar coordinates. Indeed if r = e s and θ S n 1 then Geometric Jacobi fields 0 = e n+ R n e n. Some Jacobi field, i.e., solution of the homogeneous problem L τ ω = 0, can be explicitly computed when they correspond to a smooth one parameter family of constant mean curvature hypersurfaces C λ, for λ ( 1, 1) to which D τ belongs, for example C 0 = D τ. For λ small enough, the hypersurface C λ can be written (at least locally) as a normal graph over D τ, for some function w λ. Differentiation with respect to λ, at λ = 0, yields a Jacobi field. Multiplying the result by ϕ n τ, we obtain a solution of L τ ω = 0 which, with slight abuse of terminology, will be referred to as a Jacobi field. In the special case where the one parameter family of constant mean curvature hypersurfaces is given by the action of a one parameter family of rigid motions, the corresponding Jacobi field can be obtained by projecting onto the normal bundle of D τ, the Killing vector field associated to the one parameter family of rigid motions under 36

37 consideration. We now describe these Jacobi fields as well as another independent Jacobi field which arise by changing the Delaunay parameter τ. Translation along the axis To begin with, for τ (, 0) (0, τ ), we define Φ 0,+ τ to be the Jacobi field corresponding to the translation of D τ along the x n+1 axis. As explained above, this Jacobi field can be obtained by projecting onto the normal bundle, the constant Killing field Since we conclude that χ(x, x n+1 ) := (0, 1) R n R. χ N τ = s σ τ = ϕ 1 τ s ϕ τ Φ 0,+ τ = ϕ n 4 s ϕ (4.7) is a solution of L τ ω = 0. It is easy to check that Φτ 0,+ only depends on s and is periodic, in particular, this implies that this Jacobi field is bounded. Translations orthogonal to the axis Since we have n directions orthogonal to x n+1, there are n linearly independent Jacobi fields which are obtained by translating D τ in a direction orthogonal to its axis. Let χ(x, x n+1 ) := (a, 0) R n R denote the constant Killing field which correspond to the translations in the (a, 0) direction. Then, we have ( χ N τ = τ e σ ± e (1 n)σ) a θ Taking a to be any vector of the canonical basis of R n, we get for j = 1,..., n ( ) Φ j,+ τ = ϕ n ± τ n ϕ n e j (4.8) which are solution of L τ ω = 0. Again, we see that Φ j, τ all j = 1,..., n. is periodic (hence bounded) for Rotations of the axis Next, we compute the Jacobi fields which correspond to the rotation of D τ in a direction orthogonal to the axis. Given a R n, We define the Killing field χ(x, x n+1 ) := ( x n+1 a, a x) R n R and we obtain Φ j, τ ( s κ κ χ N τ = ϕ ) + sϕ a θ Taking a to be any vector of the canonical basis of R n, we obtain for j = 1,..., n ( ) (s, θ) = ϕ n 4 ϕ s ϕ + κ s κ e j (4.9) which are solutions of L τ = 0. Observe that Φ j,+ τ is not periodic (hence not bounded), but grows linearly. In particular, there a constant c > 0 (depending on τ) such that for all (s, θ) R S n 1. Φ j,+ τ c (1 + s ), 37

38 Change of Delaunay parameter Finally, The Jacobi field corresponding to a change of parameter τ (, 0) (0, τ ) will be denoted by Φτ 0,. It can be obtained by writing, for η small enough, the constant mean curvature hypersurface D τ+η as a normal graph over D τ for some function ω η and differentiating ω η with respect to η at η = 0. More precisely, for τ close to τ there exists a local diffeomorphism Φ τ such that In particular, we get differentiating ω τ X τ Φ τ = X τ + ω τ N τ. ω τ = (X τ Φ τ X τ ) N τ, with respect to τ at τ = τ we get τ ω τ τ =τ = τ X τ N τ + (DX τ ( τ Φ τ )) N τ. Since (DX τ ( τ Φ τ )) N τ 0, the corresponding Jacobi field takes the form Φ 0, τ = ϕ n 4 ( τ κ s ϕ τ ϕ s κ). (4.10) Observe that, in the first and in the last study, we have assumed that τ τ. Some changes are needed to handle the case where τ = τ. Indeed, D τ is the cylinder S n 1 ( n 1 n ) R and hence it is invariant under translation along its axis and the projection of the constant Killing vector field (0, 1) along the normal gives 0! This fact is a consequence of the parameterization of the set of Delaunays hypersurfaces we have chosen. Therefore, we can give an intrinsic definition of the Jacobi fields Φτ 0,± by considering a different parameterization of the set of Delaunay hypersurfaces in the neighborhood of the cylinder S n 1 ( n 1 n ) R. So far we have considered the hypersurfaces parameterized by ( ) X(s, θ) = τ e στ (s+t) θ, κ τ (s + t) where t R and τ (0, τ ]. Now, we define the parameters η := π τ t and e := τ κ(s τ ) τ so that (e, η) [0, 1) S 1 are in some sense polar coordinates in the space of parameters It follows from Lemma 3.4, that σ τ = σ Then, a simple exercise shows that e n 1 cos( n 1 s) + O C (e ). Φ 0,+ τ = cos( n 1 s) and Φ 0, τ = sin( n 1 s) and these two Jacobi fields are periodic (hence bounded) corresponding to differentiation with respect to e and t respectively. 38

