MATH529 Fundamentals of Optimization Unconstrained Optimization II

Similar documents
Functions of Several Variables

MAT 419 Lecture Notes Transcribed by Eowyn Cenek 6/1/2012

Math (P)refresher Lecture 8: Unconstrained Optimization

Chapter 2: Unconstrained Extrema

MATH 5720: Unconstrained Optimization Hung Phan, UMass Lowell September 13, 2018

MATH 4211/6211 Optimization Basics of Optimization Problems

Here each term has degree 2 (the sum of exponents is 2 for all summands). A quadratic form of three variables looks as

Chapter 13. Convex and Concave. Josef Leydold Mathematical Methods WS 2018/19 13 Convex and Concave 1 / 44

The general programming problem is the nonlinear programming problem where a given function is maximized subject to a set of inequality constraints.

Mathematical Economics: Lecture 16

SIMPLE MULTIVARIATE OPTIMIZATION

The Derivative. Appendix B. B.1 The Derivative of f. Mappings from IR to IR

MATHEMATICAL ECONOMICS: OPTIMIZATION. Contents

JUST THE MATHS UNIT NUMBER 1.6. ALGEBRA 6 (Formulae and algebraic equations) A.J.Hobson

Lecture Unconstrained optimization. In this lecture we will study the unconstrained problem. minimize f(x), (2.1)

1 Overview. 2 A Characterization of Convex Functions. 2.1 First-order Taylor approximation. AM 221: Advanced Optimization Spring 2016

Math 273a: Optimization Basic concepts

OR MSc Maths Revision Course

Functions of Several Variables

Chapter 11. Taylor Series. Josef Leydold Mathematical Methods WS 2018/19 11 Taylor Series 1 / 27

MATH529 Fundamentals of Optimization Constrained Optimization I

CHAPTER 4: HIGHER ORDER DERIVATIVES. Likewise, we may define the higher order derivatives. f(x, y, z) = xy 2 + e zx. y = 2xy.

Introduction to Unconstrained Optimization: Part 2

Monotone Function. Function f is called monotonically increasing, if. x 1 x 2 f (x 1 ) f (x 2 ) x 1 < x 2 f (x 1 ) < f (x 2 ) x 1 x 2

We have seen that for a function the partial derivatives whenever they exist, play an important role. This motivates the following definition.

Computational Optimization. Mathematical Programming Fundamentals 1/25 (revised)

1 Convexity, concavity and quasi-concavity. (SB )

Paul Schrimpf. October 18, UBC Economics 526. Unconstrained optimization. Paul Schrimpf. Notation and definitions. First order conditions

g(t) = f(x 1 (t),..., x n (t)).

Course Summary Math 211

Preliminary draft only: please check for final version

Optimization. Escuela de Ingeniería Informática de Oviedo. (Dpto. de Matemáticas-UniOvi) Numerical Computation Optimization 1 / 30

Chapter 7. Extremal Problems. 7.1 Extrema and Local Extrema

Solution Methods. Richard Lusby. Department of Management Engineering Technical University of Denmark

Extreme Abridgment of Boyd and Vandenberghe s Convex Optimization

3 Applications of partial differentiation

1 Introduction to Optimization

EC /11. Math for Microeconomics September Course, Part II Problem Set 1 with Solutions. a11 a 12. x 2

Lec3p1, ORF363/COS323

REVIEW OF DIFFERENTIAL CALCULUS

Lecture 3: Basics of set-constrained and unconstrained optimization

8.7 Taylor s Inequality Math 2300 Section 005 Calculus II. f(x) = ln(1 + x) f(0) = 0

Week 4: Calculus and Optimization (Jehle and Reny, Chapter A2)

MA102: Multivariable Calculus

ISM206 Lecture Optimization of Nonlinear Objective with Linear Constraints

12. Cholesky factorization

1 Directional Derivatives and Differentiability

Unconstrained Optimization

Numerical Optimization

OPER 627: Nonlinear Optimization Lecture 2: Math Background and Optimality Conditions

Analysis/Calculus Review Day 3

Linear Models Review

Optimization and Calculus

Lines, parabolas, distances and inequalities an enrichment class

Constrained Optimization and Lagrangian Duality

NONLINEAR. (Hillier & Lieberman Introduction to Operations Research, 8 th edition)

Lecture 2 - Unconstrained Optimization Definition[Global Minimum and Maximum]Let f : S R be defined on a set S R n. Then

Real Symmetric Matrices and Semidefinite Programming

The coordinates of the vertex of the corresponding parabola are p, q. If a > 0, the parabola opens upward. If a < 0, the parabola opens downward.

Support Vector Machine (SVM) & Kernel CE-717: Machine Learning Sharif University of Technology. M. Soleymani Fall 2012

Scientific Computing: Optimization

Note: Every graph is a level set (why?). But not every level set is a graph. Graphs must pass the vertical line test. (Level sets may or may not.

