DOCSLIB.ORG

Some Basic Topological and Algebraic Properties of Convex Sets in What

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Some Basic Topological and Algebraic Properties of Convex Sets in What Some basic topological and algebraic properties of convex sets In what follows, C is a convex set in a vector space X. Topological interior. Let us start with the following easy observation. Observation 0.1. Let C be a convex set in a normed space X. If C contains an 0 open ball B (z0; r), then 0¡ ¢ B (1 ¡ t)z0 + ty; (1 ¡ t)r ½ C whenever y 2 C; t 2 [0; 1). 0 Proof. Fix y 2 C, t 2 [0; 1). Since C is convex, it contains the set (1 ¡ t)B (z0; r) + ty = B0((1 ¡ t)x + ty; (1 ¡ t)r). ¤ Corollary 0.2. Let C be a convex set in a normed space X. If x 2 int(C) and y 2 C, then [x; y) ½ int(C). In particular, int(C) is a convex set. Theorem 0.3. Let C be a convex set in a normed space X. (a) If either int(C) 6= ; or C is ¯nite-dimensional, then int(C) = int(C). (b) If int(C) 6= ;, then C = int(C). Proof. First part of (a). Suppose that int(C) 6= ;. Thus C contains an open ball 0 B (z0; r). One inclusion is obvious. Let x 2 int(C). There exists % > 0 such 0 0 that B (x; %) ½ C. If x = z0, we are done. If x 6= z0, ¯x y 2 B (x; %) such that x 2 (z0; y), that is, there exists t 2 (0; 1) with x = (1 ¡ t)z0 + ty. Since y 2 C, there 0 0 0 0 exists y 2 C be such that ky ¡y k < r. For x = (1¡t)z0 +ty , Observation 0.1 says that B0(x0; (1¡t)r) ½ C. Then x 2 int(C), since kx¡x0k = (1¡t)ky¡y0k < (1¡t)r. Second part of (a). Let C be ¯nite-dimensional. If a®(C) = X, then int(C) 6= ; by the \relative interior theorem"; hence we can apply the ¯rst part of (a). If a®(C) 6= X, the a±ne hull of C is closed (since it is a translate of a ¯nite-dimensional linear set) and has no interior points. Since int(C) ½ int(C) ½ int(a®(C)) = int(a®(C)) = ;, we have int(C) = int(C) = ;. (b) Obviously, C ⊃ int(C). If x 2 C, ¯x z0 2 int(C). Arbitrarily near to x we can ¯nd some y 2 C; but since [z0; y) ½ int(C) (Corollary 0.2), arbitrarily near to x we can ¯nd points of int(C). Thus x 2 int(C). ¤ Algebraic interior. Let C be a convex set in a vector space X, x 2 C. We say that x is an algebraically interior point of C if for every v 2 X there exists t > 0 such that x + tv 2 C. In this case, we shall write x 2 a-int(C). (For a nonconvex set E ½ X, an algebraically interior point is a point x 2 E such that for each v 2 X there exists t > 0 such that [x; x + tv] ½ E.) In other words, x 2 a-int(C) if and only if x 2 intL(C \ L) for each line L ½ X passing through x. The following proposition is an analogue of Corollary 0.2. Proposition 0.4. Let C be a convex set in a vector space X. If x 2 a-int(C) and y 2 C, then [x; y) 2 a-int(C). In particular, the set a-int(C) is convex. 1 2 Proof. Let 0 < ¸ < 1, z = (1 ¡ ¸)x + ¸y, v 2 X. Take t > 0 such that x + tv 2 C. Then z +(1¡¸)tv 2 C since z +(1¡¸)tv = (1¡¸)(x+tv)+¸y. Now, the assertion easily follows. ¤ If X is a normed space, then obviously int(C) ½ a-int(C). The following theo- rem states some su±cient conditions for equality of the two sets. Before stating it we recall the following fundamental theorem of R.L. Baire, which we state in the following form. Theorem 0.5 (Baire Category Theorem). Let T be either a complete metric space or a Hausdor® locally compact topologicalS space. If Fn (n 2 N) are closed sets without interior points, then also the union n2N Fn has no interior points. Exercise 0.6. Use the Baire Category Theorem to prove the following two facts. 1. Every compact countable set in a Hausdor® topological space has at least one isolated point. 