Μια εισαγωγή στα Μαθηματικά για Οικονομολόγους

Σχετικά έγγραφα
1. For each of the following power series, find the interval of convergence and the radius of convergence:


Introduction of Numerical Analysis #03 TAGAMI, Daisuke (IMI, Kyushu University)

1. Matrix Algebra and Linear Economic Models

Nowhere-zero flows Let be a digraph, Abelian group. A Γ-circulation in is a mapping : such that, where, and : tail in X, head in

Statistical Inference I Locally most powerful tests

n r f ( n-r ) () x g () r () x (1.1) = Σ g() x = Σ n f < -n+ r> g () r -n + r dx r dx n + ( -n,m) dx -n n+1 1 -n -1 + ( -n,n+1)

SOLUTIONS TO MATH38181 EXTREME VALUES AND FINANCIAL RISK EXAM

Last Lecture. Biostatistics Statistical Inference Lecture 19 Likelihood Ratio Test. Example of Hypothesis Testing.

Biorthogonal Wavelets and Filter Banks via PFFS. Multiresolution Analysis (MRA) subspaces V j, and wavelet subspaces W j. f X n f, τ n φ τ n φ.

Other Test Constructions: Likelihood Ratio & Bayes Tests

Reminders: linear functions

MATH 38061/MATH48061/MATH68061: MULTIVARIATE STATISTICS Solutions to Problems on Matrix Algebra

IIT JEE (2013) (Trigonomtery 1) Solutions

Chapter 6: Systems of Linear Differential. be continuous functions on the interval

Tridiagonal matrices. Gérard MEURANT. October, 2008

2 Composition. Invertible Mappings

Solutions: Homework 3

Every set of first-order formulas is equivalent to an independent set

Ordinal Arithmetic: Addition, Multiplication, Exponentiation and Limit

COMMON RANDOM FIXED POINT THEOREMS IN SYMMETRIC SPACES

The challenges of non-stable predicates

Fourier Series. constant. The ;east value of T>0 is called the period of f(x). f(x) is well defined and single valued periodic function

Homework for 1/27 Due 2/5

Supplementary Materials: Trading Computation for Communication: Distributed Stochastic Dual Coordinate Ascent

ΚΟΙΛΕΣ KAI ΟΙΟΝΕΙ-ΚΟΙΛΕΣ ΣΥΝΑΡΤΗΣΕΙΣ. R και καθε αριθμο α οριζουμε

α β

Lecture 17: Minimum Variance Unbiased (MVUB) Estimators

New bounds for spherical two-distance sets and equiangular lines

Math 446 Homework 3 Solutions. (1). (i): Reverse triangle inequality for metrics: Let (X, d) be a metric space and let x, y, z X.

Στα επόμενα θεωρούμε ότι όλα συμβαίνουν σε ένα χώρο πιθανότητας ( Ω,,P) Modes of convergence: Οι τρόποι σύγκλισης μιας ακολουθίας τ.μ.

Fractional Colorings and Zykov Products of graphs

ΚΟΙΛΕΣ KAI ΟΙΟΝΕΙ ΚΟΙΛΕΣ ΣΥΝΑΡΤΗΣΕΙΣ. R και καθε αριθμο οριζουμε. Την καμπυλη αδιαφοριας(indifference curve,level set) της f

On Inclusion Relation of Absolute Summability

SCITECH Volume 13, Issue 2 RESEARCH ORGANISATION Published online: March 29, 2018

ΠΑΝΕΠΙΣΤΗΜΙΟ ΙΩΑΝΝΙΝΩΝ ΤΜΗΜΑ ΟΙΚΟΝΟΜΙΚΩΝ ΕΠΙΣΤΗΜΩΝ ΜΕΤΑΠΤΥΧΙΑΚΟ ΠΡΟΓΡΑΜΜΑ ΣTHN ΟΙΚΟΝΟΜΙΚΗ ΑΝΑΛΥΣΗ ΕΙΔΙΚΑ ΘΕΜΑΤΑ ΜΙΚΡΟΟΙΚΟΝΟΜΙΚΗΣ.