39 3 Indicial roots associated to a n-delaunay Because of the rotational invariance of the operator L τ we can introduce the eigenfunction decomposition with respect to the cross-sectional Laplace-Beltrami operator S n 1. In this way we obtain the sequence of operators ( ) n L τ,j = s λ j + n(n + ) ϕ + 4 n(3n ) τ n ϕ n (4.11) 4 for j N. By definition, the indicial roots of the operator L τ,j characterize the rate of growth (or rate of decay) of the solutions of the homogeneous equation L τ,j ω = 0 at infinity (see [16]). To explain this, observe that the potential in the expression of L τ,j is given by δj n(n + ) + ϕ n(3n ) + τ n ϕ n 4 4 where δ j has been defined in (1.), and hence this potential is periodic of period s τ. We define a by matrix T j (τ) M (R) such that, for all ω solution of L τ,j ω = 0 in R, we have ( ) ( ) ω ω (s s ω τ ) = T j (τ) (0). s ω Let λ (τ, j) and λ + (τ, j) denote the roots of the characteristic polynomial of T j (τ). Assume that v 1 and v are the solutions of L τ,j v i = 0 with v 1 (0) = s v (0) = 1 and v (0) = s v 1 (0) = 0. We denote by W the Wronskian of v 1, v W := v 1 s v v s v 1 The Wronskian of v 1 and v being constant, we evaluate it at s = 0 and s = s τ, and we obtain det (T j (τ)) = λ (τ, j)λ + (τ, j) = 1. Observe that the matrix T j (τ) has real entries, hence This being understood, we can write Tr (T j (τ)) = λ (τ, j) + λ + (τ, j) R λ + (τ, j) = µ e iξ and λ (τ, j) = 1 µ e iξ (4.1) where ξ := ξ(τ, j) [0, π) and µ := µ(τ, j) 1 satisfy (µ 1) sin ξ = 0. Now, we distinguish two cases according to the value of ξ. Case 1 Assume that ξ (0, π) (π, π). Then λ + (τ) = e iξ and λ (τ) = e iξ are distinct. In particular, the matrix T j (τ) can be diagonalized there exists two independent solutions of L τ,j ω = 0, whose Cauchy data at t = 0 correspond, to the eigenvectors of T j (τ). These two solutions are bounded. 39

Outils de Recherche Opérationnelle en Génie MTH Astuce de modélisation en Programmation Linéaire

Outils de Recherche Opérationnelle en Génie MTH Astuce de modélisation en Programmation Linéaire Outils de Recherche Opérationnelle en Génie MTH 8414 Astuce de modélisation en Programmation Linéaire Résumé Les problèmes ne se présentent pas toujours sous une forme qui soit naturellement linéaire.

More information

The conformal theory of Alexandrov embedded constant mean curvature surfaces in R 3

The conformal theory of Alexandrov embedded constant mean curvature surfaces in R 3 The conformal theory of Alexandrov embedded constant mean curvature surfaces in R 3 Rafe Mazzeo Stanford University Frank Pacard Université de Paris XII Daniel Pollack University of Washington Abstract

More information

Apprentissage automatique Méthodes à noyaux - motivation

Apprentissage automatique Méthodes à noyaux - motivation Apprentissage automatique Méthodes à noyaux - motivation MODÉLISATION NON-LINÉAIRE prédicteur non-linéaire On a vu plusieurs algorithmes qui produisent des modèles linéaires (régression ou classification)

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

Constant mean curvature surfaces with Delaunay ends

Constant mean curvature surfaces with Delaunay ends Constant mean curvature surfaces with Delaunay ends Rafe Mazzeo Stanford University Frank Pacard Université Paris XII October 12, 2007 1 Introduction and statement of the results In this paper we shall

More information

Kato s inequality when u is a measure. L inégalité de Kato lorsque u est une mesure

Kato s inequality when u is a measure. L inégalité de Kato lorsque u est une mesure Kato s inequality when u is a measure L inégalité de Kato lorsque u est une mesure Haïm Brezis a,b, Augusto C. Ponce a,b, a Laboratoire Jacques-Louis Lions, Université Pierre et Marie Curie, BC 187, 4

More information

Exercise sheet n Compute the eigenvalues and the eigenvectors of the following matrices. C =

Exercise sheet n Compute the eigenvalues and the eigenvectors of the following matrices. C = L2 - UE MAT334 Exercise sheet n 7 Eigenvalues and eigenvectors 1. Compute the eigenvalues and the eigenvectors of the following matrices. 1 1 1 2 3 4 4 1 4 B = 1 1 1 1 1 1 1 1 1 C = Which of the previous

More information

Free Boundary Minimal Surfaces in the Unit 3-Ball

Free Boundary Minimal Surfaces in the Unit 3-Ball Free Boundary Minimal Surfaces in the Unit 3-Ball T. Zolotareva (joint work with A. Folha and F. Pacard) CMLS, Ecole polytechnique December 15 2015 Free boundary minimal surfaces in B 3 Denition : minimal

More information

arxiv:math/ v1 [math.dg] 28 Mar 2006

arxiv:math/ v1 [math.dg] 28 Mar 2006 Riemann minimal surfaces in higher dimensions arxiv:math/060366v1 [math.dg] 8 Mar 006 S. Kaabachi and F. Pacard February, 008 Abstract. We prove the existence of a one parameter family of minimal embedded

More information

Apprentissage automatique Machine à vecteurs de support - motivation

Apprentissage automatique Machine à vecteurs de support - motivation Apprentissage automatique Machine à vecteurs de support - motivation RÉGRESSION À NOYAU régression à noyau Algorithme de régression à noyau entraînement : prédiction : a = (K + λi N ) 1 t. y(x) =k(x) T

More information

A note on the moving hyperplane method

A note on the moving hyperplane method 001-Luminy conference on Quasilinear Elliptic and Parabolic Equations and Systems, Electronic Journal of Differential Equations, Conference 08, 00, pp 1 6. http://ejde.math.swt.edu or http://ejde.math.unt.edu

More information

arxiv: v2 [math.dg] 13 Feb 2018

arxiv: v2 [math.dg] 13 Feb 2018 HOLOMORPHIC CARTAN GEOMETRIES ON COMPLEX TORI INDRANIL BISWAS AND SORIN DUMITRESCU arxiv:1710.05874v2 [math.dg] 13 Feb 2018 Abstract. In [DM] it was asked whether all flat holomorphic Cartan geometries(g,

More information

Bifurcating nodoids. Rafe Mazzeo and Frank Pacard

Bifurcating nodoids. Rafe Mazzeo and Frank Pacard Bifurcating nodoids Rafe Mazzeo and Frank Pacard Abstract. All complete, axially symmetric surfaces of constant mean curvature in R 3 lie in the one-parameter family D τ of Delaunay surfaces. The elements

More information

ANNALES SCIENTIFIQUES L ÉCOLE NORMALE SUPÉRIEURE. Cluster ensembles, quantization and the dilogarithm. Vladimir V. FOCK & Alexander B.

