An Easy Path to Convex Analysis and Applications
|
|
- Wesley Bates
- 6 years ago
- Views:
Transcription
1 M & C Morgan & Claypool Publishers An Easy Path to Convex Analysis and Applications Boris S. Mordukhovich Nguyen Mau Nam SYNTHESIS LECTURES ON MATHEMATICS AND STATISTICS Steven G. Krantz, Series Editor
2
3 An Easy Path to Convex Analysis and Applications
4 Synthesis Lectures on Mathematics and Statistics Editor Steven G. Krantz, Washington University, St. Louis An Easy Path to Convex Analysis and Applications Boris S. Mordukhovich and Nguyen Mau Nam 2014 Applications of Affine and Weyl Geometry Eduardo García-Río, Peter Gilkey, Stana Nikčević, and Ramón Vázquez-Lorenzo 2013 Essentials of Applied Mathematics for Engineers and Scientists, Second Edition Robert G. Watts 2012 Chaotic Maps: Dynamics, Fractals, and Rapid Fluctuations Goong Chen and Yu Huang 2011 Matrices in Engineering Problems Marvin J. Tobias 2011 e Integral: A Crux for Analysis Steven G. Krantz 2011 Statistics is Easy! Second Edition Dennis Shasha and Manda Wilson 2010 Lectures on Financial Mathematics: Discrete Asset Pricing Greg Anderson and Alec N. Kercheval 2010
5 Jordan Canonical Form: eory and Practice Steven H. Weintraub 2009 iii e Geometry of Walker Manifolds Miguel Brozos-Vázquez, Eduardo García-Río, Peter Gilkey, Stana Nikčević, and Ramón Vázquez-Lorenzo 2009 An Introduction to Multivariable Mathematics Leon Simon 2008 Jordan Canonical Form: Application to Differential Equations Steven H. Weintraub 2008 Statistics is Easy! Dennis Shasha and Manda Wilson 2008 A Gyrovector Space Approach to Hyperbolic Geometry Abraham Albert Ungar 2008
6 Copyright 2014 by Morgan & Claypool All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means electronic, mechanical, photocopy, recording, or any other except for brief quotations in printed reviews, without the prior permission of the publisher. An Easy Path to Convex Analysis and Applications Boris S. Mordukhovich and Nguyen Mau Nam ISBN: ISBN: paperback ebook DOI /S00554ED1V01Y201312MAS014 A Publication in the Morgan & Claypool Publishers series SYNTHESIS LECTURES ON MATHEMATICS AND STATISTICS Lecture #14 Series Editor: Steven G. Krantz, Washington University, St. Louis Series ISSN Synthesis Lectures on Mathematics and Statistics Print Electronic
7 An Easy Path to Convex Analysis and Applications Boris S. Mordukhovich Wayne State University Nguyen Mau Nam Portland State University SYNTHESIS LECTURES ON MATHEMATICS AND STATISTICS #14 & M C Morgan & claypool publishers
8 ABSTRACT Convex optimization has an increasing impact on many areas of mathematics, applied sciences, and practical applications. It is now being taught at many universities and being used by researchers of different fields. As convex analysis is the mathematical foundation for convex optimization, having deep knowledge of convex analysis helps students and researchers apply its tools more effectively. e main goal of this book is to provide an easy access to the most fundamental parts of convex analysis and its applications to optimization. Modern techniques of variational analysis are employed to clarify and simplify some basic proofs in convex analysis and build the theory of generalized differentiation for convex functions and sets in finite dimensions. We also present new applications of convex analysis to location problems in connection with many interesting geometric problems such as the Fermat-Torricelli problem, the Heron problem, the Sylvester problem, and their generalizations. Of course, we do not expect to touch every aspect of convex analysis, but the book consists of sufficient material for a first course on this subject. It can also serve as supplemental reading material for a course on convex optimization and applications. KEYWORDS Affine set, Carathéodory theorem, convex function, convex set, directional derivative, distance function, Fenchel conjugate, Fermat-Torricelli problem, generalized differentiation, Helly theorem, minimal time function, Nash equilibrium, normal cone, Radon theorem, optimal value function, optimization, smallest enclosing ball problem, set-valued mapping, subdifferential, subgradient, subgradient algorithm, support function, Weiszfeld algorithm
9 vii e first author dedicates this book to the loving memory of his father SHOLIM MORDUKHOVICH ( ), a kind man and a brave warrior. e second author dedicates this book to the memory of his parents.
10
11 ix Contents Preface xi Acknowledgments xiii List of Symbols xv 1 Convex Sets and Functions Preliminaries Convex Sets Convex Functions Relative Interiors of Convex Sets e Distance Function Exercises for Chapter Subdifferential Calculus Convex Separation Normals to Convex Sets Lipschitz Continuity of Convex Functions Subgradients of Convex Functions Basic Calculus Rules Subgradients of Optimal Value Functions Subgradients of Support Functions Fenchel Conjugates Directional Derivatives Subgradients of Supremum Functions Exercises for Chapter Remarkable Consequences of Convexity Characterizations of Differentiability Carathéodory eorem and Farkas Lemma Radon eorem and Helly eorem
12 x 3.4 Tangents to Convex Sets Mean Value eorems Horizon Cones Minimal Time Functions and Minkowski Gauge Subgradients of Minimal Time Functions Nash Equilibrium Exercises for Chapter Applications to Optimization and Location Problems Lower Semicontinuity and Existence of Minimizers Optimality Conditions Subgradient Methods in Convex Optimization e Fermat-Torricelli Problem A Generalized Fermat-Torricelli Problem A Generalized Sylvester Problem Exercises for Chapter Solutions and Hints for Exercises Bibliography Authors Biographies Index
13 xi Preface Some geometric properties of convex sets and, to a lesser extent, of convex functions had been studied before the 1960s by many outstanding mathematicians, first of all by Hermann Minkowski and Werner Fenchel. At the beginning of the 1960s convex analysis was greatly developed in the works of R. Tyrrell Rockafellar and Jean-Jacques Moreau who initiated a systematic study of this new field. As a fundamental part of variational analysis, convex analysis contains a generalized differentiation theory that can be used to study a large class of mathematical models with no differentiability assumptions imposed on their initial data. e importance of convex analysis for many applications in which convex optimization is the first to name has been well recognized. e presence of convexity makes it possible not only to comprehensively investigate qualitative properties of optimal solutions and derive necessary and sufficient conditions for optimality but also to develop effective numerical algorithms to solve convex optimization problems, even with nondifferentiable data. Convex analysis and optimization have an increasing impact on many areas of mathematics and applications including control systems, estimation and signal processing, communications and networks, electronic circuit design, data analysis and modeling, statistics, economics and finance, etc. ere are several fundamental books devoted to different aspects of convex analysis and optimization. Among them we mention Convex Analysis by Rockafellar [26], Convex Analysis and Minimization Algorithms (in two volumes) by Hiriart-Urruty and Lemaréchal [8] and its abridge version [9], Convex Analysis and Nonlinear Optimization by Borwein and Lewis [4], Introductory Lectures on Convex Optimization by Nesterov [21], and Convex Optimization by Boyd and Vandenberghe [3] as well as other books listed in the bibliography below. In this big picture of convex analysis and optimization, our book serves as a bridge for students and researchers who have just started using convex analysis to reach deeper topics in the field. We give detailed proofs for most of the results presented in the book and also include many figures and exercises for better understanding the material. e powerful geometric approach developed in modern variational analysis is adopted and simplified in the convex case in order to provide the reader with an easy path to access generalized differentiation of convex objects in finite dimensions. In this way, the book also serves as a starting point for the interested reader to continue the study of nonconvex variational analysis and applications. It can be of interest from this viewpoint to experts in convex and variational analysis. Finally, the application part of this book not only concerns the classical topics of convex optimization related to optimality conditions and subgradient algorithms but also presents some recent while easily understandable qualitative and numerical results for important location problems.
