Introduction to Optimization Techniques

Similar documents
Real Numbers R ) - LUB(B) may or may not belong to B. (Ex; B= { y: y = 1 x, - Note that A B LUB( A) LUB( B)

Introduction to Optimization Techniques. How to Solve Equations

Chapter 3 Inner Product Spaces. Hilbert Spaces

Definition 4.2. (a) A sequence {x n } in a Banach space X is a basis for X if. unique scalars a n (x) such that x = n. a n (x) x n. (4.

Lecture Notes for Analysis Class

PAPER : IIT-JAM 2010

HILBERT SPACE GEOMETRY

FUNDAMENTALS OF REAL ANALYSIS by

A) is empty. B) is a finite set. C) can be a countably infinite set. D) can be an uncountable set.

Abstract Vector Spaces. Abstract Vector Spaces

Math Solutions to homework 6

Lecture 10: Bounded Linear Operators and Orthogonality in Hilbert Spaces

Properties of Fuzzy Length on Fuzzy Set

Solutions to home assignments (sketches)

} is said to be a Cauchy sequence provided the following condition is true.

We are mainly going to be concerned with power series in x, such as. (x)} converges - that is, lims N n

CHAPTER 5. Theory and Solution Using Matrix Techniques

MAS111 Convergence and Continuity

1 Introduction. 1.1 Notation and Terminology

Archimedes - numbers for counting, otherwise lengths, areas, etc. Kepler - geometry for planetary motion

Riesz-Fischer Sequences and Lower Frame Bounds

, then cv V. Differential Equations Elements of Lineaer Algebra Name: Consider the differential equation. and y2 cos( kx)

Matrix Algebra from a Statistician s Perspective BIOS 524/ Scalar multiple: ka

Apply change-of-basis formula to rewrite x as a linear combination of eigenvectors v j.

LECTURE 8: ORTHOGONALITY (CHAPTER 5 IN THE BOOK)

lim za n n = z lim a n n.

Introduction to Functional Analysis

Functional Analysis I

A NOTE ON SPECTRAL CONTINUITY. In Ho Jeon and In Hyoun Kim

Brief Review of Functions of Several Variables

Homework 2. Show that if h is a bounded sesquilinear form on the Hilbert spaces X and Y, then h has the representation

Mathematical Foundations -1- Sets and Sequences. Sets and Sequences

Algebra of Least Squares

ABOUT CHAOS AND SENSITIVITY IN TOPOLOGICAL DYNAMICS

62. Power series Definition 16. (Power series) Given a sequence {c n }, the series. c n x n = c 0 + c 1 x + c 2 x 2 + c 3 x 3 +

ON THE FUZZY METRIC SPACES

5 Birkhoff s Ergodic Theorem

Chapter 6 Infinite Series

Differentiable Convex Functions

Math 61CM - Solutions to homework 3

Ma 4121: Introduction to Lebesgue Integration Solutions to Homework Assignment 5

Metric Space Properties

Theorem 3. A subset S of a topological space X is compact if and only if every open cover of S by open sets in X has a finite subcover.

ON WELLPOSEDNESS QUADRATIC FUNCTION MINIMIZATION PROBLEM ON INTERSECTION OF TWO ELLIPSOIDS * M. JA]IMOVI], I. KRNI] 1.

AN EXTENSION OF SIMONS INEQUALITY AND APPLICATIONS. Robert DEVILLE and Catherine FINET

2 Banach spaces and Hilbert spaces

Geometry of LS. LECTURE 3 GEOMETRY OF LS, PROPERTIES OF σ 2, PARTITIONED REGRESSION, GOODNESS OF FIT

INTEGRATION THEORY AND FUNCTIONAL ANALYSIS MM-501

Singular Continuous Measures by Michael Pejic 5/14/10

Sequences and Series of Functions

Axioms of Measure Theory

A gentle introduction to Measure Theory

NBHM QUESTION 2007 Section 1 : Algebra Q1. Let G be a group of order n. Which of the following conditions imply that G is abelian?