2. The algebraic dimension of a Banach space is either ¯nite or uncountable. Theorem 0.7. Let C be a convex set in a normed linear space X. Then int(C) = a-int(C) provided at least one of the following conditions is satis¯ed. (a) int(C) 6= ;; (b) C is ¯nite-dimensional; (c) X is Banach and C is an Fσ-set. Proof. (a) Let x 2 a-int(C). Fix z0 2 int(C). If x 6= z0, there exists y 2 C such that x 2 (z0; y). Then x 2 int(C) by Corollary 0.2. (b) Let x 2 a-int(C). This implies that a®(C) = X and hence X is ¯nite- dimensional. By the \relative interior theorem", int(C) = ri(C) is nonempty. Apply the part (a). S (c) Let x 2 a-int(C). We can suppose that x = 0. Write C = k Fk where each Fk is a closed set. Then [ [ [ X = tC = nC = nFk : t>0 n n;k By the Baire Category Theorem, there exist n; k such that nFk has a nonempty interior. Thus int(Fk) 6= ; and hence int(C) 6= ;. Apply (a). ¤ Directions from a point. Let C be a convex set in a vector space X, x0 2 C. Let us consider two sets of directions: the set c(x0;C) of directions of the half-lines from x0 to the points of C, and the set i(x0;C) of directions of the lines L such that L \ C is a nondegenerate segment containing x0 in its relative interior. More precisely, c(x0;C) = fv 2 V : x0 + tv 2 C for some t > 0g ; i(x0;C) = fv 2 V : x0 § tv 2 C for some t > 0g : Let us state main properties of these two sets. 3 Proposition 0.8. Let X; C; x0 be as above. (a) x0 2 x0 + i(x0;C) ½ x0 + c(x0;C) ½ a®(C). (b) c(x0;C) = cone(C ¡ x0). (c) c(x0;C) is a convex cone,£ and i(¤x0;C) is a linear set. (d) i(x0;C) = c(x0;C) \ ¡c(x0;C) . (e) x0 + i(x0;C) is£ the largest a±ne set¤ A such that x0 2 intA(C \ A). (f) a®(C) = x0 + c(x0;C) ¡ c(x0;C) . Proof. (b) follows directly form the de¯nitions; hence c(x0;C) is a convex cone. Obviously, we have 0 2 i(x0;C) ½ c(x0;C). Since x0 + c(x0;C) is the set of all closed half-lines from x0 to the points of C (or fx0g if there are no such half-lines), it is contained in a®(C); hence (a) holds. The part (d) follows easily from de¯nitions. Since K \ (¡K) is a linear set whenever K is a convex cone, we have proved (c), too. Let us show (e). Clearly, x0 + i(x0;C) is an a±ne set. Let A be an a±ne set such Ã! that x0 2 int(C \ A). For every point x 2 A n x0 the set C \ x0x is a nondegenerate segment containing x0 in its relative interior. Since x0 + i(x0;C) is the union of all possible lines L with this property, we have A ½ x0 + i(x0;C). It remains to prove (f). By translation, we can suppose that x0 = 0. Thus, by (b), we have to show that, if 0 2 C, then span(C) = cone(C) ¡ cone(C). The inclusion P\⊃" is obvious. To show the other one, consider a (¯nite) linear combination x = i ®ici with ®i 6= 0, ci 2 C. Then X X x = ®ici ¡ (¡®i)ci 2 cone(C) ¡ cone(C) : ®i>0 ®i<0 ¤ Algebraic relative interior. Let us recall that a relative interior of a convex set C in a normed space is the set ri(C) = inta®(C)(C). Recall also the fact, proved in the chapter on ¯nite-dimensional spaces, that every ¯nite-dimensional convex set has a nonempty relative interior. The algebraic relative interior of C is de¯ned analogously in the following de¯nition. De¯nition 0.9. Let C be a convex set in a vector space X. The algebraic relative interior of C is the set a-ri(C) = a-inta®(C)(C). Observation 0.10. Let C be a convex set in a vector space X, x 2 C. Then x 2 a-ri(C) () x + i(x; C) = a®(C) () x + c(x; C) = a®(C) : Proof. Easy exercise. ¤ Observe that Theorem 0.7 immediately gives the following Corollary 0.11. Let C be a convex set in a normed space. Then a-ri(C) = ri(C) whenever at least one of the following conditions is satis¯ed: (a) ri(C) 6= ;; (b) dim(C) < 1; (c) X is a Banach space, a®(C) is closed, and C is an Fσ-set. 4 Inside points. Let C be a convex set in a normed space X. For x 2 C, let Lx be the family of all lines L ½ X, for which C \L is a nondegenerate segment containing x in its relative interior.