14 xii PREFACE e book consists of four chapters and is organized as follows. In Chapter 1 we study fundamental properties of convex sets and functions while paying particular attention to classes of convex functions important in optimization. Chapter 2 is mainly devoted to developing basic calculus rules for normals to convex sets and subgradients of convex functions that are in the mainstream of convex theory. Chapter 3 concerns some additional topics of convex analysis that are largely used in applications. Chapter 4 is fully devoted to applications of basic results of convex analysis to problems of convex optimization and selected location problems considered from both qualitative and numerical viewpoints. Finally, we present at the end of the book Solutions and Hints to selected exercises. Exercises are given at the end of each chapter while figures and examples are provided throughout the whole text. e list of references contains books and selected papers, which are closely related to the topics considered in the book and may be helpful to the reader for advanced studies of convex analysis, its applications, and further extensions. Since only elementary knowledge in linear algebra and basic calculus is required, this book can be used as a textbook for both undergraduate and graduate level courses in convex analysis and its applications. In fact, the authors have used these lecture notes for teaching such classes in their universities as well as while visiting some other schools. We hope that the book will make convex analysis more accessible to large groups of undergraduate and graduate students, researchers in different disciplines, and practitioners. Boris S. Mordukhovich and Nguyen Mau Nam December 2013
15 xiii Acknowledgments e first author acknowledges continued support by grants from the National Science Foundation and the second author acknowledges support by the Simons Foundation. Boris S. Mordukhovich and Nguyen Mau Nam December 2013
16
17 xv List of Symbols R the real numbers R D.1; 1 the extended real line R C the nonnegative real numbers R > the positive real numbers N the positive integers co convex hull of aff affine hull of int interior of ri relative interior of span linear subspace generated by cone cone generated by K convex cone generated by dim dimension of closure of bd boundary of IB closed unit ball IB. NxI r/ closed ball with center Nx and radius r dom f domain of f epi f epigraph of f gph F graph of mapping F LŒa; b line connecting a and b d.xi / distance from x to.xi / projection of x to hx; yi inner product of x and y A adjoint/transpose of linear mapping/matrix A N. NxI / normal cone to at Nx ker A kernel of linear mapping A D F. Nx; Ny/ coderivative to F at. Nx; Ny/ T. NxI / tangent cone to at Nx F 1 horizon/asymptotic cone of F T F minimal time function defined by constant dynamic F and target set L.x; / Lagrangian Minkowski gauge of F F
18 xvi LIST OF SYMBOLS F.I / generalized projection defined by the minimal time function G F. Nx; t/ generalized ball defined by dynamic F with center Nx and radius t
19 C H A P T E R 1 Convex Sets and Functions 1 is chapter presents definitions, examples, and basic properties of convex sets and functions in the Euclidean space R n and also contains some related material. 1.1 PRELIMINARIES We start with reviewing classical notions and properties of the Euclidean space R n. e proofs of the results presented in this section can be found in standard books on advanced calculus and linear algebra. Let us denote by R n the set of all ntuples of real numbers x D.x 1 ; : : : ; x n /. en R n is a linear space with the following operations:.x 1 ; : : : ; x n / C.y 1 ; : : : ; y n / D.x 1 C y 1 ; : : : ; x n C y n /;.x 1 ; : : : ; x n / D.x 1 ; : : : ; x n /; where.x 1 ; : : : ; x n /;.y 1 ; : : : ; y n / 2 R n and 2 R. e zero element of R n and the number zero of R are often denoted by the same notation 0 if no confusion arises. For any x D.x 1 ; : : : ; x n / 2 R n, we identify it with the column vector x D Œx 1 ; : : : ; x n T, where the symbol T stands for vector transposition. Given x D.x 1 ; : : : ; x n / 2 R n and y D.y 1 ; : : : ; y n / 2 R n, the inner product of x and y is defined by hx; yi WD nx x i y i : e following proposition lists some important properties of the inner product in R n. Proposition 1.1 For x; y; z 2 R n and 2 R, we have: (i) hx; xi 0, and hx; xi D 0 if and only if x D 0. (ii) hx; yi D hy; xi. (iii) hx; yi D hx; yi. (iv) hx; y C zi D hx; yi C hx; zi. e Euclidean norm of x D.x 1 ; : : : ; x n / 2 R n is defined by q kxk WD x1 2 C : : : C x2 n :
20 2 1. CONVEX SETS AND FUNCTIONS It follows directly from the definition that kxk D p hx; xi. Proposition 1.2 For any x; y 2 R n and 2 R, we have: (i) kxk 0, and kxk D 0 if and only if x D 0. (ii) kxk D jj kxk. (iii) kx C yk kxk C kyk.the triangle inequality/. (iv) jhx; yij kxk kyk.the Cauchy-Schwarz inequality/. Using the Euclidean norm allows us to introduce the balls in R n, which can be used to define other topological notions in R n. Definition 1.3 e CLOSED BALL centered at Nx with radius r 0 and the CLOSED UNIT BALL of R n are defined, respectively, by IB. NxI r/ WD x 2 R n ˇˇ kx Nxk r and IB WD x 2 R n ˇˇ kxk 1 : It is easy to see that IB D IB.0I 1/ and IB. NxI r/ D Nx C rib. Definition 1.4 Let R n. en Nx is an INTERIOR POINT of if there is ı > 0 such that IB. NxI ı/ : e set of all interior points of is denoted by int. Moreover, is said to be OPEN if every point of is its interior point. We get that is open if and only if for every Nx 2 there is ı > 0 such that IB. NxI ı/. It is obvious that the empty set ; and the whole space R n are open. Furthermore, any open ball B. NxI r/ WD fx 2 R n j kx Nxk < rg centered at Nx with radius r is open. Definition 1.5 A set R n is CLOSED if its complement c D R n n is open in R n. It follows that the empty set, the whole space, and any ball IB. NxI r/ are closed in R n. Proposition 1.6 (i) e union of any collection of open sets in R n is open. (ii) e intersection of any finite collection of open sets in R n is open. (iii) e intersection of any collection of closed sets in R n is closed. (iv) e union of any finite collection of closed sets in R n is closed. Definition 1.7 Let fx k g be a sequence in R n. We say that fx k g CONVERGES to Nx if kx k Nxk! 0 as k! 1. In this case we write lim x k D Nx: k!1
21 1.1. PRELIMINARIES 3 is notion allows us to define the following important topological concepts for sets. Definition 1.8 Let be a nonempty subset of R n. en: (i) e CLOSURE of, denoted by or cl, is the collection of limits of all convergent sequences belonging to. (ii) e BOUNDARY of, denoted by bd, is the set n int. We can see that the closure of is the intersection of all closed sets containing and that the interior of is the union of all open sets contained in. It follows from the definition that Nx 2 if and only if for any ı > 0 we have IB. NxI ı/ \ ;. Furthermore, Nx 2 bd if and only if for any ı > 0 the closed ball IB. NxI / intersects both sets and its complement c. Definition 1.9 Let fx k g be a sequence in R n and let fk`g be a strictly increasing sequence of positive integers. en the new sequence fx k`g is called a SUBSEQUENCE of fx k g. We say that a set is bounded if it is contained in a ball centered at the origin with some radius r > 0, i.e., IB.0I r/. us a sequence fx k g is bounded if there is r > 0 with kx k k r for all k 2 N : e following important result is known as the Bolzano-Weierstrass theorem. eorem 1.10 Any bounded sequence in R n contains a convergent subsequence. e next concept plays a very significant role in analysis and optimization. Definition 1.11 We say that a set is COMPACT in R n if every sequence in contains a subsequence converging to some point in. e following result is a consequence of the Bolzano-Weierstrass theorem. eorem 1.12 A subset of R n is compact if and only if it is closed and bounded. For subsets, 1, and 2 of R n and for 2 R, we define the operations: 1 C 2 WD x C y ˇˇ x 2 1; y 2 2 ; WD x ˇˇ x 2 : e next proposition can be proved easily. Proposition 1.13 Let 1 and 2 be two subsets of R n. (i) If 1 is open or 2 is open, then 1 C 2 is open. (ii) If 1 is closed and 2 is compact, then 2 C 2 is closed.
22 4 1. CONVEX SETS AND FUNCTIONS Recall now the notions of bounds for subsets of the real line. Definition 1.14 have Let D be a subset of the real line. A number m 2 R is a LOWER BOUND of D if we x m for all x 2 D: If the set D has a lower bound, then it is BOUNDED BELOW. Similarly, a number M 2 R is an UPPER BOUND of D if x M for all x 2 D; and D is BOUNDED ABOVE if it has an upper bound. Furthermore, we say that the set D is BOUNDED if it is simultaneously bounded below and above. Now we are ready to define the concepts of infimum and supremum of sets. Definition 1.15 Let D R be nonempty and bounded below. e infimum of D, denoted by inf D, is the greatest lower bound of D. When D is nonempty and bounded above, its supremum, denoted by sup D, is the least upper bound of D. If D is not bounded below.resp. above/, then we set inf D WD 1.resp. sup D WD 1/. We also use the convention that inf ; WD 1 and sup ; WD 1. e following fundamental axiom ensures that these notions are well-defined. Completeness Axiom. For every nonempty subset D of R that is bounded above, the least upper bound of D exists as a real number. Using the Completeness Axiom, it is easy to see that if a nonempty set is bounded below, then its greatest lower bound exists as a real number. roughout the book we consider for convenience extended-real-valued functions, which take values in R WD.1; 1. e usual conventions of extended arithmetics are that a C 1 D 1 for any a 2 R, 1 C 1 D 1, and t 1 D 1 for t > 0. Definition 1.16 Let f W! R be an extended-real-valued function and let Nx 2 with f. Nx/ < 1. en f is CONTINUOUS at Nx if for any > 0 there is ı > 0 such that jf.x/ f. Nx/j < whenever kx Nxk < ı; x 2 : We say that f is continuous on if it is continuous at every point of. It is obvious from the definition that if f W! R is continuous at Nx (with f. Nx/ < 1), then it is finite on the intersection of and a ball centered at Nx with some radius r > 0. Furthermore, f W! R is continuous at Nx (with f. Nx/ < 1) if and only if for every sequence fx k g in converging to Nx the sequence ff.x k /g converges to f. Nx/.
23 1.2. CONVEX SETS 5 Definition 1.17 Let f W! R and let Nx 2 with f. Nx/ < 1. We say that f has a LOCAL MIN- IMUM at Nx relative to if there is ı > 0 such that f.x/ f. Nx/ for all x 2 IB. NxI ı/ \ : We also say that f has a GLOBAL/ABSOLUTE MINIMUM at Nx relative to if f.x/ f. Nx/ for all x 2 : e notions of local and global maxima can be defined similarly. Finally in this section, we formulate a fundamental result of mathematical analysis and optimization known as the Weierstrass existence theorem. eorem 1.18 Let f W! R be a continuous function, where is a nonempty, compact subset of R n. en there exist Nx 2 and Nu 2 such that f. Nx/ D inf f.x/ ˇˇ x 2 and f. Nu/ D sup f.x/ ˇˇ x 2 : In Section 4.1 we present some unilateral versions of eorem CONVEX SETS We start the study of convexity with sets and then proceed to functions. Geometric ideas play an underlying role in convex analysis, its extensions, and applications. us we implement the geometric approach in this book. Given two elements a and b in R n, define the interval/line segment Œa; b WD a C.1 /b ˇˇ 2 Œ0; 1 : Note that if a D b, then this interval reduces to a singleton Œa; b D fag. Definition 1.19 A subset of R n is CONVEX if Œa; b whenever a; b 2. Equivalently, is convex if a C.1 /b 2 for all a; b 2 and 2 Œ0; 1. Given! 1 ; : : : ;! m 2 R n, the element x D P m i! i, where P m i D 1 and i 0 for some m 2 N, is called a convex combination of! 1 ; : : : ;! m. Proposition 1.20 elements. A subset of R n is convex if and only if it contains all convex combinations of its
24 6 1. CONVEX SETS AND FUNCTIONS Figure 1.1: Convex set and nonconvex set. Proof. e sufficient condition is trivial. To justify the necessity, we show by induction that any convex combination x D P m i! i of elements in is an element of. is conclusion follows directly from the definition for m D 1; 2. Fix now a positive integer m 2 and suppose that every convex combination of k 2 N elements from, where k m, belongs to. Form the convex combination y WD mc1 X i! i ; mc1 X i D 1; i 0 and observe that if mc1 D 1, then 1 D 2 D : : : D m D 0, so y D! mc1 2. In the case where mc1 < 1 we get the representations mx i D 1 mc1 which imply in turn the inclusion and mx i 1 mc1 D 1; It yields therefore the relationships y D.1 mc1 / mx z WD mx i 1 mc1! i 2 : i 1 mc1! i C mc1! mc1 D.1 mc1 /z C mc1! mc1 2 and thus completes the proof of the proposition. Proposition 1.21 Let 1 be a convex subset of R n and let 2 be a convex subset of R p. en the Cartesian product 1 2 is a convex subset of R n R p.