Tutorials in Optimization. Richard Socher

Performance Surfaces and Optimum Points

Appendix A Taylor Approximations and Definite Matrices

Preliminaries Lectures. Dr. Abdulla Eid. Department of Mathematics MATHS 101: Calculus I

(1) Recap of Differential Calculus and Integral Calculus (2) Preview of Calculus in three dimensional space (3) Tools for Calculus 3

CS100: DISCRETE STRUCTURES. Lecture 3 Matrices Ch 3 Pages:

Math 10C - Fall Final Exam

Solutionbank Edexcel AS and A Level Modular Mathematics

SECTION A. f(x) = ln(x). Sketch the graph of y = f(x), indicating the coordinates of any points where the graph crosses the axes.

Linear algebra and applications to graphs Part 1

MATH Max-min Theory Fall 2016

AM 205: lecture 18. Last time: optimization methods Today: conditions for optimality

EC /11. Math for Microeconomics September Course, Part II Lecture Notes. Course Outline

Quadratic Programming

In view of (31), the second of these is equal to the identity I on E m, while this, in view of (30), implies that the first can be written

CHAPTER 2: CONVEX SETS AND CONCAVE FUNCTIONS. W. Erwin Diewert January 31, 2008.

9.5. Polynomial and Rational Inequalities. Objectives. Solve quadratic inequalities. Solve polynomial inequalities of degree 3 or greater.

Recitation 1. Gradients and Directional Derivatives. Brett Bernstein. CDS at NYU. January 21, 2018

Calculus 2502A - Advanced Calculus I Fall : Local minima and maxima

Optimization Tutorial 1. Basic Gradient Descent

Calculus and optimization

E 600 Chapter 3: Multivariate Calculus

Table of mathematical symbols - Wikipedia, the free encyclopedia

MAT 473 Intermediate Real Analysis II

Math 443 Differential Geometry Spring Handout 3: Bilinear and Quadratic Forms This handout should be read just before Chapter 4 of the textbook.

ARE211, Fall2015. Contents. 2. Univariate and Multivariate Differentiation (cont) Taylor s Theorem (cont) 2

Boston College. Math Review Session (2nd part) Lecture Notes August,2007. Nadezhda Karamcheva www2.bc.

6 Optimization. The interior of a set S R n is the set. int X = {x 2 S : 9 an open box B such that x 2 B S}

Quadratic Formula: - another method for solving quadratic equations (ax 2 + bx + c = 0)

Differentiable Functions

Structural and Multidisciplinary Optimization. P. Duysinx and P. Tossings

A A x i x j i j (i, j) (j, i) Let. Compute the value of for and

STA 4273H: Statistical Machine Learning

Chapter 8 ~ Quadratic Functions and Equations In this chapter you will study... You can use these skills...

Nonlinear Optimization

ECON 5111 Mathematical Economics

APPENDIX : PARTIAL FRACTIONS

Extrema of Functions of Several Variables

Transcription:

MATH529 Fundamentals of Optimization Unconstrained Optimization II Marco A. Montes de Oca Mathematical Sciences, University of Delaware, USA 1 / 31

Recap 2 / 31

Example Find the local and global minimizers and maximizers on R of f (x) = 3x 4 4x 3 + 1. 3 / 31

Graph of f (x) = 3x 4 4x 3 + 1. 4 / 31

Two theorems summarize the basic facts about global optimization of one variable functions. Theorem (1st order condition (Necessary, but not sufficient)) Suppose that f (x) is a differentiable function on R (or the function s domain I ). If x is a global minimizer of f (x), then f (x ) = 0. 5 / 31

Two theorems summarize the basic facts about global optimization of one variable functions. Theorem (1st order condition (Necessary, but not sufficient)) Suppose that f (x) is a differentiable function on R (or the function s domain I ). If x is a global minimizer of f (x), then f (x ) = 0. Theorem (2nd order condition (Sufficient, but not necessary)) Suppose that f (x), f (x), and f (x) are all continuous on R (or I ) and that x is a critical point of f (x). a) If f (x) 0 for all x R (or I ), then x is a global minimizer of f (x) on R (or I ). b) If f (x) > 0 for all x I such that x x, then x is a strict global minimizer of f (x) on R (or I ). 6 / 31

Local optimization is easier to verify. Theorem (1st order condition (Necessary, but not sufficient)) Suppose that f (x) is a differentiable function on R (or I ). If x is a local minimizer of f (x), then f (x ) = 0. 7 / 31

Local optimization is easier to verify. Theorem (1st order condition (Necessary, but not sufficient)) Suppose that f (x) is a differentiable function on R (or I ). If x is a local minimizer of f (x), then f (x ) = 0. Theorem (2nd order condition (Sufficient, but not necessary)) Suppose that f (x), f (x), and f (x) are all continuous on R (or I ) and that x is a critical point of f (x). If f (x ) > 0, then x is a strict local minimizer of f (x). 8 / 31

Exercise Find the local and global minimizers and maximizers on I = ( 1, 1) of f (x) = ln(1 x 2 ). 9 / 31

What about functions of many variables? 10 / 31

What about functions of many variables? Extend theorems that allow us to identify and classifly local minimizers of one variable functions to multivariable cases. 11 / 31