C.S. 430 Assignment 6, Sample Solutions

5. Choice under Uncertainty

Chapter 6: Systems of Linear Differential. be continuous functions on the interval

The ε-pseudospectrum of a Matrix

Homework 8 Model Solution Section

Congruence Classes of Invertible Matrices of Order 3 over F 2

Optimal Parameter in Hermitian and Skew-Hermitian Splitting Method for Certain Two-by-Two Block Matrices

FREE VIBRATION OF A SINGLE-DEGREE-OF-FREEDOM SYSTEM Revision B

Lecture 21: Properties and robustness of LSE

Ψηφιακή Επεξεργασία Εικόνας

SOLUTIONS TO MATH38181 EXTREME VALUES AND FINANCIAL RISK EXAM

A Note on Intuitionistic Fuzzy. Equivalence Relation

On Generating Relations of Some Triple. Hypergeometric Functions

SUPPLEMENT TO ROBUSTNESS, INFINITESIMAL NEIGHBORHOODS, AND MOMENT RESTRICTIONS (Econometrica, Vol. 81, No. 3, May 2013, )

Matrices and Determinants

SUPERPOSITION, MEASUREMENT, NORMALIZATION, EXPECTATION VALUES. Reading: QM course packet Ch 5 up to 5.6

On Certain Subclass of λ-bazilevič Functions of Type α + iµ

Sequent Calculi for the Modal µ-calculus over S5. Luca Alberucci, University of Berne. Logic Colloquium Berne, July 4th 2008

LAD Estimation for Time Series Models With Finite and Infinite Variance

ORDINAL ARITHMETIC JULIAN J. SCHLÖDER

Inverse trigonometric functions & General Solution of Trigonometric Equations

Presentation of complex number in Cartesian and polar coordinate system

CHAPTER 103 EVEN AND ODD FUNCTIONS AND HALF-RANGE FOURIER SERIES

SCHOOL OF MATHEMATICAL SCIENCES G11LMA Linear Mathematics Examination Solutions

A study on generalized absolute summability factors for a triangular matrix

Solution Series 9. i=1 x i and i=1 x i.

The Simply Typed Lambda Calculus

Uniform Convergence of Fourier Series Michael Taylor

6.1. Dirac Equation. Hamiltonian. Dirac Eq.

Parametrized Surfaces

Πρόβλημα 1: Αναζήτηση Ελάχιστης/Μέγιστης Τιμής

EE512: Error Control Coding

Example Sheet 3 Solutions

Three Classical Tests; Wald, LM(Score), and LR tests

Phys460.nb Solution for the t-dependent Schrodinger s equation How did we find the solution? (not required)

ST5224: Advanced Statistical Theory II

Finite Field Problems: Solutions

Quadratic Expressions

The Equivalence Theorem in Optimal Design

CHAPTER 25 SOLVING EQUATIONS BY ITERATIVE METHODS

ANSWERSHEET (TOPIC = DIFFERENTIAL CALCULUS) COLLECTION #2. h 0 h h 0 h h 0 ( ) g k = g 0 + g 1 + g g 2009 =?

Supplemental Material: Scaling Up Sparse Support Vector Machines by Simultaneous Feature and Sample Reduction

Lecture 2: Dirac notation and a review of linear algebra Read Sakurai chapter 1, Baym chatper 3

ω ω ω ω ω ω+2 ω ω+2 + ω ω ω ω+2 + ω ω+1 ω ω+2 2 ω ω ω ω ω ω ω ω+1 ω ω2 ω ω2 + ω ω ω2 + ω ω ω ω2 + ω ω+1 ω ω2 + ω ω+1 + ω ω ω ω2 + ω

Numerical Analysis FMN011

Lecture 13 - Root Space Decomposition II

Homework 3 Solutions

Partial Differential Equations in Biology The boundary element method. March 26, 2013

Ενότητα 3: Ακρότατα συναρτήσεων μίας ή πολλών μεταβλητών. Νίκος Καραμπετάκης Τμήμα Μαθηματικών

Lecture 15 - Root System Axiomatics

Exercises 10. Find a fundamental matrix of the given system of equations. Also find the fundamental matrix Φ(t) satisfying Φ(0) = I. 1.