A Characterization of Compact Operators by Orthogonality

TENSOR PRODUCTS AND PARTIAL TRACES

PRELIM PROBLEM SOLUTIONS

MAT1026 Calculus II Basic Convergence Tests for Series

ANSWERS TO MIDTERM EXAM # 2

Linear Elliptic PDE s Elliptic partial differential equations frequently arise out of conservation statements of the form

Common Coupled Fixed Point of Mappings Satisfying Rational Inequalities in Ordered Complex Valued Generalized Metric Spaces

MATH 205 HOMEWORK #2 OFFICIAL SOLUTION. (f + g)(x) = f(x) + g(x) = f( x) g( x) = (f + g)( x)

Research Article Approximate Riesz Algebra-Valued Derivations

x x x Using a second Taylor polynomial with remainder, find the best constant C so that for x 0,

5.1. The Rayleigh s quotient. Definition 49. Let A = A be a self-adjoint matrix. quotient is the function. R(x) = x,ax, for x = 0.

Topologie. Musterlösungen

Homework 4. x n x X = f(x n x) +

Mathematical Methods for Physics and Engineering

LinearAlgebra DMTH502

A brief introduction to linear algebra

5. Matrix exponentials and Von Neumann s theorem The matrix exponential. For an n n matrix X we define

VECTOR SEMINORMS, SPACES WITH VECTOR NORM, AND REGULAR OPERATORS

Filter banks. Separately, the lowpass and highpass filters are not invertible. removes the highest frequency 1/ 2and

CHAPTER I: Vector Spaces

1 lim. f(x) sin(nx)dx = 0. n sin(nx)dx

Math 299 Supplement: Real Analysis Nov 2013

Chapter 7 Isoperimetric problem

Linear Transformations

Sequences and Series

Ma 530 Introduction to Power Series

REAL ANALYSIS II: PROBLEM SET 1 - SOLUTIONS

On n-collinear elements and Riesz theorem

Machine Learning Theory Tübingen University, WS 2016/2017 Lecture 11

TERMWISE DERIVATIVES OF COMPLEX FUNCTIONS

Convergence of random variables. (telegram style notes) P.J.C. Spreij

REGULARIZATION OF CERTAIN DIVERGENT SERIES OF POLYNOMIALS

The value of Banach limits on a certain sequence of all rational numbers in the interval (0,1) Bao Qi Feng

Modern Algebra. Previous year Questions from 2017 to Ramanasri

Equivalent Banach Operator Ideal Norms 1

Math 341 Lecture #31 6.5: Power Series

Lecture 3 : Random variables and their distributions

In number theory we will generally be working with integers, though occasionally fractions and irrationals will come into play.

Chapter 3. Strong convergence. 3.1 Definition of almost sure convergence

The second is the wish that if f is a reasonably nice function in E and φ n

Sh. Al-sharif - R. Khalil

AH Checklist (Unit 3) AH Checklist (Unit 3) Matrices

Product measures, Tonelli s and Fubini s theorems For use in MAT3400/4400, autumn 2014 Nadia S. Larsen. Version of 13 October 2014.

Chapter 2. Periodic points of toral. automorphisms. 2.1 General introduction

Chapter 10: Power Series

Infinite Series and Improper Integrals

Integrable Functions. { f n } is called a determining sequence for f. If f is integrable with respect to, then f d does exist as a finite real number

Transcription:

Itroductio to Optimizatio Techiques Basic Cocepts of Aalysis - Real Aalysis, Fuctioal Aalysis 1

Basic Cocepts of Aalysis Liear Vector Spaces Defiitio: A vector space X is a set of elemets called vectors together with two operatios. The first operatio is additio which associates with ay two vectors x, y X a vector x y X, the sum of x ad y. The secod operatio is scalar multiplicatio which associates with ay vector x X ad ay scalar a vector x ; the scalar multiple of x by. The set X ad the operatios of additio ad scalar multiplicatio are assumed to satisfy the followig axioms: 1. x y yx. (commutative law) 2. ( x y) z x( yz). (associative law) 3. There is a ull vector i X such that x x for all x i X. 2

Basic Cocepts of Aalysis 4. ( x y) xy. (distributive laws) 5. ( ) xxx. 6. ( ) x ( x). (associative law) 7. 0 x, 1 x x. Propositio: I ay vector space, the followig properties hold: 1. x y xz implies y z. 2. xy ad 0 imply x y. (cacellatio laws) 3. xx ad x imply. 4. ( ) xxx. 5. ( x y) xy. 6. 0 0. (distributive laws) 3