Load more

Recommended publications
  • On Quasi Norm Attaining Operators Between Banach Spaces
    ON QUASI NORM ATTAINING OPERATORS BETWEEN BANACH SPACES GEUNSU CHOI, YUN SUNG CHOI, MINGU JUNG, AND MIGUEL MART´IN Abstract. We provide a characterization of the Radon-Nikod´ymproperty in terms of the denseness of bounded linear operators which attain their norm in a weak sense, which complement the one given by Bourgain and Huff in the 1970's. To this end, we introduce the following notion: an operator T : X ÝÑ Y between the Banach spaces X and Y is quasi norm attaining if there is a sequence pxnq of norm one elements in X such that pT xnq converges to some u P Y with }u}“}T }. Norm attaining operators in the usual (or strong) sense (i.e. operators for which there is a point in the unit ball where the norm of its image equals the norm of the operator) and also compact operators satisfy this definition. We prove that strong Radon-Nikod´ymoperators can be approximated by quasi norm attaining operators, a result which does not hold for norm attaining operators in the strong sense. This shows that this new notion of quasi norm attainment allows to characterize the Radon-Nikod´ymproperty in terms of denseness of quasi norm attaining operators for both domain and range spaces, completing thus a characterization by Bourgain and Huff in terms of norm attaining operators which is only valid for domain spaces and it is actually false for range spaces (due to a celebrated example by Gowers of 1990). A number of other related results are also included in the paper: we give some positive results on the denseness of norm attaining Lipschitz maps, norm attaining multilinear maps and norm attaining polynomials, characterize both finite dimensionality and reflexivity in terms of quasi norm attaining operators, discuss conditions to obtain that quasi norm attaining operators are actually norm attaining, study the relationship with the norm attainment of the adjoint operator and, finally, present some stability results.
    [Show full text]
  • Uniform Boundedness Principle for Unbounded Operators
    UNIFORM BOUNDEDNESS PRINCIPLE FOR UNBOUNDED OPERATORS C. GANESA MOORTHY and CT. RAMASAMY Abstract. A uniform boundedness principle for unbounded operators is derived. A particular case is: Suppose fTigi2I is a family of linear mappings of a Banach space X into a normed space Y such that fTix : i 2 Ig is bounded for each x 2 X; then there exists a dense subset A of the open unit ball in X such that fTix : i 2 I; x 2 Ag is bounded. A closed graph theorem and a bounded inverse theorem are obtained for families of linear mappings as consequences of this principle. Some applications of this principle are also obtained. 1. Introduction There are many forms for uniform boundedness principle. There is no known evidence for this principle for unbounded operators which generalizes classical uniform boundedness principle for bounded operators. The second section presents a uniform boundedness principle for unbounded operators. An application to derive Hellinger-Toeplitz theorem is also obtained in this section. A JJ J I II closed graph theorem and a bounded inverse theorem are obtained for families of linear mappings in the third section as consequences of this principle. Go back Let us assume the following: Every vector space X is over R or C. An α-seminorm (0 < α ≤ 1) is a mapping p: X ! [0; 1) such that p(x + y) ≤ p(x) + p(y), p(ax) ≤ jajαp(x) for all x; y 2 X Full Screen Close Received November 7, 2013. 2010 Mathematics Subject Classification. Primary 46A32, 47L60. Key words and phrases.