25 1.2. CONVEX SETS 7 Proof. Fix a D.a 1 ; a 2 /; b D.b 1 ; b 2 / 2 1 2, and 2.0; 1/. en we have a 1 ; b and a 2 ; b e convexity of 1 and 2 gives us which implies therefore that a i C.1 /b i 2 i for i D 1; 2; a C.1 /b D a 1 C.1 /b 1 ; a 2 C.1 /b : us the Cartesian product 1 2 is convex. Let us continue with the definition of affine mappings. Definition 1.22 A mapping B W R n! R p is AFFINE if there exist a linear mapping A W R n! R p and an element b 2 R p such that B.x/ D A.x/ C b for all x 2 R n. Every linear mapping is affine. Moreover, B W R n! R p is affine if and only if B.x C.1 /y/ D B.x/ C.1 /B.y/ for all x; y 2 R n and 2 R : Now we show that set convexity is preserved under affine operations. Proposition 1.23 Let B W R n! R p be an affine mapping. Suppose that is a convex subset of R n and is a convex subset of R p. en B. / is a convex subset of R p and B 1./ is a convex subset of R n. Proof. Fix any a; b 2 B. / and 2.0; 1/. en a D B.x/ and b D B.y/ for x; y 2. Since is convex, we have x C.1 /y 2. en a C.1 /b D B.x/ C.1 /B.y/ D B.x C.1 /y/ 2 B. /; which justifies the convexity of the image B. /. Taking now any x; y 2 B 1./ and 2.0; 1/, we get B.x/; B.y/ 2. is gives us B.x/ C.1 /B.y/ D B x C.1 /y 2 by the convexity of. us we have x C.1 /y 2 B 1./, which verifies the convexity of the inverse image B 1./. Proposition 1.24 Let 1; 2 R n be convex and let 2 R. en both sets 1 C 2 and 1 are also convex in R n. Proof. It follows directly from the definitions.
26 8 1. CONVEX SETS AND FUNCTIONS Next we proceed with intersections of convex sets. Proposition 1.25 subset of R n. Let f g 2I be a collection of convex subsets of R n. en T 2I is also a convex Proof. Taking any a; b 2 T 2I, we get that a; b 2 for all 2 I. e convexity of each ensures that a C.1 /b 2 for any 2.0; 1/. us a C.1 /b 2 T 2I and the intersection T 2I is convex. Definition 1.26 Let be a subset of R n. e CONVEX HULL of is defined by co WD \ o nc ˇ C is convex and C : e next proposition follows directly from the definition and Proposition Figure 1.2: Nonconvex set and its convex hull. Proposition 1.27 e convex hull co is the smallest convex set containing. Proof. e convexity of the set co follows from Proposition On the other hand, for any convex set C that contains we clearly have co C, which verifies the conclusion. Proposition 1.28 For any subset of R n, its convex hull admits the representation n X m co D i a i ˇˇˇ mx o i D 1; i 0; a i 2 ; m 2 N :
27 1.2. CONVEX SETS 9 Proof. Denoting by C the right-hand side of the representation to prove, we obviously have C. Let us check that the set C is convex. Take any a; b 2 C and get px a WD ia i ; qx b WD ˇj b j ; j D1 where a i ; b j 2, i; ˇj 0 with P p i D P q j D1 ˇj D 1, and p; q 2 N. It follows easily that for every number 2.0; 1/, we have en the resulting equality a C.1 /b D px i C j D1 px ia i C qx.1 /ˇj b j : j D1 qx px.1 /ˇj D i C.1 / qx ˇj D 1 ensures that a C.1 /b 2 C, and thus co C by the definition of co. Fix now any a D P m ia i 2 C with a i 2 co for i D 1; : : : ; m. Since the set co is convex, we conclude j D1 by Proposition 1.20 that a 2 co and thus co D C. Proposition 1.29 e interior int and closure of a convex set R n are also convex. Proof. Fix a; b 2 int and 2.0; 1/. en find an open set V such that a 2 V and so a C.1 /b 2 V C.1 /b : Since V C.1 /b is open, we get a C.1 /b 2 int, and thus the set int is convex. To verify the convexity of, we fix a; b 2 and 2.0; 1/ and then find sequences fa k g and fb k g in converging to a and b, respectively. By the convexity of, the sequence fa k C.1 /b k g lies entirely in and converges to a C.1 /b. is ensures the inclusion a C.1 /b 2 and thus justifies the convexity of the closure. To proceed further, for any a; b 2 R n, define the interval Œa; b/ WD a C.1 /b ˇˇ 2.0; 1 : We can also define the intervals.a; b and.a; b/ in a similar way. Lemma 1.30 For a convex set R n with nonempty interior, take any a 2 int and b 2. en Œa; b/ int.
28 10 1. CONVEX SETS AND FUNCTIONS Proof. Since b 2, for any > 0, we have b 2 C IB. Take now 2.0; 1 and let x WD a C.1 /b. Choosing > 0 such that a C 2 IB gives us x C IB D a C.1 /b C IB a C.1 / C IB C IB D a C.1 / C.1 /IB C IB h a C 2 i IB C.1 / C.1 / : is shows that x 2 int and thus verifies the inclusion Œa; b/ int. Now we establish relationships between taking the interior and closure of convex sets. Proposition 1.31 Let R n be a convex set with nonempty interior. en we have: (i) int D and.ii/ int D int. Proof. (i) Obviously, int. For any b 2 and a 2 int, define the sequence fx k g by x k WD 1 k a C 1 1 b; k 2 N: k Lemma 1.30 ensures that x k 2 int. Since x k! b, we have b 2 int and thus verify (i). (ii) Since the inclusion int int is obvious, it remains to prove the opposite inclusion int int. To proceed, fix any b 2 int and a 2 int. If > 0 is sufficiently small, then c WD b C.b a/ 2, and hence b D 1 C a C 1 c 2.a; c/ int ; 1 C which verifies that int int and thus completes the proof. 1.3 CONVEX FUNCTIONS is section collects basic facts about general (extended-real-valued) convex functions including their analytic and geometric characterizations, important properties as well as their specifications for particular subclasses. We also define convex set-valued mappings and use them to study a remarkable class of optimal value functions employed in what follows. Definition 1.32 Let f W! R be an extended-real-valued function define on a convex set R n. en the function f is CONVEX on if f x C.1 /y f.x/ C.1 /f.y/ for all x; y 2 and 2.0; 1/: (1.1)
29 If the inequality in (1.1) is strict for x y, then f is STRICTLY CONVEX on. Given a function f W! R, the extension of f to R n is defined by 1.3. CONVEX FUNCTIONS 11 ef.x/ WD ( f.x/ if x 2 ; 1 otherwise: Obviously, if f is convex on, where is a convex set, then e f is convex on R n. Furthermore, if f W R n! R is a convex function, then it is also convex on every convex subset of R n. is allows to consider without loss of generality extended-real-valued convex functions on the whole space R n. Figure 1.3: Convex function and nonconvex function. Definition 1.33 e DOMAIN and EPIGRAPH of f W R n! R are defined, respectively, by dom f WD x 2 R n ˇˇ f.x/ < 1 and epi f WD.x; t/ 2 R n R ˇˇ x 2 R n ; t f.x/ D.x; t/ 2 R n R ˇˇ x 2 dom f; t f.x/ : Let us illustrate the convexity of functions by examples. Example 1.34 e following functions are convex: (i) f.x/ WD ha; xi C b for x 2 R n, where a 2 R n and b 2 R. (ii) g.x/ WD kxk for x 2 R n. (iii) h.x/ WD x 2 for x 2 R.