ANNALES SCIENTIFIQUES L ÉCOLE NORMALE SUPÉRIEURE. Cluster ensembles, quantization and the dilogarithm. Vladimir V. FOCK & Alexander B. ISSN 0012-9593 ASENAH quatrième série - tome 42 fascicule 6 novembre-décembre 2009 ANNALES SCIENTIFIQUES de L ÉCOLE NORMALE SUPÉRIEURE Vladimir V. FOCK & Alexander B. GONCHAROV Cluster ensembles, quantization

More information

Stable bundles on non-algebraic surfaces giving rise to compact moduli spaces

Stable bundles on non-algebraic surfaces giving rise to compact moduli spaces Analytic Geometry/ Géométrie analytique Stable bundles on non-algebraic surfaces giving rise to compact moduli spaces Matei Toma Abstract We prove the existence of a class of holomorphic vector bundles

More information

Random variables. Florence Perronnin. Univ. Grenoble Alpes, LIG, Inria. September 28, 2018

Random variables. Florence Perronnin. Univ. Grenoble Alpes, LIG, Inria. September 28, 2018 Random variables Florence Perronnin Univ. Grenoble Alpes, LIG, Inria September 28, 2018 Florence Perronnin (UGA) Random variables September 28, 2018 1 / 42 Variables aléatoires Outline 1 Variables aléatoires

More information

ON THE ASYMPTOTICS OF GREEN S FUNCTIONS OF ELLIPTIC OPERATORS WITH CONSTANT COEFFICIENTS. Shmuel Agmon

ON THE ASYMPTOTICS OF GREEN S FUNCTIONS OF ELLIPTIC OPERATORS WITH CONSTANT COEFFICIENTS. Shmuel Agmon Séminaires & Congrès 9, 2004, p. 13 23 ON THE ASYMPTOTICS OF GREEN S FUNCTIONS OF ELLIPTIC OPERATORS WITH CONSTANT COEFFICIENTS by Shmuel Agmon Abstract. In this paper we discuss the following problem.

More information

The Moduli Space Of Complete Embedded Constant Mean Curvature Surfaces

The Moduli Space Of Complete Embedded Constant Mean Curvature Surfaces University of Massachusetts Amherst From the SelectedWorks of Robert Kusner 1996 The Moduli Space Of Complete Embedded Constant Mean Curvature Surfaces Robert Kusner, University of Massachusetts - Amherst

More information

THE RESOLUTION OF SAFFARI S PHASE PROBLEM. Soit K n := { p n : p n (z) = n

THE RESOLUTION OF SAFFARI S PHASE PROBLEM. Soit K n := { p n : p n (z) = n Comptes Rendus Acad. Sci. Paris Analyse Fonctionnelle Titre français: Solution du problème de la phase de Saffari. THE RESOLUTION OF SAFFARI S PHASE PROBLEM Tamás Erdélyi Abstract. We prove a conjecture

More information

On the uniform Poincaré inequality

On the uniform Poincaré inequality On the uniform Poincaré inequality Abdesslam oulkhemair, Abdelkrim Chakib To cite this version: Abdesslam oulkhemair, Abdelkrim Chakib. On the uniform Poincaré inequality. Communications in Partial Differential

More information

Fatou s Theorem and minimal graphs

Fatou s Theorem and minimal graphs Fatou s Theorem and minimal graphs José M. Espinar 1, Harold Rosenberg May 15, 2009 Institut de Mathématiques, Université Paris VII, 175 Rue du Chevaleret, 75013 Paris, France; e-mail: jespinar@ugr.es

More information

arxiv:cs/ v1 [cs.dm] 21 Apr 2005

arxiv:cs/ v1 [cs.dm] 21 Apr 2005 arxiv:cs/0504090v1 [cs.dm] 21 Apr 2005 Abstract Discrete Morse Theory for free chain complexes Théorie de Morse pour des complexes de chaines libres Dmitry N. Kozlov Eidgenössische Technische Hochschule,

More information

It s a Small World After All Calculus without s and s

It s a Small World After All Calculus without s and s It s a Small World After All Calculus without s and s Dan Sloughter Department of Mathematics Furman University November 18, 2004 Smallworld p1/39 L Hôpital s axiom Guillaume François Antoine Marquis de

More information

Numerical solution of the Monge-Ampère equation by a Newton s algorithm

Numerical solution of the Monge-Ampère equation by a Newton s algorithm Numerical solution of the Monge-Ampère equation by a Newton s algorithm Grégoire Loeper a, Francesca Rapetti b a Département de Mathématiques, Ecole Polytechnique Fédérale de Lausanne, 1015 Lausanne, CH

More information

Holomorphic torsion on Hermitian symmetric spaces

Holomorphic torsion on Hermitian symmetric spaces Géométrie différentielle/differential geometry Holomorphic torsion on Hermitian symmetric spaces Kai Köhler Abstract We calculate explicitly the equivariant holomorphic Ray-Singer torsion for all equivariant