    [Show full text]
  • On the Boundary Between Mereology and Topology
    On the Boundary Between Mereology and Topology Achille C. Varzi Istituto per la Ricerca Scientifica e Tecnologica, I-38100 Trento, Italy [email protected] (Final version published in Roberto Casati, Barry Smith, and Graham White (eds.), Philosophy and the Cognitive Sciences. Proceedings of the 16th International Wittgenstein Symposium, Vienna, Hölder-Pichler-Tempsky, 1994, pp. 423–442.) 1. Introduction Much recent work aimed at providing a formal ontology for the common-sense world has emphasized the need for a mereological account to be supplemented with topological concepts and principles. There are at least two reasons under- lying this view. The first is truly metaphysical and relates to the task of charac- terizing individual integrity or organic unity: since the notion of connectedness runs afoul of plain mereology, a theory of parts and wholes really needs to in- corporate a topological machinery of some sort. The second reason has been stressed mainly in connection with applications to certain areas of artificial in- telligence, most notably naive physics and qualitative reasoning about space and time: here mereology proves useful to account for certain basic relation- ships among things or events; but one needs topology to account for the fact that, say, two events can be continuous with each other, or that something can be inside, outside, abutting, or surrounding something else. These motivations (at times combined with others, e.g., semantic transpar- ency or computational efficiency) have led to the development of theories in which both mereological and topological notions play a pivotal role. How ex- actly these notions are related, however, and how the underlying principles should interact with one another, is still a rather unexplored issue.
    [Show full text]
  • Baire Category Theorem and Uniform Boundedness Principle)
    Functional Analysis (WS 19/20), Problem Set 3 (Baire Category Theorem and Uniform Boundedness Principle) Baire Category Theorem B1. Let (X; k·kX ) be an innite dimensional Banach space. Prove that X has uncountable Hamel basis. Note: This is Problem A2 from Problem Set 1. B2. Consider subset of bounded sequences 1 A = fx 2 l : only nitely many xk are nonzerog : Can one dene a norm on A so that it becomes a Banach space? Consider the same question with the set of polynomials dened on interval [0; 1]. B3. Prove that the set L2(0; 1) has empty interior as the subset of Banach space L1(0; 1). B4. Let f : [0; 1) ! [0; 1) be a continuous function such that for every x 2 [0; 1), f(kx) ! 0 as k ! 1. Prove that f(x) ! 0 as x ! 1. B5.( Uniform Boundedness Principle) Let (X; k · kX ) be a Banach space and (Y; k · kY ) be a normed space. Let fTαgα2A be a family of bounded linear operators between X and Y . Suppose that for any x 2 X, sup kTαxkY < 1: α2A Prove that . supα2A kTαk < 1 Uniform Boundedness Principle U1. Let F be a normed space C[0; 1] with L2(0; 1) norm. Check that the formula 1 Z n 'n(f) = n f(t) dt 0 denes a bounded linear functional on . Verify that for every , F f 2 F supn2N j'n(f)j < 1 but . Why Uniform Boundedness Principle is not satised in this case? supn2N k'nk = 1 U2.( pointwise convergence of operators) Let (X; k · kX ) be a Banach space and (Y; k · kY ) be a normed space.
    [Show full text]
  • FUNCTIONAL ANALYSIS 1. Banach and Hilbert Spaces in What
    FUNCTIONAL ANALYSIS PIOTR HAJLASZ 1. Banach and Hilbert spaces In what follows K will denote R of C. Definition. A normed space is a pair (X, k · k), where X is a linear space over K and k · k : X → [0, ∞) is a function, called a norm, such that (1) kx + yk ≤ kxk + kyk for all x, y ∈ X; (2) kαxk = |α|kxk for all x ∈ X and α ∈ K; (3) kxk = 0 if and only if x = 0. Since kx − yk ≤ kx − zk + kz − yk for all x, y, z ∈ X, d(x, y) = kx − yk defines a metric in a normed space. In what follows normed paces will always be regarded as metric spaces with respect to the metric d. A normed space is called a Banach space if it is complete with respect to the metric d. Definition. Let X be a linear space over K (=R or C). The inner product (scalar product) is a function h·, ·i : X × X → K such that (1) hx, xi ≥ 0; (2) hx, xi = 0 if and only if x = 0; (3) hαx, yi = αhx, yi; (4) hx1 + x2, yi = hx1, yi + hx2, yi; (5) hx, yi = hy, xi, for all x, x1, x2, y ∈ X and all α ∈ K. As an obvious corollary we obtain hx, y1 + y2i = hx, y1i + hx, y2i, hx, αyi = αhx, yi , Date: February 12, 2009. 1 2 PIOTR HAJLASZ for all x, y1, y2 ∈ X and α ∈ K. For a space with an inner product we define kxk = phx, xi . Lemma 1.1 (Schwarz inequality).