30 12 1. CONVEX SETS AND FUNCTIONS Indeed, the function f in (i) is convex since f x C.1 /y D ha; x C.1 /yi C b D ha; xi C.1 /ha; yi C b D.ha; xi C b/ C.1 /.ha; yi C b/ D f.x/ C.1 /f.y/ for all x; y 2 R n and 2.0; 1/: e function g in (ii) is convex since for x; y 2 R n and 2.0; 1/, we have g x C.1 /y D kx C.1 /yk kxk C.1 /kyk D g.x/ C.1 /g.y/; which follows from the triangle inequality and the fact that k uk D j j kuk whenever 2 R and u 2 R n. e convexity of the simplest quadratic function h in (iii) follows from a more general result for the quadratic function on R n given in the next example. Example 1.35 Let A be an n n symmetric matrix. It is called positive semidefinite if hau; ui 0 for all u 2 R n. Let us check that A is positive semidefinite if and only if the function f W R n! R defined by f.x/ WD 1 2 hax; xi; x 2 Rn ; is convex. Indeed, a direct calculation shows that for any x; y 2 R n and 2.0; 1/ we have f.x/ C.1 /f.y/ f x C.1 /y D 1.1 /ha.x y/; x yi: (1.2) 2 If the matrix A is positive semidefinite, then ha.x y/; x yi 0, so the function f is convex by (1.2). Conversely, assuming the convexity of f and using equality (1.2) for x D u and y D 0 verify that A is positive semidefinite. e following characterization of convexity is known as the Jensen inequality. eorem 1.36 A function f W R n! R is convex if and only if for any numbers i 0 as i D 1; : : : ; m with P m i D 1 and for any elements x i 2 R n, i D 1; : : : ; m, it holds that X m f i x i mx i f.x i /: (1.3) Proof. Since (1.3) immediately implies the convexity of f, we only need to prove that any convex function f satisfies the Jensen inequality (1.3). Arguing by induction and taking into account that for m D 1 inequality (1.3) holds trivially and for m D 2 inequality (1.3) holds by the definition of convexity, we suppose that it holds for an integer m WD k with k 2. Fix numbers i 0, i D 1; : : : ; k C 1, with P kc1 i D 1 and elements x i 2 R n, i D 1; : : : ; k C 1. We obviously have the relationship kx i D 1 kc1 :
31 1.3. CONVEX FUNCTIONS 13 If kc1 D 1, then i D 0 for all i D 1; : : : ; k and (1.3) obviously holds for m WD k C 1 in this case. Supposing now that 0 kc1 < 1, we get kx i 1 kc1 D 1 and by direct calculations based on convexity arrive at f kc1 X i x i D f h P k.1 kc1 / i ix i C kc1 x kc1 1 kc1.1 kc1 /f kx D.1 kc1 /f.1 kc1 / P k ix i 1 kc1 kx kc1 X D i f.x i /: C kc1 f.x kc1 / i x i C kc1 f.x kc1 / 1 kc1 i 1 kc1 f.x i / C kc1 f.x kc1 / is justifies inequality (1.3) and completes the proof of the theorem. e next theorem gives a geometric characterization of the function convexity via the convexity of the associated epigraphical set. eorem 1.37 A function f W R n! R is convex if and only if its epigraph epi f is a convex subset of the product space R n R. Proof. Assuming that f is convex, fix pairs.x 1 ; t 1 /;.x 2 ; t 2 / 2 epi f and a number 2.0; 1/. en we have f.x i / t i for i D 1; 2. us the convexity of f ensures that is implies therefore that f x 1 C.1 /x 2 f.x1 / C.1 /f.x 2 / t 1 C.1 /t 2 :.x 1 ; t 1 / C.1 /.x 2 ; t 2 / D.x 1 C.1 /x 2 ; t 1 C.1 /t 2 / 2 epi f; and thus the epigraph epi f is a convex subset of R n R. Conversely, suppose that the set epi f is convex and fix x 1 ; x 2 2 dom f and a number 2.0; 1/. en.x 1 ; f.x 1 //;.x 2 ; f.x 2 // 2 epi f. is tells us that x1 C.1 /x 2 ; f.x 1 / C.1 /f.x 2 / D x 1 ; f.x 1 / C.1 / x 2 ; f.x 2 / 2 epi f
32 14 1. CONVEX SETS AND FUNCTIONS and thus implies the inequality f x 1 C.1 /x 2 f.x1 / C.1 /f.x 2 /; which justifies the convexity of the function f. Figure 1.4: Epigraphs of convex function and nonconvex function. Now we show that convexity is preserved under some important operations. Proposition 1.38 Let f i W R n! R be convex functions for all i D 1; : : : ; m. en the following functions are convex as well: (i) e multiplication by scalars f for any > 0. (ii) e sum function P m f i. (iii) e maximum function max 1im f i. Proof. e convexity of f in (i) follows directly from the definition. It is sufficient to prove (ii) and (iii) for m D 2, since the general cases immediately follow by induction. (ii) Fix any x; y 2 R n and 2.0; 1/. en we have f1 C f 2 x C.1 /y D f1 x C.1 /y C f2 x C.1 /y which verifies the convexity of the sum function f 1 C f 2. f 1.x/ C.1 /f 1.y/ C f 2.x/ C.1 /f 2.y/ D.f 1 C f 2 /.x/ C.1 /.f 1 C f 2 /.y/; (iii) Denote g WD maxff 1 ; f 2 g and get for any x; y 2 R n and 2.0; 1/ that f i x C.1 /y fi.x/ C.1 /f i.y/ g.x/ C.1 /g.y/ for i D 1; 2. is directly implies that g x C.1 /y D max f 1 x C.1 /y ; f2 x C.1 /y g.x/ C.1 /g.y/; which shows that the maximum function g.x/ D maxff 1.x/; f 2.x/g is convex.
33 1.3. CONVEX FUNCTIONS 15 e next result concerns the preservation of convexity under function compositions. Proposition 1.39 Let f W R n! R be convex and let ' W R! R be nondecreasing and convex on a convex set containing the range of the function f. en the composition ' ı f is convex. Proof. Take any x 1 ; x 2 2 R n and 2.0; 1/. en we have by the convexity of f that f x 1 C.1 /x 2 f.x1 / C.1 /f.x 2 /: Since ' is nondecreasing and it is also convex, it follows that ' ı f x1 C.1 /x 2 D ' f x1 C.1 /x 2 ' f.x 1 / C.1 /f.x 2 / ' f.x 1 / C.1 /' f.x 2 / D.' ı f /.x 1 / C.1 /.' ı f /.x 2 /; which verifies the convexity of the composition ' ı f. Now we consider the composition of a convex function and an affine mapping. Proposition 1.40 Let B W R n! R p be an affine mapping and let f W R p! R be a convex function. en the composition f ı B is convex. Proof. Taking any x; y 2 R n and 2.0; 1/, we have f ı B x C.1 /y D f.b.x C.1 /y// D f B.x/ C.1 /B.y/ f B.x/ C.1 /f B.y/ D.f ı B/.x/ C.1 /.f ı B/.y/ and thus justify the convexity of the composition f ı B. e following simple consequence of Proposition 1.40 is useful in applications. Corollary 1.41 Let f W R n! R be a convex function. For any Nx; d 2 R n, the function 'W R! R defined by '.t/ WD f. Nx C td/ is convex as well. Conversely, if for every Nx; d 2 R n the function ' defined above is convex, then f is also convex. Proof. Since B.t/ D Nx C td is an affine mapping, the convexity of ' immediately follows from Proposition To prove the converse implication, take any x 1 ; x 2 2 R n, 2.0; 1/ and let Nx WD x 2, d WD x 1 x 2. Since '.t/ D f. Nx C td/ is convex, we have f x 1 C.1 /x 2 D f x2 C.x 1 x 2 / D './ D '.1/ C.1 /.0/ '.1/ C.1 /'.0/ D f.x 1 / C.1 /f.x 2 /; which verifies the convexity of the function f.
34 16 1. CONVEX SETS AND FUNCTIONS e next proposition is trivial while useful in what follows. Proposition 1.42 Let f W R n R p! R be convex. For. Nx; Ny/ 2 R n R p, the functions '.y/ WD f. Nx; y/ and.x/ WD f.x; Ny/ are also convex. Now we present an important extension of Proposition 1.38(iii). Proposition 1.43 Let f i W R n! R for i 2 I be a collection of convex functions with a nonempty index set I. en the supremum function f.x/ WD sup i2i f i.x/ is convex. Proof. Fix x 1 ; x 2 2 R n and 2.0; 1/. For every i 2 I, we have f i x1 C.1 /x 2 fi.x 1 / C.1 /f i.x 2 / f.x 1 / C.1 /f.x 2 /; which implies in turn that f x 1 C.1 /x 2 D sup f i x1 C.1 /x 2 f.x1 / C.1 /f.x 2 /: i2i is justifies the convexity of the supremum function. Our next intention is to characterize convexity of smooth functions of one variable. To proceed, we begin with the following lemma. Lemma 1.44 Given a convex function f W R! R, assume that its domain is an open interval I. For any a; b 2 I and a < x < b, we have the inequalities f.x/ f.a/ x a f.b/ f.a/ b a f.b/ f.x/ : b x Proof. Fix a; b; x as above and form the numbers t WD x a 2.0; 1/. en b a f.x/ D f a C.x a/ D f a C x a b a D f a C t.b a/ D f tb C.1 t/a : b a is gives us the inequalities f.x/ tf.b/ C.1 t/f.a/ and f.x/ f.a/ tf.b/ C.1 t/f.a/ f.a/ D t f.b/ f.a/ D x a f.b/ f.a/ ; b a which can be equivalently written as f.x/ f.a/ x a f.b/ f.a/ : b a
35 1.3. CONVEX FUNCTIONS 17 Similarly, we have the estimate f.x/ f.b/ tf.b/ C.1 t/f.a/ f.b/ D.t 1/ f.b/ f.a/ D x b f.b/ f.a/ ; b a which finally implies that f.b/ f.a/ b a and thus completes the proof of the lemma. f.b/ f.x/ b x eorem 1.45 Suppose that f W R! R is differentiable on its domain, which is an open interval I. en f is convex if and only if the derivative f 0 is nondecreasing on I. Proof. Suppose that f is convex and fix a < b with a; b 2 I. By Lemma 1.44, we have f.x/ f.a/ x a f.b/ f.a/ b a for any x 2.a; b/. is implies by the derivative definition that Similarly, we arrive at the estimate f 0.a/ f.b/ f.a/ b a f.b/ f.a/ : b a f 0.b/ and conclude that f 0.a/ f 0.b/, i.e., f 0 is a nondecreasing function. To prove the converse, suppose that f 0 is nondecreasing on I and fix x 1 < x 2 with x 1 ; x 2 2 I and t 2.0; 1/. en x 1 < x t < x 2 for x t WD tx 1 C.1 t/x 2 : By the mean value theorem, we find c 1 ; c 2 such that x 1 < c 1 < x t < c 2 < x 2 and f.x t / f.x 1 / D f 0.c 1 /.x t x 1 / D f 0.c 1 /.1 t/.x 2 x 1 /; f.x t / f.x 2 / D f 0.c 2 /.x t x 2 / D f 0.c 2 /t.x 1 x 2 /: is can be equivalently rewritten as tf.x t / tf.x 1 / D f 0.c 1 /t.1 t/.x 2 x 1 /;.1 t/f.x t /.1 t/f.x 2 / D f 0.c 2 /t.1 t/.x 1 x 2 /: Since f 0.c 1 / f 0.c 2 /, adding these equalities yields f.x t / tf.x 1 / C.1 t/f.x 2 /; which justifies the convexity of the function f.