More information

Γ -convergence and Sobolev norms

Γ -convergence and Sobolev norms C. R. Acad. Sci. Paris, Ser. I 345 (2007) 679 684 http://france.elsevier.com/direct/crass1/ Partial Differential Equations Γ -convergence and Sobolev norms Hoai-Minh Nguyen Rutgers University, Department

More information

SIMPLE QUASICRYSTALS ARE SETS OF STABLE SAMPLING

SIMPLE QUASICRYSTALS ARE SETS OF STABLE SAMPLING SIMPLE QUASICRYSTALS ARE SETS OF STABLE SAMPLING BASARAB MATEI AND YVES MEYER Version française abrégée Soit K R n un ensemble compact et soit E K L 2 (R n ) le sous-espace de L 2 (R n ) composé de toutes

More information

CALCULUS ON MANIFOLDS. 1. Riemannian manifolds Recall that for any smooth manifold M, dim M = n, the union T M =

CALCULUS ON MANIFOLDS. 1. Riemannian manifolds Recall that for any smooth manifold M, dim M = n, the union T M = CALCULUS ON MANIFOLDS 1. Riemannian manifolds Recall that for any smooth manifold M, dim M = n, the union T M = a M T am, called the tangent bundle, is itself a smooth manifold, dim T M = 2n. Example 1.

More information

A set of formulas for primes

A set of formulas for primes A set of formulas for primes by Simon Plouffe December 31, 2018 Abstract In 1947, W. H. Mills published a paper describing a formula that gives primes : if A 1.3063778838630806904686144926 then A is always

More information

Constant mean curvature spheres in Riemannian manifolds

Constant mean curvature spheres in Riemannian manifolds Constant mean curvature spheres in Riemannian manifolds F. Pacard and X. Xu Abstract We prove the existence of embedded spheres with large constant mean curvature in any compact Riemannian manifold M,

More information

On the coverage probability of the Clopper-Pearson confidence interval

On the coverage probability of the Clopper-Pearson confidence interval On the coverage probability of the Clopper-Pearson confidence interval Sur la probabilité de couverture de l intervalle de confiance de Clopper-Pearson Rapport Interne GET / ENST Bretagne Dominique Pastor

More information

ANNALES. FLORENT BALACHEFF, ERAN MAKOVER, HUGO PARLIER Systole growth for finite area hyperbolic surfaces

ANNALES. FLORENT BALACHEFF, ERAN MAKOVER, HUGO PARLIER Systole growth for finite area hyperbolic surfaces ANNALES DE LA FACULTÉ DES SCIENCES Mathématiques FLORENT BALACHEFF, ERAN MAKOVER, HUGO PARLIER Systole growth for finite area hyperbolic surfaces Tome XXIII, n o 1 (2014), p. 175-180.

More information

Répartition géographique des investissements directs étrangers en Chine

Répartition géographique des investissements directs étrangers en Chine Répartition géographique des investissements directs étrangers en Chine Qixu Chen To cite this version: Qixu Chen. Répartition géographique des investissements directs étrangers en Chine. Economies et

More information

Connected sum constructions in geometry and nonlinear analysis. Frank Pacard

Connected sum constructions in geometry and nonlinear analysis. Frank Pacard Connected sum constructions in geometry and nonlinear analysis Frank Pacard January 19, 2008 2 Contents 1 Laplace-Beltrami operator 5 1.1 Definition................................ 5 1.2 Spectrum of the

More information

Expression of Dirichlet boundary conditions in terms of the strain tensor in linearized elasticity

Expression of Dirichlet boundary conditions in terms of the strain tensor in linearized elasticity Expression of Dirichlet boundary conditions in terms of the strain tensor in linearized elasticity Philippe Ciarlet a, Cristinel Mardare b a Department of Mathematics, City University of Hong Kong, 83

More information

arxiv: v1 [math.dg] 28 Jun 2008

arxiv: v1 [math.dg] 28 Jun 2008 Limit Surfaces of Riemann Examples David Hoffman, Wayne Rossman arxiv:0806.467v [math.dg] 28 Jun 2008 Introduction The only connected minimal surfaces foliated by circles and lines are domains on one of

More information

for all subintervals I J. If the same is true for the dyadic subintervals I D J only, we will write ϕ BMO d (J). In fact, the following is true

for all subintervals I J. If the same is true for the dyadic subintervals I D J only, we will write ϕ BMO d (J). In fact, the following is true 3 ohn Nirenberg inequality, Part I A function ϕ L () belongs to the space BMO() if sup ϕ(s) ϕ I I I < for all subintervals I If the same is true for the dyadic subintervals I D only, we will write ϕ BMO

More information

NOTE ON THE NODAL LINE OF THE P-LAPLACIAN. 1. Introduction In this paper we consider the nonlinear elliptic boundary-value problem

NOTE ON THE NODAL LINE OF THE P-LAPLACIAN. 1. Introduction In this paper we consider the nonlinear elliptic boundary-value problem 2005-Oujda International Conference on Nonlinear Analysis. Electronic Journal of Differential Equations, Conference 14, 2006, pp. 155 162. ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu

More information

Linear Ordinary Differential Equations

Linear Ordinary Differential Equations MTH.B402; Sect. 1 20180703) 2 Linear Ordinary Differential Equations Preliminaries: Matrix Norms. Denote by M n R) the set of n n matrix with real components, which can be identified the vector space R

More information

Bounding Stability Constants for Affinely Parameter-Dependent Operators

Bounding Stability Constants for Affinely Parameter-Dependent Operators Bounding Stability Constants for Affinely Parameter-Dependent Operators Robert O Connor a a RWTH Aachen University, Aachen, Germany Received *****; accepted after revision +++++ Presented by Abstract In

More information

A DIFFERENT APPROACH TO MULTIPLE CORRESPONDENCE ANALYSIS (MCA) THAN THAT OF SPECIFIC MCA. Odysseas E. MOSCHIDIS 1

A DIFFERENT APPROACH TO MULTIPLE CORRESPONDENCE ANALYSIS (MCA) THAN THAT OF SPECIFIC MCA. Odysseas E. MOSCHIDIS 1 Math. Sci. hum / Mathematics and Social Sciences 47 e année, n 86, 009), p. 77-88) A DIFFERENT APPROACH TO MULTIPLE CORRESPONDENCE ANALYSIS MCA) THAN THAT OF SPECIFIC MCA Odysseas E. MOSCHIDIS RÉSUMÉ Un

More information

Φ B. , into the page. 2π ln(b/a).