    [Show full text]
  • POLARS and DUAL CONES 1. Convex Sets
    POLARS AND DUAL CONES VERA ROSHCHINA Abstract. The goal of this note is to remind the basic definitions of convex sets and their polars. For more details see the classic references [1, 2] and [3] for polytopes. 1. Convex sets and convex hulls A convex set has no indentations, i.e. it contains all line segments connecting points of this set. Definition 1 (Convex set). A set C ⊆ Rn is called convex if for any x; y 2 C the point λx+(1−λ)y also belongs to C for any λ 2 [0; 1]. The sets shown in Fig. 1 are convex: the first set has a smooth boundary, the second set is a Figure 1. Convex sets convex polygon. Singletons and lines are also convex sets, and linear subspaces are convex. Some nonconvex sets are shown in Fig. 2. Figure 2. Nonconvex sets A line segment connecting two points x; y 2 Rn is denoted by [x; y], so we have [x; y] = f(1 − α)x + αy; α 2 [0; 1]g: Likewise, we can consider open line segments (x; y) = f(1 − α)x + αy; α 2 (0; 1)g: and half open segments such as [x; y) and (x; y]. Clearly any line segment is a convex set. The notion of a line segment connecting two points can be generalised to an arbitrary finite set n n by means of a convex combination. Given a finite set fx1; x2; : : : ; xpg ⊂ R , a point x 2 R is the convex combination of x1; x2; : : : ; xp if it can be represented as p p X X x = αixi; αi ≥ 0 8i = 1; : : : ; p; αi = 1: i=1 i=1 Given an arbitrary set S ⊆ Rn, we can consider all possible finite convex combinations of points from this set.
    [Show full text]
  • A Study of Optimization in Hilbert Space
    California State University, San Bernardino CSUSB ScholarWorks Theses Digitization Project John M. Pfau Library 1998 A study of optimization in Hilbert Space John Awunganyi Follow this and additional works at: https://scholarworks.lib.csusb.edu/etd-project Part of the Mathematics Commons Recommended Citation Awunganyi, John, "A study of optimization in Hilbert Space" (1998). Theses Digitization Project. 1459. https://scholarworks.lib.csusb.edu/etd-project/1459 This Project is brought to you for free and open access by the John M. Pfau Library at CSUSB ScholarWorks. It has been accepted for inclusion in Theses Digitization Project by an authorized administrator of CSUSB ScholarWorks. For more information, please contact [email protected]. A STUDY OF OPTIMIZATIONIN HELBERT SPACE A Project Presented to the Faculty of California State University, San Bernardino In Partial Fulfillment ofthe Requirements for the Degree Master ofArts in Mathematics by John Awunganyi June 1998 A STUDY OF OPTIMIZATION IN HILBERT SPACE A Project Presented to the Faculty of California State University, San Bernardino by John Awunganyi June 1998 Approved by; Dr. Chetan Prakash, Chair, Mathematics Date r. Wenxang Wang Dr. Christopher Freiling Dr. Terry Hallett, Graduate Coordinator Dr. Peter Williams, Department Chair ABSTRACT The primary objective ofthis project is to demonstrate that a certain field ofoptimization can be effectively unified by afew geometric principles ofcomplete normed linear space theory. By employing these principles, important and complex finite - dimensional problems can be interpreted and solved by methods springing from geometric insight. Concepts such as distance, orthogonality, and convexity play a fundamental and indispensable role in this development. Viewed in these terms, seemingly diverse problems and techniques often are found to be closely related.