36 18 1. CONVEX SETS AND FUNCTIONS Corollary 1.46 Let f W R! R be twice differentiable on its domain, which is an open interval I. en f is convex if and only if f 00.x/ 0 for all x 2 I. Proof. Since f 00.x/ 0 for all x 2 I if and only if the derivative function f 0 is nondecreasing on this interval. en the conclusion follows directly from eorem Example 1.47 Consider the function 8 < 1 if x > 0; f.x/ WD x : 1 otherwise: To verify its convexity, we get that f 00.x/ D 2 > 0 for all x belonging to the domain of f, which x3 is I D.0; 1/. us this function is convex on R by Corollary Next we define the notion of set-valued mappings (or multifunctions), which plays an important role in modern convex analysis, its extensions, and applications. Definition 1.48 We say that F is a SET-VALUED MAPPING between R n and R p and denote it by F W R n! R p if F.x/ is a subset of R p for every x 2 R n. e DOMAIN and GRAPH of F are defined, respectively, by dom F WD x 2 R n ˇˇ F.x/ ; and gph F WD.x; y/ 2 R n R p ˇˇ y 2 F.x/ : Any single-valued mapping F W R n! R p is a particular set-valued mapping where the set F.x/ is a singleton for every x 2 R n. It is essential in the following definition that the mapping F is set-valued. Definition 1.49 Let F W R n! R p and let ' W R n R p! R. e OPTIMAL VALUE or MARGINAL FUNCTION associated with F and ' is defined by.x/ WD inf '.x; y/ ˇˇ y 2 F.x/ ; x 2 R n : (1.4) roughout this section we assume that.x/ > 1 for every x 2 R n. Proposition 1.50 Assume that 'W R n R p! R is a convex function and that F W R n! R p is of convex graph. en the optimal value function in (1.4) is convex.
37 1.3. CONVEX FUNCTIONS 19 Figure 1.5: Graph of set-valued mapping. Proof. Take x 1 ; x 2 2 dom, 2.0; 1/. For any > 0, find y i 2 F.x i / such that It directly implies the inequalities '.x i ; y i / <.x i / C for i D 1; 2: '.x 1 ; y 1 / <.x 1 / C ;.1 /'.x 2 ; y 2 / <.1 /.x 2 / C.1 /: Summing up these inequalities and employing the convexity of ' yield ' x 1 C.1 /x 2 ; y 1 C.1 /y 2 '.x1 ; y 1 / C.1 /'.x 2 ; y 2 / <.x 1 / C.1 /.x 2 / C : Furthermore, the convexity of gph F gives us x1 C.1 /x 2 ; y 1 C.1 /y 2 D.x1 ; y 1 / C.1 /.x 2 ; y 2 / 2 gph F; and therefore y 1 C.1 /y 2 2 F.x 1 C.1 /x 2 /. is implies that x 1 C.1 /x 2 ' x1 C.1 /x 2 ; y 1 C.1 /y 2 <.x1 / C.1 /.x 2 / C : Letting finally! 0 ensures the convexity of the optimal value function. Using Proposition 1.50, we can verify convexity in many situations. For instance, given two convex functions f i W R n! R, i D 1; 2, let '.x; y/ WD f 1.x/ C y and F.x/ WD Œf 2.x/; 1/. en the function ' is convex and set gph F D epi f 2 is convex as well, and hence we justify the convexity of the sum.x/ D inf '.x; y/ D f 1.x/ C f 2.x/: y2f.x/
38 20 1. CONVEX SETS AND FUNCTIONS Another example concerns compositions. Let f W R p! R be convex and let B W R n! R p be affine. Define '.x; y/ WD f.y/ and F.x/ WD fb.x/g. Observe that ' is convex while F is of convex graph. us we have the convex composition.x/ D inf y2f.x/ '.x; y/ D f B.x/ ; x 2 R n : e examples presented above recover the results obtained previously by direct proofs. Now we establish via Proposition 1.50 the convexity of three new classes of functions. Proposition 1.51 Let ' W R p! R be convex and let B W R p! R n be affine. Consider the setvalued inverse image mapping B 1 W R n! R p, define f.x/ WD inf '.y/ ˇˇ y 2 B 1.x/ ; x 2 R n ; and suppose that f.x/ > 1 for all x 2 R n. en f is a convex function. Proof. Let '.x; y/ '.y/ and F.x/ WD B 1.x/. en the set gph F D.u; v/ 2 R n R p ˇˇ B.v/ D u is obviously convex. Since we have the representation f.x/ D inf '.y/; x 2 R n ; y2f.x/ the convexity of f follows directly from Proposition Proposition 1.52 For convex functions f 1 ; f 2 W R n! R, define the INFIMAL CONVOLUTION.f 1 f 2 /.x/ WD inf f 1.x 1 / C f 2.x 2 / ˇˇ x1 C x 2 D x and suppose that.f 1 f 2 /.x/ > 1 for all x 2 R n. en f 1 f 2 is also convex. Proof. Define ' W R n R n! R by '.x 1 ; x 2 / WD f 1.x 1 / C f 2.x 2 / and B W R n R n! R n by B.x 1 ; x 2 / WD x 1 C x 2. We have inf '.x 1 ; x 2 / ˇˇ.x1 ; x 2 / 2 B 1.x/ D.f 1 f 2 /.x/ for all x 2 R n ; which implies the convexity of.f 1 f 2 / by Proposition 1.51.
39 Definition RELATIVE INTERIORS OF CONVEX SETS 21 A function g W R p! R is called NONDECREASING COMPONENTWISE if xi y i for all i D 1; : : : ; p H) g.x 1 ; : : : ; x p / g.y 1 ; : : : ; y p / : Now we are ready to present the final consequence of Proposition 1.50 in this section that involves the composition. Proposition 1.54 Define h W R n! R p by h.x/ WD.f 1.x/; : : : ; f p.x//, where f i W R n! R for i D 1; : : : ; p are convex functions. Suppose that g W R p! R is convex and nondecreasing componentwise. en the composition g ı h W R n! R is a convex function. Proof. Let F W R n! R p be a set-valued mapping defined by en the graph of F is represented by F.x/ WD Œf 1.x/; 1/ Œf 2.x/; 1/ : : : Œf p.x/; 1/: gph F D.x; t 1 ; : : : ; t p / 2 R n R p ˇˇ ti f i.x/ : Since all f i are convex, the set gph F is convex as well. Define further ' W R n R p! R by '.x; y/ WD g.y/ and observe, since g is increasing componentwise, that inf '.x; y/ ˇˇ y 2 F.x/ D g f1.x/; : : : ; f p.x/ D.g ı h/.x/; which ensures the convexity of the composition g ı h by Proposition RELATIVE INTERIORS OF CONVEX SETS We begin this section with the definition and properties of affine sets. Given two elements a and b in R n, the line connecting them is Note that if a D b, then LŒa; b D fag. LŒa; b WD a C.1 /b ˇˇ 2 R : Definition 1.55 A subset of R n is AFFINE if for any a; b 2 we have LŒa; b. For instance, any point, line, and plane in R 3 are affine sets. e empty set and the whole space are always affine. It follows from the definition that the intersection of any collection of affine sets is affine. is leads us to the construction of the affine hull of a set. Definition 1.56 e AFFINE HULL of a set R n is aff WD \ C ˇˇ C is affine and C :
40 22 1. CONVEX SETS AND FUNCTIONS An element x in R n of the form mx x D i! i with mx i D 1; m 2 N; is called an affine combination of! 1 ; : : : ;! m. e proof of the next proposition is straightforward and thus is omitted. Proposition 1.57 e following assertions hold: (i) A set is affine if and only if contains all affine combinations of its elements. (ii) Let, 1, and 2 be affine subsets of R n. en the sum 1 C 2 and the scalar product for any 2 R are also affine subsets of R n. (iii) Let B W R n! R p be an affine mapping. If is an affine subset of R n and is an affine subset of R p, then the image B. / is an affine subset of R p and the inverse image B 1./ is an affine subset of R n. (iv) Given R n, its affine hull is the smallest affine set containing. We have n X m aff D i! i ˇˇˇ mx o i D 1;! i 2 ; m 2 N : (v) A set is a.linear/ subspace if and only if is an affine set containing the origin. Next we consider relationships between affine sets and (linear) subspaces. Lemma 1.58! 2. A nonempty subset of R n is affine if and only if! is a subspace of R n for any Proof. Suppose that a nonempty set R n is affine. It follows from Proposition 1.57(v) that! is a subspace for any! 2. Conversely, fix! 2 and suppose that! is a subspace denoted by L. en the set D! C L is obviously affine. e preceding lemma leads to the following notion. Definition 1.59 An affine set is PARALLEL to a subspace L if D! C L for some! 2. e next proposition justifies the form of the parallel subspace. Proposition 1.60 Let be a nonempty, affine subset of R n. en it is parallel to the unique subspace L of R n given by L D.
41 1.4. RELATIVE INTERIORS OF CONVEX SETS 23 Proof. Given a nonempty, affine set, fix! 2 and come up to the linear subspace L WD! parallel to. To justify the uniqueness of such L, take any! 1 ;! 2 2 and any subspaces L 1 ; L 2 R n such that D! 1 C L 1 D! 2 C L 2. en L 1 D! 2! 1 C L 2. Since 0 2 L 1, we have! 1! 2 2 L 2. is implies that! 2! 1 2 L 2 and thus L 1 D! 2! 1 C L 2 L 2. Similarly, we get L 2 L 1, which justifies that L 1 D L 2. It remains to verify the representation L D. Let D! C L with the unique subspace L and some! 2. en L D!. Fix any x D u! with u;! 2 and observe that! is a subspace parallel to. Hence! D L by the uniqueness of L proved above. is ensures that x 2! D L and thus L. e uniqueness of the parallel subspace shows that the next notion is well defined. Definition 1.61 e DIMENSION OF AN AFFINE SET ; R n is the dimension of the linear subspace parallel to. Furthermore, the DIMENSION OF A CONVEX SET ; R n is the dimension of its affine hull aff. To proceed further, we need yet another definition important in what follows. Definition 1.62 e elements v 0 ; : : : ; v m in R n, m 1, are AFFINELY INDEPENDENT if h X m i v i D 0; id0 mx id0 i i D 0 H) i D 0 for all i D 0; : : : ; m : It is easy to observe the following relationship with the linear independence. Proposition 1.63 e elements v 0 ; : : : ; v m in R n are affinely independent if and only if the elements v 1 v 0 ; : : : ; v m v 0 are linearly independent. Proof. Suppose that v 0 ; : : : ; v m are affinely independent and consider the system mx mx i.v i v 0 / D 0; i.e., 0 v 0 C i v i D 0; where 0 WD P m i. Since the elements v 0 ; : : : ; v m are affinely independent and P m id0 i D 0, we have that i D 0 for all i D 1; : : : ; m. us v 1 v 0 ; : : : ; v m v 0 are linearly independent. e proof of the converse statement is straightforward. Recall that the span of some set C, span C, is the linear subspace generated by C. Lemma 1.64 Let WD afffv 0 ; : : : ; v m g, where v i 2 R n for all i D 0; : : : ; m. en the span of the set fv 1 v 0 ; : : : ; v m v 0 g is the subspace parallel to.