Φ B. , into the page. 2π ln(b/a). Chapitre 29 Induction électromagnétique [13 au 15 juin] DEVOIR : 29.8; 29.20; 29.22; 29.30; 29.36 29.1. Expériences d'induction Il n est pas nécessaire de lire cette section. Ce qu il faut retenir de la

More information

Brunn Minkowski Theory in Minkowski space-times. François Fillastre Université de Cergy Pontoise France

Brunn Minkowski Theory in Minkowski space-times. François Fillastre Université de Cergy Pontoise France Brunn Minkowski Theory in Minkowski space-times François Fillastre Université de Cergy Pontoise France Convex bodies A convex body is a (non empty) compact convex set of R d+1 Convex bodies A convex body

More information

THE OLYMPIAD CORNER No. 305

THE OLYMPIAD CORNER No. 305 THE OLYMPIAD CORNER / 67 THE OLYMPIAD CORNER No. 305 Nicolae Strungaru The solutions to the problems are due to the editor by 1 January 014. Each problem is given in English and French, the official languages

More information

An explicit formula for ndinv, a new statistic for two-shuffle parking functions

An explicit formula for ndinv, a new statistic for two-shuffle parking functions FPSAC 2012, Nagoya, Japan DMTCS proc AR, 2012, 147 156 An explicit formula for ndinv, a new statistic for two-shuffle parking functions Angela Hicks and Yeonkyung Kim Mathematics Department, University

More information

( ) 2 ( kg) ( 9.80 m/s 2

( ) 2 ( kg) ( 9.80 m/s 2 Chapitre 1 Charges et champs électriques [9 au 1 mai] DEVOIR : 1.78, 1.84, 1.56, 1.90, 1.71 1.1. Charge électrique et structure de la matière À lire rapidement. Concepts déjà familiers. 1.. Conducteurs,

More information

Régularité des équations de Hamilton-Jacobi du premier ordre et applications aux jeux à champ moyen

Régularité des équations de Hamilton-Jacobi du premier ordre et applications aux jeux à champ moyen Régularité des équations de Hamilton-Jacobi du premier ordre et applications aux jeux à champ moyen Daniela Tonon en collaboration avec P. Cardaliaguet et A. Porretta CEREMADE, Université Paris-Dauphine,

More information

Docteur de l Université Henri Poincaré, Nancy-I. Roger NAKAD. Sous-variétés spéciales des variétés spinorielles complexes

Docteur de l Université Henri Poincaré, Nancy-I. Roger NAKAD. Sous-variétés spéciales des variétés spinorielles complexes UFR S.T.M.I.A. École Doctorale IAEM Lorraine Université Henri Poincaré - Nancy I D.F.D. Mathématiques Thèse présentée pour l obtention du titre de Docteur de l Université Henri Poincaré, Nancy-I en Mathématiques

More information

Small and big sets in analysis

Small and big sets in analysis Small and big sets in analysis Frédéric Bayart Université Blaise Pascal, Clermont-Ferrand May 2015 What is a small set? Aaronszajn null, capacity zero, cube null, Dirichlet, first category, Γ-null, Gauss

More information

ASYMPTOTIC BEHAVIOR OF STRUCTURES MADE OF CURVED RODS

ASYMPTOTIC BEHAVIOR OF STRUCTURES MADE OF CURVED RODS ASYMPTOTIC BEHAVIOR OF STRUCTURES MADE OF CURVED RODS G. GRISO Abstract. In this paper we study the asymptotic behavior of a structure made of curved rods of thickness when 0. This study is carried on

More information

Basic building blocks for a triple-double intermediate format

Basic building blocks for a triple-double intermediate format Laboratoire de l Informatique du Parallélisme École Normale Supérieure de Lyon Unité Mixte de Recherche CNRS-INRIA-ENS LYON-UCBL n o 5668 Basic building blocks for a triple-double intermediate format Christoph

More information

arxiv: v2 [gr-qc] 9 Sep 2010

arxiv: v2 [gr-qc] 9 Sep 2010 arxiv:1006.2933v2 [gr-qc] 9 Sep 2010 Global geometry of T 2 symmetric spacetimes with weak regularity Philippe G. LeFloch a and Jacques Smulevici b a Laboratoire Jacques-Louis Lions & Centre National de

More information

i=1 α i. Given an m-times continuously

i=1 α i. Given an m-times continuously 1 Fundamentals 1.1 Classification and characteristics Let Ω R d, d N, d 2, be an open set and α = (α 1,, α d ) T N d 0, N 0 := N {0}, a multiindex with α := d i=1 α i. Given an m-times continuously differentiable

More information

MGDA II: A direct method for calculating a descent direction common to several criteria

MGDA II: A direct method for calculating a descent direction common to several criteria MGDA II: A direct method for calculating a descent direction common to several criteria Jean-Antoine Désidéri To cite this version: Jean-Antoine Désidéri. MGDA II: A direct method for calculating a descent

More information

MATH 205C: STATIONARY PHASE LEMMA

MATH 205C: STATIONARY PHASE LEMMA MATH 205C: STATIONARY PHASE LEMMA For ω, consider an integral of the form I(ω) = e iωf(x) u(x) dx, where u Cc (R n ) complex valued, with support in a compact set K, and f C (R n ) real valued. Thus, I(ω)