    [Show full text]
  • On the Ekeland Variational Principle with Applications and Detours
    Lectures on The Ekeland Variational Principle with Applications and Detours By D. G. De Figueiredo Tata Institute of Fundamental Research, Bombay 1989 Author D. G. De Figueiredo Departmento de Mathematica Universidade de Brasilia 70.910 – Brasilia-DF BRAZIL c Tata Institute of Fundamental Research, 1989 ISBN 3-540- 51179-2-Springer-Verlag, Berlin, Heidelberg. New York. Tokyo ISBN 0-387- 51179-2-Springer-Verlag, New York. Heidelberg. Berlin. Tokyo No part of this book may be reproduced in any form by print, microfilm or any other means with- out written permission from the Tata Institute of Fundamental Research, Colaba, Bombay 400 005 Printed by INSDOC Regional Centre, Indian Institute of Science Campus, Bangalore 560012 and published by H. Goetze, Springer-Verlag, Heidelberg, West Germany PRINTED IN INDIA Preface Since its appearance in 1972 the variational principle of Ekeland has found many applications in different fields in Analysis. The best refer- ences for those are by Ekeland himself: his survey article [23] and his book with J.-P. Aubin [2]. Not all material presented here appears in those places. Some are scattered around and there lies my motivation in writing these notes. Since they are intended to students I included a lot of related material. Those are the detours. A chapter on Nemyt- skii mappings may sound strange. However I believe it is useful, since their properties so often used are seldom proved. We always say to the students: go and look in Krasnoselskii or Vainberg! I think some of the proofs presented here are more straightforward. There are two chapters on applications to PDE.
    [Show full text]
  • Nonlinear Monotone Operators with Values in 9(X, Y)
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector JOURNAL OF MATHEMATICAL ANALYSIS AND APPLICATIONS 140, 83-94 (1989) Nonlinear Monotone Operators with Values in 9(X, Y) N. HADJISAVVAS, D. KRAVVARITIS, G. PANTELIDIS, AND I. POLYRAKIS Department of Mathematics, National Technical University of Athens, Zografou Campus, 15773 Athens, Greece Submitted by C. Foias Received April 2, 1987 1. INTRODUCTION Let X be a topological vector space, Y an ordered topological vector space, and 9(X, Y) the space of all linear and continuous mappings from X into Y. If T is an operator from X into 9(X, Y) (generally multivalued) with domain D(T), T is said to be monotone if for all xieD(T) and Aie T(x,), i= 1, 2. In the scalar case Y = R, this definition coincides with the well known definition of monotone operator (cf. [ 12,4]). An important subclass of monotone operators T: X+ 9(X, Y) consists of the subdifferentials of convex operators from X into Y, which have been studied by Valadier [ 171, Kusraev and Kutateladze [ 111, Papageorgiou [ 131, and others. Some properties of monotone operators have been investigated by Kirov [7, 81, in connection with the study of the differentiability of convex operators. In this paper we begin our investigation of the properties of monotone operators from X into 3(X, Y) by introducing, in Section 3, the notion of hereditary order convexity of subsets of 9(X, Y). This notion plays a key role in our discussion and expresses a kind of separability between sets and points, Its interest relies on the fact that the images of maximal monotone operators are hereditarily order-convex.