42 24 1. CONVEX SETS AND FUNCTIONS Proof. Denote by L the subspace parallel to. en v 0 D L and therefore v i v 0 2 L for all i D 1; : : : ; m. is gives span v i v 0 ˇˇ i D 1; : : : ; m L: To prove the converse inclusion, fix any v 2 L and get v C v 0 2. us we have mx v C v 0 D i v i ; id0 mx i D 1: id0 is implies the relationship mx v D i.v i v 0 / 2 span v i v 0 ˇˇ i D 1; : : : ; m ; which justifies the converse inclusion and hence completes the proof. e proof of the next proposition is rather straightforward. Proposition 1.65 e elements v 0 ; : : : ; v m are affinely independent in R n if and only if its affine hull WD afffv 0 ; : : : ; v m g is m-dimensional. Proof. Suppose that v 0 ; : : : ; v m are affinely independent. en Lemma 1.64 tells us that the subspace L WD spanfv i v 0 j i D 1; : : : ; mg is parallel to. e linear independence of v 1 v 0 ; : : : ; v m v 0 by Proposition 1.63 means that the subspace L is m-dimensional and so is. e proof of the converse statement is also straightforward. Affinely independent systems lead us to the construction of simplices. Definition 1.66 Let v 0 ; : : : ; v m be affinely independent in R n. en the set m WD co v i ˇˇ i D 0; : : : ; m is called an m-simplex in R n with the vertices v i, i D 0; : : : ; m. An important role of simplex vertices is revealed by the following proposition. Proposition 1.67 Consider an m-simplex m with vertices v i for i D 0; : : : ; m. For every v 2 m, there is a unique element. 0 ; : : : ; m / 2 R mc1 C such that mx v D i v i ; id0 mx i D 1: id0
Applications of Affine and Weyl Geometry
Applications of Affine and Weyl Geometry Synthesis Lectures on Mathematics and Statistics Editor Steven G. Krantz, Washington University, St. Louis Applications of Affine and Weyl Geometry Eduardo García-Río,
More informationJordan Canonical Form. Theory and Practice
Jordan Canonical Form Theory and Practice Synthesis Lectures on Mathematics and Statistics Editor Steven G. Krantz, Washington University, St. Louis Jordan Canonical Form: Theory and Practice Steven H.
More informationHelly's Theorem and its Equivalences via Convex Analysis
Portland State University PDXScholar University Honors Theses University Honors College 2014 Helly's Theorem and its Equivalences via Convex Analysis Adam Robinson Portland State University Let us know
More informationGEOMETRIC APPROACH TO CONVEX SUBDIFFERENTIAL CALCULUS October 10, Dedicated to Franco Giannessi and Diethard Pallaschke with great respect
GEOMETRIC APPROACH TO CONVEX SUBDIFFERENTIAL CALCULUS October 10, 2018 BORIS S. MORDUKHOVICH 1 and NGUYEN MAU NAM 2 Dedicated to Franco Giannessi and Diethard Pallaschke with great respect Abstract. In
More informationOptimality Conditions for Nonsmooth Convex Optimization
Optimality Conditions for Nonsmooth Convex Optimization Sangkyun Lee Oct 22, 2014 Let us consider a convex function f : R n R, where R is the extended real field, R := R {, + }, which is proper (f never
More informationDedicated to Michel Théra in honor of his 70th birthday
VARIATIONAL GEOMETRIC APPROACH TO GENERALIZED DIFFERENTIAL AND CONJUGATE CALCULI IN CONVEX ANALYSIS B. S. MORDUKHOVICH 1, N. M. NAM 2, R. B. RECTOR 3 and T. TRAN 4. Dedicated to Michel Théra in honor of
More informationOptimization and Optimal Control in Banach Spaces
Optimization and Optimal Control in Banach Spaces Bernhard Schmitzer October 19, 2017 1 Convex non-smooth optimization with proximal operators Remark 1.1 (Motivation). Convex optimization: easier to solve,
More informationChapter 2 Convex Analysis
Chapter 2 Convex Analysis The theory of nonsmooth analysis is based on convex analysis. Thus, we start this chapter by giving basic concepts and results of convexity (for further readings see also [202,
More informationExtreme Abridgment of Boyd and Vandenberghe s Convex Optimization
Extreme Abridgment of Boyd and Vandenberghe s Convex Optimization Compiled by David Rosenberg Abstract Boyd and Vandenberghe s Convex Optimization book is very well-written and a pleasure to read. The
More informationMAT 570 REAL ANALYSIS LECTURE NOTES. Contents. 1. Sets Functions Countability Axiom of choice Equivalence relations 9
MAT 570 REAL ANALYSIS LECTURE NOTES PROFESSOR: JOHN QUIGG SEMESTER: FALL 204 Contents. Sets 2 2. Functions 5 3. Countability 7 4. Axiom of choice 8 5. Equivalence relations 9 6. Real numbers 9 7. Extended
More informationSubdifferential representation of convex functions: refinements and applications
Subdifferential representation of convex functions: refinements and applications Joël Benoist & Aris Daniilidis Abstract Every lower semicontinuous convex function can be represented through its subdifferential
More informationMath 341: Convex Geometry. Xi Chen
Math 341: Convex Geometry Xi Chen 479 Central Academic Building, University of Alberta, Edmonton, Alberta T6G 2G1, CANADA E-mail address: xichen@math.ualberta.ca CHAPTER 1 Basics 1. Euclidean Geometry
More informationPOLARS AND DUAL CONES
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.
More informationA SET OF LECTURE NOTES ON CONVEX OPTIMIZATION WITH SOME APPLICATIONS TO PROBABILITY THEORY INCOMPLETE DRAFT. MAY 06
A SET OF LECTURE NOTES ON CONVEX OPTIMIZATION WITH SOME APPLICATIONS TO PROBABILITY THEORY INCOMPLETE DRAFT. MAY 06 CHRISTIAN LÉONARD Contents Preliminaries 1 1. Convexity without topology 1 2. Convexity
More informationInequality Constraints
Chapter 2 Inequality Constraints 2.1 Optimality Conditions Early in multivariate calculus we learn the significance of differentiability in finding minimizers. In this section we begin our study of the
More informationTopology. Xiaolong Han. Department of Mathematics, California State University, Northridge, CA 91330, USA address:
Topology Xiaolong Han Department of Mathematics, California State University, Northridge, CA 91330, USA E-mail address: Xiaolong.Han@csun.edu Remark. You are entitled to a reward of 1 point toward a homework
More informationChapter 1 Preliminaries
Chapter 1 Preliminaries 1.1 Conventions and Notations Throughout the book we use the following notations for standard sets of numbers: N the set {1, 2,...} of natural numbers Z the set of integers Q the
More informationConvex Functions. Pontus Giselsson
Convex Functions Pontus Giselsson 1 Today s lecture lower semicontinuity, closure, convex hull convexity preserving operations precomposition with affine mapping infimal convolution image function supremum
More informationON GENERALIZED-CONVEX CONSTRAINED MULTI-OBJECTIVE OPTIMIZATION
ON GENERALIZED-CONVEX CONSTRAINED MULTI-OBJECTIVE OPTIMIZATION CHRISTIAN GÜNTHER AND CHRISTIANE TAMMER Abstract. In this paper, we consider multi-objective optimization problems involving not necessarily
More informationMetric Spaces and Topology
Chapter 2 Metric Spaces and Topology From an engineering perspective, the most important way to construct a topology on a set is to define the topology in terms of a metric on the set. This approach underlies
More informationOn the use of semi-closed sets and functions in convex analysis
Open Math. 2015; 13: 1 5 Open Mathematics Open Access Research Article Constantin Zălinescu* On the use of semi-closed sets and functions in convex analysis Abstract: The main aim of this short note is
More informationThe proximal mapping
The proximal mapping http://bicmr.pku.edu.cn/~wenzw/opt-2016-fall.html Acknowledgement: this slides is based on Prof. Lieven Vandenberghes lecture notes Outline 2/37 1 closed function 2 Conjugate function
More informationOn Semicontinuity of Convex-valued Multifunctions and Cesari s Property (Q)
On Semicontinuity of Convex-valued Multifunctions and Cesari s Property (Q) Andreas Löhne May 2, 2005 (last update: November 22, 2005) Abstract We investigate two types of semicontinuity for set-valued
More informationIntroduction to Mathematical Analysis I. Second Edition. Beatriz Lafferriere Gerardo Lafferriere Nguyen Mau Nam
Introduction to Mathematical Analysis I Second Edition Beatriz Lafferriere Gerardo Lafferriere Nguyen Mau Nam Introduction to Mathematical Analysis I Second Edition Beatriz Lafferriere Gerardo Lafferriere
More informationLecture Notes in Advanced Calculus 1 (80315) Raz Kupferman Institute of Mathematics The Hebrew University
Lecture Notes in Advanced Calculus 1 (80315) Raz Kupferman Institute of Mathematics The Hebrew University February 7, 2007 2 Contents 1 Metric Spaces 1 1.1 Basic definitions...........................