More information

A set of formulas for primes

A set of formulas for primes A set of formulas for primes by Simon Plouffe December 31, 2018 Abstract In 1947, W. H. Mills published a paper describing a formula that gives primes : if A 1.3063778838630806904686144926 then A is always

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

TRAVELING WAVE SOLUTIONS OF ADVECTION-DIFFUSION EQUATIONS WITH NONLINEAR DIFFUSION

TRAVELING WAVE SOLUTIONS OF ADVECTION-DIFFUSION EQUATIONS WITH NONLINEAR DIFFUSION Annales de l IHP C 00 202 3 logo IHP C TRAVELING WAVE SOLUTIONS OF ADVECTION-DIFFUSION EQUATIONS WITH NONLINEAR DIFFUSION L. Monsaingeon a, A. Novikov b, J.-M. Roquejoffre a a Institut de Mathématiques

More information

Counterexample to the infinite dimensional Carleson embedding theorem

Counterexample to the infinite dimensional Carleson embedding theorem Counterexample to the infinite dimensional Carleson embedding theorem Fedor Nazarov, Serguei Treil, and Alexander Volberg Abstract. We are going to show that the classical Carleson embedding theorem fails

More information

Alexey ZYKIN PROPRIÉTÉS ASYMPTOTIQUES DES CORPS GLOBAUX

Alexey ZYKIN PROPRIÉTÉS ASYMPTOTIQUES DES CORPS GLOBAUX UNIVERSITÉ DE LA MÉDITERRANÉE AIX MARSEILLE II Faculté des Sciences de Luminy ÉCOLE DOCTORALE DE MATHÉMATIQUES ET INFORMATIQUE E.D. 184 THÈSE présentée pour obtenir le grade de Docteur de l Université

More information

Math 350 Fall 2011 Notes about inner product spaces. In this notes we state and prove some important properties of inner product spaces.

Math 350 Fall 2011 Notes about inner product spaces. In this notes we state and prove some important properties of inner product spaces. Math 350 Fall 2011 Notes about inner product spaces In this notes we state and prove some important properties of inner product spaces. First, recall the dot product on R n : if x, y R n, say x = (x 1,...,

More information

ERRATUM TO AFFINE MANIFOLDS, SYZ GEOMETRY AND THE Y VERTEX

ERRATUM TO AFFINE MANIFOLDS, SYZ GEOMETRY AND THE Y VERTEX ERRATUM TO AFFINE MANIFOLDS, SYZ GEOMETRY AND THE Y VERTEX JOHN LOFTIN, SHING-TUNG YAU, AND ERIC ZASLOW 1. Main result The purpose of this erratum is to correct an error in the proof of the main result

More information

Titre : Feuilletages courbure moyenne constante dans les espaces asymptotiquement

Titre : Feuilletages courbure moyenne constante dans les espaces asymptotiquement Titre : Feuilletages courbure moyenne constante dans les espaces asymptotiquement hyperboliques. Résumé : Soit M, g) une variété asymptotiquement hyperbolique dont la compactification conforme est régulière.

More information

STOKES PROBLEM WITH SEVERAL TYPES OF BOUNDARY CONDITIONS IN AN EXTERIOR DOMAIN

STOKES PROBLEM WITH SEVERAL TYPES OF BOUNDARY CONDITIONS IN AN EXTERIOR DOMAIN Electronic Journal of Differential Equations, Vol. 2013 2013, No. 196, pp. 1 28. ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu ftp ejde.math.txstate.edu STOKES PROBLEM

More information

UNIVERSITE PARIS XI UFR Scientifique d Orsay THESE. Matthieu ALFARO. Systèmes de convection-réaction-diffusion et dynamique d interface

UNIVERSITE PARIS XI UFR Scientifique d Orsay THESE. Matthieu ALFARO. Systèmes de convection-réaction-diffusion et dynamique d interface N D ORDRE : 8392 UNIVERSITE PARIS XI UFR Scientifique d Orsay THESE présentée pour obtenir le grade de Docteur en sciences de l Université Paris XI Orsay, spécialité mathématiques par Matthieu ALFARO Sujet

More information

William P. Thurston. The Geometry and Topology of Three-Manifolds

William P. Thurston. The Geometry and Topology of Three-Manifolds William P. Thurston The Geometry and Topology of Three-Manifolds Electronic version 1.1 - March 00 http://www.msri.org/publications/books/gt3m/ This is an electronic edition of the 1980 notes distributed

More information

FRACTAL LAPLACIANS ON THE UNIT INTERVAL

FRACTAL LAPLACIANS ON THE UNIT INTERVAL Ann. Sci. Math. Québec 7 3, no.,. FRACTAL LAPLACIANS ON THE UNIT INTERVAL ERIK J. BIRD, SZE-MAN NGAI AND ALEXANDER TEPLYAEV RÉSUMÉ. Nous étudions les valeurs/fonctions propres du laplacien sur [, ] définies

More information

On the direct kinematics of planar parallel manipulators: special architectures and number of solutions

On the direct kinematics of planar parallel manipulators: special architectures and number of solutions On the direct kinematics of planar parallel manipulators: special architectures and number of solutions by Clément M. Gosselin and Jean-Pierre Merlet Département de Génie Mécanique Université Laval Ste-Foy,

More information

Generalized Ricci Bounds and Convergence of Metric Measure Spaces

Generalized Ricci Bounds and Convergence of Metric Measure Spaces Generalized Ricci Bounds and Convergence of Metric Measure Spaces Bornes Généralisées de la Courbure Ricci et Convergence des Espaces Métriques Mesurés Karl-Theodor Sturm a a Institut für Angewandte Mathematik,

More information

Questions d analyse et de géométrie sur les espaces métriques mesurés et les groupes