    [Show full text]
  • Contents 1. Introduction 1 2. Cones in Vector Spaces 2 2.1. Ordered Vector Spaces 2 2.2
    ORDERED VECTOR SPACES AND ELEMENTS OF CHOQUET THEORY (A COMPENDIUM) S. COBZAS¸ Contents 1. Introduction 1 2. Cones in vector spaces 2 2.1. Ordered vector spaces 2 2.2. Ordered topological vector spaces (TVS) 7 2.3. Normal cones in TVS and in LCS 7 2.4. Normal cones in normed spaces 9 2.5. Dual pairs 9 2.6. Bases for cones 10 3. Linear operators on ordered vector spaces 11 3.1. Classes of linear operators 11 3.2. Extensions of positive operators 13 3.3. The case of linear functionals 14 3.4. Order units and the continuity of linear functionals 15 3.5. Locally order bounded TVS 15 4. Extremal structure of convex sets and elements of Choquet theory 16 4.1. Faces and extremal vectors 16 4.2. Extreme points, extreme rays and Krein-Milman's Theorem 16 4.3. Regular Borel measures and Riesz' Representation Theorem 17 4.4. Radon measures 19 4.5. Elements of Choquet theory 19 4.6. Maximal measures 21 4.7. Simplexes and uniqueness of representing measures 23 References 24 1. Introduction The aim of these notes is to present a compilation of some basic results on ordered vector spaces and positive operators and functionals acting on them. A short presentation of Choquet theory is also included. They grew up from a talk I delivered at the Seminar on Analysis and Optimization. The presentation follows mainly the books [3], [9], [19], [22], [25], and [11], [23] for the Choquet theory. Note that the first two chapters of [9] contains a thorough introduction (with full proofs) to some basics results on ordered vector spaces.
    [Show full text]
  • On Barycenters of Probability Measures
    On barycenters of probability measures Sergey Berezin Azat Miftakhov Abstract A characterization is presented of barycenters of the Radon probability measures supported on a closed convex subset of a given space. A case of particular interest is studied, where the underlying space is itself the space of finite signed Radon measures on a metric compact and where the corresponding support is the convex set of probability measures. For locally compact spaces, a simple characterization is obtained in terms of the relative interior. 1. The main goal of the present note is to characterize barycenters of the Radon probability measures supported on a closed convex set. Let X be a Fr´echet space. Without loss of generality, the topology on X is generated by the translation-invariant metric ρ on X (for details see [2]). We denote the set of Radon probability measures on X by P(X). By definition, the barycen- ter a ∈ X of a measure µ ∈ P(X) is a = xµ(dx) (1) XZ if the latter integral exists in the weak sense, that is, if the following holds Λa = Λxµ(dx) (2) XZ for all Λ ∈ X∗, where X∗ is the topological dual of X. More details on such integrals can be found in [2, Chapter 3]. Note that if (1) exists, then a = xµ(dx), (3) suppZ µ arXiv:1911.07680v2 [math.PR] 20 Apr 2020 and, by the Hahn–Banach separation theorem, one has a ∈ co(supp µ), where co(·) stands for the convex hull. From now on, we will use the bar over a set in order to denote its topological closure.
    [Show full text]
  • The Ekeland Variational Principle, the Bishop-Phelps Theorem, and The
    The Ekeland Variational Principle, the Bishop-Phelps Theorem, and the Brøndsted-Rockafellar Theorem Our aim is to prove the Ekeland Variational Principle which is an abstract result that found numerous applications in various fields of Mathematics. As its application to Convex Analysis, we provide a proof of the famous Bishop- Phelps Theorem and some related results. Let us recall that the epigraph and the hypograph of a function f : X ! [−∞; +1] (where X is a set) are the following subsets of X × R: epi f = f(x; t) 2 X × R : t ≥ f(x)g ; hyp f = f(x; t) 2 X × R : t ≤ f(x)g : Notice that, if f > −∞ on X then epi f 6= ; if and only if f is proper, that is, f is finite at at least one point. 0.1. The Ekeland Variational Principle. In what follows, (X; d) is a metric space. Given λ > 0 and (x; t) 2 X × R, we define the set Kλ(x; t) = f(y; s): s ≤ t − λd(x; y)g = f(y; s): λd(x; y) ≤ t − sg ⊂ X × R: Notice that Kλ(x; t) = hyp[t − λ(x; ·)]. Let us state some useful properties of these sets. Lemma 0.1. (a) Kλ(x; t) is closed and contains (x; t). (b) If (¯x; t¯) 2 Kλ(x; t) then Kλ(¯x; t¯) ⊂ Kλ(x; t). (c) If (xn; tn) ! (x; t) and (xn+1; tn+1) 2 Kλ(xn; tn) (n 2 N), then \ Kλ(x; t) = Kλ(xn; tn) : n2N Proof. (a) is quite easy.
    [Show full text]