Basic Cocepts of Aalysis Examples of liear vector spaces Let R ( C ): -dimesioal real (complex) coordiate space xv, x(,,,, ) { } 1 2 k k k 1 The collectio of all ifiite sequeces of real umbers forms a vector space. V Cab [, ]: Vector space of real-valued cotiuous fuctios o [ ab, ] with a orm x max x( t) a t b l p, L p spaces (see pages 13 & 14) 4

Basic Cocepts of Aalysis Norm, Ier product Defiitio: A ormed liear vector space is a vector space X o which there is defied a real-valued fuctio which maps each elemet x i X ito a real umber x called the orm of x. The orm satisfies the followig axioms: 1. x 0 for all xx, x 0 iff x (Null vector) 2. x y x + y for each x, yx 3. x x for all scalar ad each xx 5

Basic Cocepts of Aalysis Defiitio: A pre-hilbert space is a liear vector space X together with a ier product defied o X X. Correspodig to each pair of vectors x, y i X the ier product [ xy, ] of x ad y is a scalar. The ier product satisfies the followig axioms: 1. [ x, y] [ y, x]: cojugate of [ y, x] 2. [ x y, z] [ x, z] [ y, z] 3. [ xy, ] [ xy, ] 4. [ xx, ] 0 ad [ xx, ] 0 iff x 6

7 C s 7 C s 1. Covexity 2. Closedess 3. Covergece 4. Cotiuity 5. Cauchy sequece 6. Completeess 7. Compactess 7

7 C s 1. Covexity Defiitio : A set K i a liear vector space is said to be covex if, give x1, x2 K, all poit of the form x1 (1 ) x2k for all 0,1 2. Closedess x X P 0 p P x p P P P Defiitio : A poit is said to be a closure poit of a set if, give, there is a poit satisfyig. The collectio of all closure poits of is called the closure of ad is deoted. Defiitio : A set P is said to be closed if P P. 8

7 C s P X p P P 0 x x p P P P Defiitio : Let be a subset of a ormed space. The poit is said to be a iterior poit of if there is a such that all vectors satisfyig are also members of. The collectio of all iterior poits of is called the iterior of ad is deoted. P P P Defiitio : A set is said to be ope if. P S( x, ) : ope sphere cetered at x with radius Sx (, ) { y: x y } 9

A poit is a iterior poit S( x, ) P 3. Covergece 7 C s x P 0 if there is such that Defiitio : I a ormed liear space a ifiite sequece of vectors is said to coverge to a vector if the sequece { x } x x x x. write of real umbers coverges to zero. I this case, we x Defiitio : A trasformatio from a vector space ito the space of real (or complex) scalars is said to be a fuctioal o X X 10

7 C s 4. Cotiuity T 0 Y x0 X 0 xx0 T( x) T( x ). Defiitio : A trasformatio mappig a ormed space ito a ormed space is cotiuous at if for every there is a such that implies that If is cotiuous at each poit, we say is cotiuous. 0 T x0 X T X 11

7 C s 5. Cauchy sequece Defiitio: A sequece { x } i a ormed space is said to be a Cauchy sequece if x x 0 as m, ; i.e., give 0, there is a iteger such that x x for all m, N. Every coverget sequece m N m Cauchy sequece 12

7 C s 6. Completeess Defiitio: A ormed liear vector space X is complete if every Cauchy sequece from X has a limit i X. A complete ormed liear vector space is called a Baach space. Examples of Baach spaces: C[0,1], l, L p p cosists of all sequeces of scalars {,, } for which lp 1 2 p p, x i i p i1 i1 1/ p 13

7 C s L [ cosists of those real-valued measurable fuctios o p a, b ] x the iterval [ ab, ] for which xt () p is Lebesque itegrable. The orm o this space is defied as b p 1/ () a x x t dt p 7. Compactess Defiitio: A set K i a ormed space X is said to be compact if, give a arbitrary sequece { x i } i K, there is a subsequece { } covergig to a elemet x K. p x i I fiite dimesios, compactess closed ad boudedess 14