Questions d analyse et de géométrie sur les espaces métriques mesurés et les groupes UNIVERSITÉ DE CERGY-PONTOISE ET UNIVERSITÉ DE NEUCHATEL THÈSE DE DOCTORAT Spécialité Mathématiques Ecole doctorale Economie et Mathématiques de Paris-Ouest. Présentée par Romain Tessera Questions d analyse

More information

Scalar curvature and the Thurston norm

Scalar curvature and the Thurston norm Scalar curvature and the Thurston norm P. B. Kronheimer 1 andt.s.mrowka 2 Harvard University, CAMBRIDGE MA 02138 Massachusetts Institute of Technology, CAMBRIDGE MA 02139 1. Introduction Let Y be a closed,

More information

DIFFERENTIAL GEOMETRY, LECTURE 16-17, JULY 14-17

DIFFERENTIAL GEOMETRY, LECTURE 16-17, JULY 14-17 DIFFERENTIAL GEOMETRY, LECTURE 16-17, JULY 14-17 6. Geodesics A parametrized line γ : [a, b] R n in R n is straight (and the parametrization is uniform) if the vector γ (t) does not depend on t. Thus,

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

THREE-MANIFOLDS OF CONSTANT VECTOR CURVATURE ONE

THREE-MANIFOLDS OF CONSTANT VECTOR CURVATURE ONE THREE-MANIFOLDS OF CONSTANT VECTOR CURVATURE ONE BENJAMIN SCHMIDT AND JON WOLFSON ABSTRACT. A Riemannian manifold has CVC(ɛ) if its sectional curvatures satisfy sec ε or sec ε pointwise, and if every tangent

More information

Black Holes and Thermodynamics I: Classical Black Holes

Black Holes and Thermodynamics I: Classical Black Holes Black Holes and Thermodynamics I: Classical Black Holes Robert M. Wald General references: R.M. Wald General Relativity University of Chicago Press (Chicago, 1984); R.M. Wald Living Rev. Rel. 4, 6 (2001).

More information

Nonlinear semigroups generated by j-elliptic functionals

Nonlinear semigroups generated by j-elliptic functionals Nonlinear semigroups generated by j-elliptic functionals Ralph Chill TU Dresden, Institut für Analysis, Fachrichtung Mathematik, 01062 Dresden, Germany. ralph.chill@tu-dresden.de Daniel Hauer School of

More information

Exercises in Geometry II University of Bonn, Summer semester 2015 Professor: Prof. Christian Blohmann Assistant: Saskia Voss Sheet 1

Exercises in Geometry II University of Bonn, Summer semester 2015 Professor: Prof. Christian Blohmann Assistant: Saskia Voss Sheet 1 Assistant: Saskia Voss Sheet 1 1. Conformal change of Riemannian metrics [3 points] Let (M, g) be a Riemannian manifold. A conformal change is a nonnegative function λ : M (0, ). Such a function defines

More information

ISOPERIMETRIC REGIONS IN ROTATIONALLY SYMMETRIC CONVEX BODIES

ISOPERIMETRIC REGIONS IN ROTATIONALLY SYMMETRIC CONVEX BODIES ISOPERIMETRIC REGIONS IN ROTATIONALLY SYMMETRIC CONVEX BODIES CÉSAR ROSALES Abstract. We consider the isoperimetric problem of minimizing perimeter for given volume in a strictly convex domain Ω R n+1

More information

DETERMINING HIGH VOLTAGE CABLE CONDUCTOR TEMPERATURES. Guy Van der Veken. Euromold, Belgium. INVESTIGATIONS. INTRODUCTION.

DETERMINING HIGH VOLTAGE CABLE CONDUCTOR TEMPERATURES. Guy Van der Veken. Euromold, Belgium. INVESTIGATIONS. INTRODUCTION. DETERMINING HIGH VOLTAGE CABLE CONDUCTOR TEMPERATURES. Guy Van der Veken. Euromold, Belgium. INTRODUCTION. INVESTIGATIONS. Type tests on MV cable accessories are described in CENELEC HD68 and HD69 documents.

More information

CHAPTER 3. Gauss map. In this chapter we will study the Gauss map of surfaces in R 3.

CHAPTER 3. Gauss map. In this chapter we will study the Gauss map of surfaces in R 3. CHAPTER 3 Gauss map In this chapter we will study the Gauss map of surfaces in R 3. 3.1. Surfaces in R 3 Let S R 3 be a submanifold of dimension 2. Let {U i, ϕ i } be a DS on S. For any p U i we have a

More information

LINEAR FLOW IN POROUS MEDIA WITH DOUBLE PERIODICITY

LINEAR FLOW IN POROUS MEDIA WITH DOUBLE PERIODICITY PORTUGALIAE MATHEMATICA Vol. 56 Fasc. 2 1999 LINEAR FLOW IN POROUS MEDIA WITH DOUBLE PERIODICITY R. Bunoiu and J. Saint Jean Paulin Abstract: We study the classical steady Stokes equations with homogeneous

More information

Periodic constant mean curvature surfaces in H 2 R

Periodic constant mean curvature surfaces in H 2 R Periodic constant mean curvature surfaces in H 2 R Laurent Mazet, M. Magdalena Rodríguez and Harold Rosenberg June 8, 2011 1 Introduction A properly embedded surface Σ in H 2 R, invariant by a non-trivial

More information

Riemannian Curvature Functionals: Lecture III

Riemannian Curvature Functionals: Lecture III Riemannian Curvature Functionals: Lecture III Jeff Viaclovsky Park City Mathematics Institute July 18, 2013 Lecture Outline Today we will discuss the following: Complete the local description of the moduli

More information

Mathématiques. Denis AUROUX

Mathématiques. Denis AUROUX Thèse présentée pour obtenir le titre de DOCTEUR DE L ÉCOLE POLYTECHNIQUE Spécialité : Mathématiques par Denis AUROUX Titre : Théorèmes de structure des variétés symplectiques compactes via des techniques

More information

Georgia Tech PHYS 6124 Mathematical Methods of Physics I

Georgia Tech PHYS 6124 Mathematical Methods of Physics I Georgia Tech PHYS 612 Mathematical Methods of Physics I Instructor: Predrag Cvitanović Fall semester 2012 Homework Set #5 due October 2, 2012 == show all your work for maximum credit, == put labels, title,

More information

Functional Analysis. Franck Sueur Metric spaces Definitions Completeness Compactness Separability...