More informationConstraint qualifications for convex inequality systems with applications in constrained optimization
Constraint qualifications for convex inequality systems with applications in constrained optimization Chong Li, K. F. Ng and T. K. Pong Abstract. For an inequality system defined by an infinite family
More informationWeek 3: Faces of convex sets
Week 3: Faces of convex sets Conic Optimisation MATH515 Semester 018 Vera Roshchina School of Mathematics and Statistics, UNSW August 9, 018 Contents 1. Faces of convex sets 1. Minkowski theorem 3 3. Minimal
More information1 Topology Definition of a topology Basis (Base) of a topology The subspace topology & the product topology on X Y 3
Index Page 1 Topology 2 1.1 Definition of a topology 2 1.2 Basis (Base) of a topology 2 1.3 The subspace topology & the product topology on X Y 3 1.4 Basic topology concepts: limit points, closed sets,
More informationTopological properties
CHAPTER 4 Topological properties 1. Connectedness Definitions and examples Basic properties Connected components Connected versus path connected, again 2. Compactness Definition and first examples Topological
More informationIntroduction to Convex and Quasiconvex Analysis
Introduction to Convex and Quasiconvex Analysis J.B.G.Frenk Econometric Institute, Erasmus University, Rotterdam G.Kassay Faculty of Mathematics, Babes Bolyai University, Cluj August 27, 2001 Abstract
More informationHandout 2: Elements of Convex Analysis
ENGG 5501: Foundations of Optimization 2018 19 First Term Handout 2: Elements of Convex Analysis Instructor: Anthony Man Cho So September 10, 2018 As briefly mentioned in Handout 1, the notion of convexity
More informationConvex Analysis and Economic Theory AY Elementary properties of convex functions
Division of the Humanities and Social Sciences Ec 181 KC Border Convex Analysis and Economic Theory AY 2018 2019 Topic 6: Convex functions I 6.1 Elementary properties of convex functions We may occasionally
More informationSet, functions and Euclidean space. Seungjin Han
Set, functions and Euclidean space Seungjin Han September, 2018 1 Some Basics LOGIC A is necessary for B : If B holds, then A holds. B A A B is the contraposition of B A. A is sufficient for B: If A holds,
More informationOptimality, identifiability, and sensitivity
Noname manuscript No. (will be inserted by the editor) Optimality, identifiability, and sensitivity D. Drusvyatskiy A. S. Lewis Received: date / Accepted: date Abstract Around a solution of an optimization
More informationIntroduction to Real Analysis Alternative Chapter 1
Christopher Heil Introduction to Real Analysis Alternative Chapter 1 A Primer on Norms and Banach Spaces Last Updated: March 10, 2018 c 2018 by Christopher Heil Chapter 1 A Primer on Norms and Banach Spaces
More informationMaths 212: Homework Solutions
Maths 212: Homework Solutions 1. The definition of A ensures that x π for all x A, so π is an upper bound of A. To show it is the least upper bound, suppose x < π and consider two cases. If x < 1, then
More information4. Convex Sets and (Quasi-)Concave Functions
4. Convex Sets and (Quasi-)Concave Functions Daisuke Oyama Mathematics II April 17, 2017 Convex Sets Definition 4.1 A R N is convex if (1 α)x + αx A whenever x, x A and α [0, 1]. A R N is strictly convex
More informationLegendre-Fenchel transforms in a nutshell
1 2 3 Legendre-Fenchel transforms in a nutshell Hugo Touchette School of Mathematical Sciences, Queen Mary, University of London, London E1 4NS, UK Started: July 11, 2005; last compiled: October 16, 2014
More informationg 2 (x) (1/3)M 1 = (1/3)(2/3)M.
COMPACTNESS If C R n is closed and bounded, then by B-W it is sequentially compact: any sequence of points in C has a subsequence converging to a point in C Conversely, any sequentially compact C R n is
More informationCourse 212: Academic Year Section 1: Metric Spaces
Course 212: Academic Year 1991-2 Section 1: Metric Spaces D. R. Wilkins Contents 1 Metric Spaces 3 1.1 Distance Functions and Metric Spaces............. 3 1.2 Convergence and Continuity in Metric Spaces.........
More informationLECTURE 25: REVIEW/EPILOGUE LECTURE OUTLINE
LECTURE 25: REVIEW/EPILOGUE LECTURE OUTLINE CONVEX ANALYSIS AND DUALITY Basic concepts of convex analysis Basic concepts of convex optimization Geometric duality framework - MC/MC Constrained optimization
More informationSome Properties of the Augmented Lagrangian in Cone Constrained Optimization
MATHEMATICS OF OPERATIONS RESEARCH Vol. 29, No. 3, August 2004, pp. 479 491 issn 0364-765X eissn 1526-5471 04 2903 0479 informs doi 10.1287/moor.1040.0103 2004 INFORMS Some Properties of the Augmented
More informationOptimization Theory. A Concise Introduction. Jiongmin Yong
October 11, 017 16:5 ws-book9x6 Book Title Optimization Theory 017-08-Lecture Notes page 1 1 Optimization Theory A Concise Introduction Jiongmin Yong Optimization Theory 017-08-Lecture Notes page Optimization
More informationIntegral Jensen inequality
Integral Jensen inequality Let us consider a convex set R d, and a convex function f : (, + ]. For any x,..., x n and λ,..., λ n with n λ i =, we have () f( n λ ix i ) n λ if(x i ). For a R d, let δ a
More informationGeneralized Differential Calculus and Applications to Optimization
Portland State University PDXScholar Dissertations and Theses Dissertations and Theses Spring 6-1-2017 Generalized Differential Calculus and Applications to Optimization R. Blake Rector Portland State
More informationAnalysis-3 lecture schemes
Analysis-3 lecture schemes (with Homeworks) 1 Csörgő István November, 2015 1 A jegyzet az ELTE Informatikai Kar 2015. évi Jegyzetpályázatának támogatásával készült Contents 1. Lesson 1 4 1.1. The Space
More informationSome Background Material
Chapter 1 Some Background Material In the first chapter, we present a quick review of elementary - but important - material as a way of dipping our toes in the water. This chapter also introduces important
More informationPARTIAL SECOND-ORDER SUBDIFFERENTIALS IN VARIATIONAL ANALYSIS AND OPTIMIZATION BORIS S. MORDUKHOVICH 1, NGUYEN MAU NAM 2 and NGUYEN THI YEN NHI 3
PARTIAL SECOND-ORDER SUBDIFFERENTIALS IN VARIATIONAL ANALYSIS AND OPTIMIZATION BORIS S. MORDUKHOVICH 1, NGUYEN MAU NAM 2 and NGUYEN THI YEN NHI 3 Abstract. This paper presents a systematic study of partial
More informationConvex Analysis and Economic Theory Winter 2018
Division of the Humanities and Social Sciences Ec 181 KC Border Convex Analysis and Economic Theory Winter 2018 Supplement A: Mathematical background A.1 Extended real numbers The extended real number
More informationOn the convexity of piecewise-defined functions
On the convexity of piecewise-defined functions arxiv:1408.3771v1 [math.ca] 16 Aug 2014 Heinz H. Bauschke, Yves Lucet, and Hung M. Phan August 16, 2014 Abstract Functions that are piecewise defined are
More informationGEORGIA INSTITUTE OF TECHNOLOGY H. MILTON STEWART SCHOOL OF INDUSTRIAL AND SYSTEMS ENGINEERING LECTURE NOTES OPTIMIZATION III
GEORGIA INSTITUTE OF TECHNOLOGY H. MILTON STEWART SCHOOL OF INDUSTRIAL AND SYSTEMS ENGINEERING LECTURE NOTES OPTIMIZATION III CONVEX ANALYSIS NONLINEAR PROGRAMMING THEORY NONLINEAR PROGRAMMING ALGORITHMS
More informationTranslative Sets and Functions and their Applications to Risk Measure Theory and Nonlinear Separation
Translative Sets and Functions and their Applications to Risk Measure Theory and Nonlinear Separation Andreas H. Hamel Abstract Recently defined concepts such as nonlinear separation functionals due to
More informationOptimality Conditions for Constrained Optimization
72 CHAPTER 7 Optimality Conditions for Constrained Optimization 1. First Order Conditions In this section we consider first order optimality conditions for the constrained problem P : minimize f 0 (x)
More informationConvex Functions and Optimization
Chapter 5 Convex Functions and Optimization 5.1 Convex Functions Our next topic is that of convex functions. Again, we will concentrate on the context of a map f : R n R although the situation can be generalized
More informationCHAPTER 2: CONVEX SETS AND CONCAVE FUNCTIONS. W. Erwin Diewert January 31, 2008.
1 ECONOMICS 594: LECTURE NOTES CHAPTER 2: CONVEX SETS AND CONCAVE FUNCTIONS W. Erwin Diewert January 31, 2008. 1. Introduction Many economic problems have the following structure: (i) a linear function
More informationMathematical Analysis Outline. William G. Faris
Mathematical Analysis Outline William G. Faris January 8, 2007 2 Chapter 1 Metric spaces and continuous maps 1.1 Metric spaces A metric space is a set X together with a real distance function (x, x ) d(x,
More informationLecture 1: Background on Convex Analysis
Lecture 1: Background on Convex Analysis John Duchi PCMI 2016 Outline I Convex sets 1.1 Definitions and examples 2.2 Basic properties 3.3 Projections onto convex sets 4.4 Separating and supporting hyperplanes
More informationChapter 2: Preliminaries and elements of convex analysis
Chapter 2: Preliminaries and elements of convex analysis Edoardo Amaldi DEIB Politecnico di Milano edoardo.amaldi@polimi.it Website: http://home.deib.polimi.it/amaldi/opt-14-15.shtml Academic year 2014-15
More informationChapter 2 Metric Spaces
Chapter 2 Metric Spaces The purpose of this chapter is to present a summary of some basic properties of metric and topological spaces that play an important role in the main body of the book. 2.1 Metrics
More informationAppendix B Convex analysis
This version: 28/02/2014 Appendix B Convex analysis In this appendix we review a few basic notions of convexity and related notions that will be important for us at various times. B.1 The Hausdorff distance
More informationConvex Analysis Background
Convex Analysis Background John C. Duchi Stanford University Park City Mathematics Institute 206 Abstract In this set of notes, we will outline several standard facts from convex analysis, the study of
More information2 Sequences, Continuity, and Limits
2 Sequences, Continuity, and Limits In this chapter, we introduce the fundamental notions of continuity and limit of a real-valued function of two variables. As in ACICARA, the definitions as well as proofs
More informationConvex Sets. Prof. Dan A. Simovici UMB
Convex Sets Prof. Dan A. Simovici UMB 1 / 57 Outline 1 Closures, Interiors, Borders of Sets in R n 2 Segments and Convex Sets 3 Properties of the Class of Convex Sets 4 Closure and Interior Points of Convex
More informationConvex Analysis and Optimization Chapter 2 Solutions
Convex Analysis and Optimization Chapter 2 Solutions Dimitri P. Bertsekas with Angelia Nedić and Asuman E. Ozdaglar Massachusetts Institute of Technology Athena Scientific, Belmont, Massachusetts http://www.athenasc.com
More informationEpiconvergence and ε-subgradients of Convex Functions
Journal of Convex Analysis Volume 1 (1994), No.1, 87 100 Epiconvergence and ε-subgradients of Convex Functions Andrei Verona Department of Mathematics, California State University Los Angeles, CA 90032,
More informationSubgradients. subgradients and quasigradients. subgradient calculus. optimality conditions via subgradients. directional derivatives
Subgradients subgradients and quasigradients subgradient calculus optimality conditions via subgradients directional derivatives Prof. S. Boyd, EE392o, Stanford University Basic inequality recall basic
More informationDocument downloaded from: This paper must be cited as:
Document downloaded from: http://hdl.handle.net/10251/50602 This paper must be cited as: Pedraza Aguilera, T.; Rodríguez López, J.; Romaguera Bonilla, S. (2014). Convergence of fuzzy sets with respect
More informationBASICS OF CONVEX ANALYSIS
BASICS OF CONVEX ANALYSIS MARKUS GRASMAIR 1. Main Definitions We start with providing the central definitions of convex functions and convex sets. Definition 1. A function f : R n R + } is called convex,
More informationB553 Lecture 3: Multivariate Calculus and Linear Algebra Review
B553 Lecture 3: Multivariate Calculus and Linear Algebra Review Kris Hauser December 30, 2011 We now move from the univariate setting to the multivariate setting, where we will spend the rest of the class.