7 C s f x0 X 0, 0 f( x) f( x0) x x0 Defiitio : A real valued fuctioal defied o a ormed space is said to be upper semicotiuous at if, give there is a such that for. Theorem: (Weierstrass) A upper semicotiuous fuctioal o a compact subset K of a ormed liear space X achieves a maximum o K. Defiitio : A complete pre-hilbert space is called a Hilbert space Defiitio : I a pre-hilbert space two vectors x, y are said to be orthogoal if [ xy, ] 0. We symbolize this by x y. A vector x is said to be orthogoal to a set S (writte x S ) if x s for each s S. X 15

Projectio Theorem Theorem : (The classical Projectio Theorem) Let H be a Hilbert space ad M a closed subspace of H. Correspodig to ay vector x H, there is a uique vector m0 M such that xm0 xm for all m M. Furthermore, a ecessary ad sufficiet coditio that m0 M be the uique miimizig vector is that x m0 be orthogoal to M x M x m 0 m 0 16

Projectio Theorem Defiitio : Give a subset S of a pre-hilbert space, the set of the all vectors orthogoal to S is called the orthogoal complemet of S ad is deoted S. Propositio : Let S ad T be subsets of a Hilbert space, the 1. S is a closed subspace. 2. S S. 3. If S T, the T S. 4. S S 5. S S, i.e., S is the smallest closed subspace cotaiig S 17

Gram-Schmidt Procedure Defiitio : A set S of vectors i a pre-hilbert space is said to be a orthogoal set if x y for each x, y S, x y. The set is said to be orthoormal if, i additio, each vector i the set has orm equal to uity. Propositio : A orthogoal set of ozero vectors is a liearly idepedet set. GSOP... Arbitrary liearly idepedet set orthoormal set. Theorem : (Gram-Schmidt) Let { x i } be a coutable or fiite sequece of liearly idepedet vectors i a pre-hilbert space X. The, there is a orthoormal sequece { e i } such that for each the space geerated by the first ei s is the space geerated by the first x i s ; i.e., for each we have S e, e,, e Sx, x,, x 1 2 1 2 18

Gram-Schmidt Procedure Proof. For the first vector, take e 1 x x S{ e} S{ x} 1 1 1 1 e 2 Form i two steps. First, put z x x, e e z2 e2 z 2 2 2 1 1 2 z 2 e 1, e2 e, 1 z2 S e e = Sx, x 1, 2 1 2 19

Gram-Schmidt Procedure z 2 x 2 x x, e e 2 2 1 1 Gram-Schmidt procedure x, e e 2 1 1 e 1 The remaiig s are defied by iductio. The vector is formed accordig to the equatio ad z e i 1 z x x, e e i i i1 e z z 20

Gram-Schmidt Procedure It is easily verified by direct computatio that i for all, ad sice it is a liear combiatio of idepedet vectors. S z e = Sx x x e1, e2,,,,, 1 2 z e i If the origial collectio of s is fiite, the process termiates; otherwise the process produces a ifiite sequece of orthoormal sequeces. x i 21

Approximatios We ivestigate the followig approximatio problem. y1, y2,..., y H Suppose, M = Sy y y x H,,..., 1 2 Give a arbitrary vector, we seek the ˆx M vector i which is closest to. x Let ˆ 1 12 2... x y y y The problem is equivalet to that of fidig, mi x 1y1 2y2... y i 1,2,..., i 22

Normal Equatios ad Gram Matrices Accordig to the projectio theorem, there exists the uique miimizig vector ˆx : orthogoal projectio of x o M. such that x xˆ M x1y12y2... y, yi 0 for Or, Equivaletly i 1, 2,..., [ y1, y1] 1 [ y2, y1] 2 [ y, y1] [ x, y1] [ y1, y2] 1 [ y2, y2] 2 [ y, y2] [ x, y2] [ y, y ] [ y, y ] [ y, y ] [ x, y ] 1 1 2 2 23

Normal Equatios ad Gram Matrices These equatios i the coefficiets are kow as Normal equatios for the miimizatio problem. Gram matrix: traspose of the coefficiet matrix of the ormal equatios G G( y, y,..., y ) 1 2 i [ y1, y1] [ y1, y2] [ y1, y] [ y2, y1] [ y2, y2] [ y2, y] [ y, y1] [ y, y2] [ y, y] 24