Μαθηματικά Πληροφορικής Συνδυαστικά Θεωρήματα σε Πεπερασμένα Σύνολα

Concrete Mathematics Exercises from 30 September 2016

Homomorphism in Intuitionistic Fuzzy Automata

Second Order Partial Differential Equations

Generating Set of the Complete Semigroups of Binary Relations

k A = [k, k]( )[a 1, a 2 ] = [ka 1,ka 2 ] 4For the division of two intervals of confidence in R +

The Heisenberg Uncertainty Principle

Second Order RLC Filters

Απόκριση σε Μοναδιαία Ωστική Δύναμη (Unit Impulse) Απόκριση σε Δυνάμεις Αυθαίρετα Μεταβαλλόμενες με το Χρόνο. Απόστολος Σ.

Econ 2110: Fall 2008 Suggested Solutions to Problem Set 8 questions or comments to Dan Fetter 1

Practice Exam 2. Conceptual Questions. 1. State a Basic identity and then verify it. (a) Identity: Solution: One identity is csc(θ) = 1

MATHEMATICS. 1. If A and B are square matrices of order 3 such that A = -1, B =3, then 3AB = 1) -9 2) -27 3) -81 4) 81

6. MAXIMUM LIKELIHOOD ESTIMATION

Transcript:

Μια εισαγωγή στα Μαθηματικά για Οικονομολόγους

Μαθηματικά Ικανές και αναγκαίες συνθήκες Έστω δυο προτάσεις Α και Β «Α είναι αναγκαία συνθήκη για την Β» «Α είναι ικανή συνθήκη για την Β» Α is ecessary for B Α is sufficiet for B A B A B «Α είναι ικανή και αναγκαία συνθήκη για την Β» A if-f B or A is ecessary ad sufficiet for B A B

Μαθηματικά Θεωρήματα και αποδείξεις Τα θεωρήματα έχουν συνήθως τη μορφή A B Α = προϋπόθεση Για να αποδείξουμε το θεώρημα θα πρέπει να δείξουμε την ορθότητα του Β = συμπέρασμα συμπεράσματος, Β, χρησιμοποιώντας την αλήθεια της προϋπόθεσης, Α, με όρους βασικής λογικής costructive proof: A B cotrapositive proof: ~ A ~ B proof by cotradictio: Υποθέσεις: Α αληθές και Β όχι αληθές τότε να καταλήξουμε σε ΑΤΟΠΟ

Μαθηματικά Θεωρία Συνόλων Ένα σύνολο S είναι υποσύνολο (subset) ενός άλλου συνόλου Τ : S T x S x T υο σύνολα είναι ισα, εάν περιέχουν ακριβώς τα ίδια στοιχεία: Κενό σύνολο S = T x S x T ad x T x S S T ad T S Συμπλήρωμα (complemet) συνόλου S T x\ x S or x T Ένωση (uio) συνόλων { } S T x\ x S ad x T Τομή (itersectio) συνόλων { }

Μαθηματικά Θεωρία Συνόλων S T (,)\ s t s S, t T Γινόμενο (product of two sets) συνόλων: { } Στη Μικροοικονομική Θεωρία περιοριζόμαστε στα υποσύνολα: ++ + Θα γράφω x 0 ή x και θα εννοώ x 0, i= 1,2.. + x>> 0 ή x και θα εννοώ x > 0, i= 1,2.. ++ i i

Μαθηματικά Covex Set Κυρτά σύνολα (covex sets) στον S είναι κυρτό σύνολο εάν για κάθε 1 x S και 2 x S 1 2 tx t x S t [ ] + (1 ), 0,1

Μαθηματικά Fuctios f : A B A = domai B = rage f ( a) = b a A = όρισμα της f, b B = τιμή, εικόνα της f στο α

Μαθηματικά Fuctios Μια συνάρτηση f : A B λέγεται 1-1 ή αμφιμονοσήμαντη, όταν αντιστοιχίζει κάθε όρισμα σε αποκλειστικά δική του τιμή ή αλλιώς, όταν διαφορετικά ορίσματα απεικονίζονται σε διαφορετικές τιμές: αν a a τότε f ( a) f( a ) Μια συνάρτηση f : A B λέγεται επί (oto), όταν δεν υπάρχει στοιχείο στο Β που να μην είναι η εικόνα κάποιου στοιχείου του Α: b B, a A: b= f( a) Μια συνάρτηση είναι αντιστρέψιμη ( f 1 ) αν είναι 1-1 και επί: 1 1 f B A f b = a b= f a : ( ) ( )