Functional Analysis. Franck Sueur Metric spaces Definitions Completeness Compactness Separability... Functional Analysis Franck Sueur 2018-2019 Contents 1 Metric spaces 1 1.1 Definitions........................................ 1 1.2 Completeness...................................... 3 1.3 Compactness......................................

More information

Thèse de Doctorat de L'Université Paris-Saclay. L'Université Paris-Sud. Inria Saclay Ile-de-France

Thèse de Doctorat de L'Université Paris-Saclay. L'Université Paris-Sud. Inria Saclay Ile-de-France NNT : 2016SACLS459 Thèse de Doctorat de L'Université Paris-Saclay préparée à L'Université Paris-Sud au sein de Inria Saclay Ile-de-France ÉCOLE DOCTORALE N 580 Sciences et technologies de l'information

More information

Delaunay Surfaces in S 3

Delaunay Surfaces in S 3 Delaunay Surfaces in S 3 Ryan Hynd ryanhynd@math.berkeley.edu Department of Mathematics University of California, Berkeley Queen Dido Conference Carthage, Tunisia May 28, 2010 Part 1: Delaunay Surfaces

More information

arxiv: v4 [math.dg] 7 Nov 2007

arxiv: v4 [math.dg] 7 Nov 2007 The Ricci iteration and its applications arxiv:0706.2777v4 [math.dg] 7 Nov 2007 Yanir A. Rubinstein Abstract. In this Note we introduce and study dynamical systems related to the Ricci operator on the

More information

Content. Content. Introduction. T. Chateau. Computer Vision. Introduction. Outil projectif permettant l acquisition d une scène 3D sur un plan 2D

Content. Content. Introduction. T. Chateau. Computer Vision. Introduction. Outil projectif permettant l acquisition d une scène 3D sur un plan 2D Content Modèle de caméra T Chateau Lamea/Gravir/ComSee, Blaie Pacal Univerit Computer Viion 2 Content La projection perpective Changement de repère objet/caméra Changement de repère caméra/image Changement

More information

DOUBLY PERIODIC SELF-TRANSLATING SURFACES FOR THE MEAN CURVATURE FLOW

DOUBLY PERIODIC SELF-TRANSLATING SURFACES FOR THE MEAN CURVATURE FLOW DOUBLY PERIODIC SELF-TRANSLATING SURFACES FOR THE MEAN CURVATURE FLOW XUAN HIEN NGUYEN Abstract. We construct new examples of self-translating surfaces for the mean curvature flow from a periodic configuration

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

SPECTRAL PROPERTIES OF THE LAPLACIAN ON BOUNDED DOMAINS

SPECTRAL PROPERTIES OF THE LAPLACIAN ON BOUNDED DOMAINS SPECTRAL PROPERTIES OF THE LAPLACIAN ON BOUNDED DOMAINS TSOGTGEREL GANTUMUR Abstract. After establishing discrete spectra for a large class of elliptic operators, we present some fundamental spectral properties

More information

The multi-terminal vertex separator problem : Complexity, Polyhedra and Algorithms

The multi-terminal vertex separator problem : Complexity, Polyhedra and Algorithms The multi-terminal vertex separator problem : Complexity, Polyhedra and Algorithms Youcef Magnouche To cite this version: Youcef Magnouche. The multi-terminal vertex separator problem : Complexity, Polyhedra

More information

Rigidity and Non-rigidity Results on the Sphere

Rigidity and Non-rigidity Results on the Sphere Rigidity and Non-rigidity Results on the Sphere Fengbo Hang Xiaodong Wang Department of Mathematics Michigan State University Oct., 00 1 Introduction It is a simple consequence of the maximum principle

More information

4.2. ORTHOGONALITY 161

4.2. ORTHOGONALITY 161 4.2. ORTHOGONALITY 161 Definition 4.2.9 An affine space (E, E ) is a Euclidean affine space iff its underlying vector space E is a Euclidean vector space. Given any two points a, b E, we define the distance

More information

INSTANTON MODULI AND COMPACTIFICATION MATTHEW MAHOWALD

INSTANTON MODULI AND COMPACTIFICATION MATTHEW MAHOWALD INSTANTON MODULI AND COMPACTIFICATION MATTHEW MAHOWALD () Instanton (definition) (2) ADHM construction (3) Compactification. Instantons.. Notation. Throughout this talk, we will use the following notation:

More information

Isometric elastic deformations

Isometric elastic deformations Isometric elastic deformations Fares Al-Azemi and Ovidiu Calin Abstract. This paper deals with the problem of finding a class of isometric deformations of simple and closed curves, which decrease the total

More information

On the minimal number of periodic Reeb orbits on a contact manifold

On the minimal number of periodic Reeb orbits on a contact manifold On the minimal number of periodic Reeb orbits on a contact manifold Jean Gutt To cite this version: Jean Gutt. On the minimal number of periodic Reeb orbits on a contact manifold. General Mathematics [math.gm].

More information

arxiv:math/ v1 [math.ap] 28 Oct 2005

arxiv:math/ v1 [math.ap] 28 Oct 2005 arxiv:math/050643v [math.ap] 28 Oct 2005 A remark on asymptotic completeness for the critical nonlinear Klein-Gordon equation Hans Lindblad and Avy Soffer University of California at San Diego and Rutgers

More information