More informationSemicontinuous functions and convexity
Semicontinuous functions and convexity Jordan Bell jordan.bell@gmail.com Department of Mathematics, University of Toronto April 3, 2014 1 Lattices If (A, ) is a partially ordered set and S is a subset
More informationConvex analysis and profit/cost/support functions
Division of the Humanities and Social Sciences Convex analysis and profit/cost/support functions KC Border October 2004 Revised January 2009 Let A be a subset of R m Convex analysts may give one of two
More informationVariational Analysis and Tame Optimization
Variational Analysis and Tame Optimization Aris Daniilidis http://mat.uab.cat/~arisd Universitat Autònoma de Barcelona April 15 17, 2010 PLAN OF THE TALK Nonsmooth analysis Genericity of pathological situations
More informationDivision of the Humanities and Social Sciences. Supergradients. KC Border Fall 2001 v ::15.45
Division of the Humanities and Social Sciences Supergradients KC Border Fall 2001 1 The supergradient of a concave function There is a useful way to characterize the concavity of differentiable functions.
More informationExtreme points of compact convex sets
Extreme points of compact convex sets In this chapter, we are going to show that compact convex sets are determined by a proper subset, the set of its extreme points. Let us start with the main definition.
More informationContents Real Vector Spaces Linear Equations and Linear Inequalities Polyhedra Linear Programs and the Simplex Method Lagrangian Duality
Contents Introduction v Chapter 1. Real Vector Spaces 1 1.1. Linear and Affine Spaces 1 1.2. Maps and Matrices 4 1.3. Inner Products and Norms 7 1.4. Continuous and Differentiable Functions 11 Chapter
More informationIntroduction to Convex Analysis Microeconomics II - Tutoring Class
Introduction to Convex Analysis Microeconomics II - Tutoring Class Professor: V. Filipe Martins-da-Rocha TA: Cinthia Konichi April 2010 1 Basic Concepts and Results This is a first glance on basic convex
More informationReal Analysis Math 131AH Rudin, Chapter #1. Dominique Abdi
Real Analysis Math 3AH Rudin, Chapter # Dominique Abdi.. If r is rational (r 0) and x is irrational, prove that r + x and rx are irrational. Solution. Assume the contrary, that r+x and rx are rational.
More informationMathematics 530. Practice Problems. n + 1 }
Department of Mathematical Sciences University of Delaware Prof. T. Angell October 19, 2015 Mathematics 530 Practice Problems 1. Recall that an indifference relation on a partially ordered set is defined
More informationAnalysis Finite and Infinite Sets The Real Numbers The Cantor Set
Analysis Finite and Infinite Sets Definition. An initial segment is {n N n n 0 }. Definition. A finite set can be put into one-to-one correspondence with an initial segment. The empty set is also considered
More informationMA651 Topology. Lecture 10. Metric Spaces.
MA65 Topology. Lecture 0. Metric Spaces. This text is based on the following books: Topology by James Dugundgji Fundamental concepts of topology by Peter O Neil Linear Algebra and Analysis by Marc Zamansky
More informationConvex Optimization Theory. Chapter 5 Exercises and Solutions: Extended Version
Convex Optimization Theory Chapter 5 Exercises and Solutions: Extended Version Dimitri P. Bertsekas Massachusetts Institute of Technology Athena Scientific, Belmont, Massachusetts http://www.athenasc.com
More informationLectures on Geometry
January 4, 2001 Lectures on Geometry Christer O. Kiselman Contents: 1. Introduction 2. Closure operators and Galois correspondences 3. Convex sets and functions 4. Infimal convolution 5. Convex duality:
More informationSMSTC (2017/18) Geometry and Topology 2.
SMSTC (2017/18) Geometry and Topology 2 Lecture 1: Differentiable Functions and Manifolds in R n Lecturer: Diletta Martinelli (Notes by Bernd Schroers) a wwwsmstcacuk 11 General remarks In this lecture
More informationIn English, this means that if we travel on a straight line between any two points in C, then we never leave C.
Convex sets In this section, we will be introduced to some of the mathematical fundamentals of convex sets. In order to motivate some of the definitions, we will look at the closest point problem from
More informationConvex Analysis and Economic Theory Fall 2018 Winter 2019
Division of the Humanities and Social Sciences Ec 181 KC Border Convex Analysis and Economic Theory Fall 2018 Winter 2019 Topic 18: Differentiability 18.1 Differentiable functions In this section I want
More informationZERO DUALITY GAP FOR CONVEX PROGRAMS: A GENERAL RESULT
ZERO DUALITY GAP FOR CONVEX PROGRAMS: A GENERAL RESULT EMIL ERNST AND MICHEL VOLLE Abstract. This article addresses a general criterion providing a zero duality gap for convex programs in the setting of
More informationMath 117: Topology of the Real Numbers
Math 117: Topology of the Real Numbers John Douglas Moore November 10, 2008 The goal of these notes is to highlight the most important topics presented in Chapter 3 of the text [1] and to provide a few
More informationMORE EXERCISES FOR SECTIONS II.1 AND II.2. There are drawings on the next two pages to accompany the starred ( ) exercises.
Math 133 Winter 2013 MORE EXERCISES FOR SECTIONS II.1 AND II.2 There are drawings on the next two pages to accompany the starred ( ) exercises. B1. Let L be a line in R 3, and let x be a point which does
More informationIntroduction to Topology
Introduction to Topology Randall R. Holmes Auburn University Typeset by AMS-TEX Chapter 1. Metric Spaces 1. Definition and Examples. As the course progresses we will need to review some basic notions about
More informationEE 546, Univ of Washington, Spring Proximal mapping. introduction. review of conjugate functions. proximal mapping. Proximal mapping 6 1
EE 546, Univ of Washington, Spring 2012 6. Proximal mapping introduction review of conjugate functions proximal mapping Proximal mapping 6 1 Proximal mapping the proximal mapping (prox-operator) of a convex
More informationMath 5052 Measure Theory and Functional Analysis II Homework Assignment 7
Math 5052 Measure Theory and Functional Analysis II Homework Assignment 7 Prof. Wickerhauser Due Friday, February 5th, 2016 Please do Exercises 3, 6, 14, 16*, 17, 18, 21*, 23*, 24, 27*. Exercises marked
More informationTopology, Math 581, Fall 2017 last updated: November 24, Topology 1, Math 581, Fall 2017: Notes and homework Krzysztof Chris Ciesielski
Topology, Math 581, Fall 2017 last updated: November 24, 2017 1 Topology 1, Math 581, Fall 2017: Notes and homework Krzysztof Chris Ciesielski Class of August 17: Course and syllabus overview. Topology
More informationAppendix PRELIMINARIES 1. THEOREMS OF ALTERNATIVES FOR SYSTEMS OF LINEAR CONSTRAINTS
Appendix PRELIMINARIES 1. THEOREMS OF ALTERNATIVES FOR SYSTEMS OF LINEAR CONSTRAINTS Here we consider systems of linear constraints, consisting of equations or inequalities or both. A feasible solution
More informationLecture Notes on Metric Spaces
Lecture Notes on Metric Spaces Math 117: Summer 2007 John Douglas Moore Our goal of these notes is to explain a few facts regarding metric spaces not included in the first few chapters of the text [1],
More informationMath 413/513 Chapter 6 (from Friedberg, Insel, & Spence)
Math 413/513 Chapter 6 (from Friedberg, Insel, & Spence) David Glickenstein December 7, 2015 1 Inner product spaces In this chapter, we will only consider the elds R and C. De nition 1 Let V be a vector
More informationL p Spaces and Convexity
L p Spaces and Convexity These notes largely follow the treatments in Royden, Real Analysis, and Rudin, Real & Complex Analysis. 1. Convex functions Let I R be an interval. For I open, we say a function
More informationREVIEW OF ESSENTIAL MATH 346 TOPICS
REVIEW OF ESSENTIAL MATH 346 TOPICS 1. AXIOMATIC STRUCTURE OF R Doğan Çömez The real number system is a complete ordered field, i.e., it is a set R which is endowed with addition and multiplication operations
More informationPart III. 10 Topological Space Basics. Topological Spaces
Part III 10 Topological Space Basics Topological Spaces Using the metric space results above as motivation we will axiomatize the notion of being an open set to more general settings. Definition 10.1.
More informationConvex Optimization Notes
Convex Optimization Notes Jonathan Siegel January 2017 1 Convex Analysis This section is devoted to the study of convex functions f : B R {+ } and convex sets U B, for B a Banach space. The case of B =
More informationTHE INVERSE FUNCTION THEOREM FOR LIPSCHITZ MAPS
THE INVERSE FUNCTION THEOREM FOR LIPSCHITZ MAPS RALPH HOWARD DEPARTMENT OF MATHEMATICS UNIVERSITY OF SOUTH CAROLINA COLUMBIA, S.C. 29208, USA HOWARD@MATH.SC.EDU Abstract. This is an edited version of a
More information