Notation: A vector in R n is an ordered n-tuple x = x 1 x 2 x 3. x n of real numbers called components of x. If x and y are vectors in R n, then their dot product or inner product is defined by y 2 x y = x T y = (x 1, x 2, x 3,..., x n ) y 3 =. y n y 1 n x i y i i=1 where x T is the transpose of x. 12 / 31

Notation: If f (x) is a function of n variables with continuous first and second partial derivatives on R n, then the gradient of f (x) is the vector f x 1 f f x n x 2 f f = x 3. 13 / 31

Notation: The Hessian of f (x), denoted by 2 f or Hf, is the symmetric n n matrix 2 f = Hf = x 2 1 x 2 x 1 x 3 x 1. x n x 1 x 1 x 2 x 2 2 x 3 x 2. x n x 2 x 1 x 3... 2 f x 1 x n x 2 x 3... x 2 3. x 2 x n... x 3 x n.... x n x 3... 2 f xn 2 14 / 31

Definition Suppose f (x) is a real-valued function defined on a subset D of R n. A point x in D is: A global minimizer for f (x) on D if f (x ) f (x) for all x D; 15 / 31

Definition Suppose f (x) is a real-valued function defined on a subset D of R n. A point x in D is: A global minimizer for f (x) on D if f (x ) f (x) for all x D; A strict global minimizer for f (x) on D if f (x ) < f (x) for all x D such that x x ; 16 / 31

Definition Suppose f (x) is a real-valued function defined on a subset D of R n. A point x in D is: A global minimizer for f (x) on D if f (x ) f (x) for all x D; A strict global minimizer for f (x) on D if f (x ) < f (x) for all x D such that x x ; A local minimizer for f (x) if there is a positive number δ such that f (x ) f (x) for all x D for which x in B(x, δ); 17 / 31

Definition Suppose f (x) is a real-valued function defined on a subset D of R n. A point x in D is: A global minimizer for f (x) on D if f (x ) f (x) for all x D; A strict global minimizer for f (x) on D if f (x ) < f (x) for all x D such that x x ; A local minimizer for f (x) if there is a positive number δ such that f (x ) f (x) for all x D for which x in B(x, δ); A strict local minimizer for f (x) if there is a positive number δ such that f (x ) < f (x) for all x D for which x in B(x, δ) and x x ; 18 / 31

Definition Suppose f (x) is a real-valued function defined on a subset D of R n. A point x in D is: A global minimizer for f (x) on D if f (x ) f (x) for all x D; A strict global minimizer for f (x) on D if f (x ) < f (x) for all x D such that x x ; A local minimizer for f (x) if there is a positive number δ such that f (x ) f (x) for all x D for which x in B(x, δ); A strict local minimizer for f (x) if there is a positive number δ such that f (x ) < f (x) for all x D for which x in B(x, δ) and x x ; A critical point (also called a stationary point) of f (x) if the first partial derivatives of f (x) exist at x and f x i = 0, for i = 1, 2, 3..., n. 19 / 31

Theorem (Multivariable Taylor s formula) Suppose that x, x are points in R n and that f (x) is a real-valued function of n variables with continuous first and second partial derivatives on some open set containing the line segment [x, x] = {w R n : w = x + t(x x ), 0 t 1} joining x and x. Then, there exists a z [x, x] such that f (x) = f (x ) + f (x ) T (x x ) + 1 2 (x x ) T Hf (z)(x x ) 20 / 31

Theorem (Local minimizer identification) Suppose that f (x) is a real-valued function for which all first partial derivatives of f (x) exist on a subset D R n. If x is an interior point of D that is a local minimizer of f (x), then f (x ) = 0. 21 / 31

Theorem (Classification of minimizers (maximizers)) Suppose that x is a critical point of a function f (x) with continuous first and second partial derivatives on R n. Then: x is a global minimizer of f (x) if (x x ) T Hf (z)(x x ) 0 for all x R n and all z [x, x]; x is a strict global minimizer of f (x) if (x x ) T Hf (z)(x x ) > 0 for all x R n such that x x and for all z [x, x]; x is a global maximizer of f (x) if (x x ) T Hf (z)(x x ) 0 for all x R n and all z [x, x]; x is a strict global maximizer of f (x) if (x x ) T Hf (z)(x x ) < 0 for all x R n such that x x and for all z [x, x]; 22 / 31

Practical ways to use the previous theorem: Conditions that involve the form (x x ) T Hf (z)(x x ), or in general v T Av, where A is a symmetric square matrix, call for methods to identify whether A (in our case the Hessian of the objective function) is positive or negative (semi)definite. 23 / 31

Quadratic forms: Let a 11 a 12... a 1n a 21 a 22... a 2n A =....... a n1 a n2... a nn The quadratic form Q A (x) = x T Ax = a 11 x 2 1 + a 12x 1 x 2 + a 13 x 1 x 3 +... + a ij x i x j +... + a ii x 2 i +... + a nn x 2 n. 24 / 31

Example Write the quadratic form associated with the following matrix: 1 0 2 3 A = 0 2 1/2 1 2 1/2 0 4. 3 1 4 5 25 / 31

Determining whether a quadratic form Q A (x) > 0 for all x R n. Example in class.... 26 / 31