Μαθηματικά Topology Μετρική είναι απλά ένα μέγεθος μέτρησης της απόστασης Μετρικός χώρος είναι ένα σύνολο στο οποίο έχει οριστεί μια έννοια απόστασης (μια μετρική). Μετρικοί χώροι Μετρική {( x, x )\ x, x } 2 = 1 2 1 2 {( x, x,.. x )\ x, i 1,2... } 1 2 i = = dx (, x) = x x 1 2 1 2 dx (, x) = ( x x) + ( x x) 1 2 2 1 2 2 1 2 1 1 2 2 dx (, x) = ( x x) +... + ( x x) 1 2 2 1 2 2 1 2 1 1

Μαθηματικά Topology Defiitio: Ope ad Closed ε-balls 1. The ope ε-ball with ceter 0 x ad radius ε > 0 (a real umber) is the subset of poits i : { ε} B x x d x x < 0 0 ε ( ) \ (, ) 2. The closed ε-ball with ceter 0 x ad radius ε > 0 (a real umber) is the subset of poits i : { ε} 0 0 Bε x x d x x ( ) \ (, )

Μαθηματικά Topology Defiitio: Ope sets i S is a ope set if, for all x S, there exists some ε > 0 such that B ( x) S ε Theorem: O Ope Sets i 1. is a ope set 2. is a ope set 3. The uio of ope sets is a ope set 4. The itersectio of ay fiite collectio of ope sets is a ope set

Μαθηματικά Topology Defiitio: Closed sets i S is a closed set if, its complemet c S, is a ope set Theorem: O Closed Sets i 1. is a closed set 2. is a closed set 3. The uio of ay fiite collectio of closed sets is a closed set 4. The itersectio of closed sets is a closed set

Μαθηματικά Topology Defiitio: Bouded sets i A set S is called bouded if it is etirely cotaied withi some ε-ball (either ope or closed). That is S is bouded if there exists some ε > 0 such that S B ( x) x. for some ε

Μαθηματικά Topology Defiitio (Heie Borel): Compact sets i A set S is called compact if it is closed ad bouded

Μαθηματικά Real Valued Fuctio Defiitio: Caushy cotiuity Let m D ad let f : every ε > 0, there is a δ > 0 such that: D. The fuctio f is cotiuous at the poit 0 0 ( δ( ) ) ε ( ( )) f B x D B f x 0 x D if for If f is cotiuous at every poit x D, the it is called a cotiuous fuctio.

Μαθηματικά Real Valued Fuctio Theorem: Cotiuity ad iverse images Let m D. The followig coditios are equivalet: 1. f : D is cotiuous 2. For every ope ball B i, f 1 ( B) is ope i D 3. For every ope S 1, f ( S) is ope i D Theorem: The Cotiuous Image of a Compact Set is a Compact Set Let m D ad let f : D be a cotiuous fuctio. If S D (closed ad bouded) the its image f ( S) is a compact set. is compact

Μαθηματικά Real Valued Fuctio Theorem (Weierstrass): Existece of Extreme Values Let f : S be a cotiuous real-valued fuctio, where S is a oempty compact subset of. The there exists a vector x S ad a vector x S such that: f ( x ) f( x) f( x), x S

Μαθηματικά Real Valued Fuctio Defiitio: Real Valued Fuctios f : D R is a real-valued fuctio if D is ay set ad R Defiitio: Icreasig Real Valued Fuctios Let f : D, where D. The ( ) ( ) 0 1 f is icreasig if f x f x wheever x x 0 1 ( ) ( ) 0 1 f is strictly icreasig if f x > f x wheever x >> x 0 1 f is strogly icreasig if 0 1 ( ) f ( x ) f x > wheever x x ad 0 1 x x 0 1

Μαθηματικά Real Valued Fuctio Defiitio: Decreasig Real Valued Fuctios Let f : D, where D. The ( ) ( ) 0 1 f is decreasig if f x f x wheever x x 0 1 ( ) ( ) 0 1 f is strictly decreasig if f x < f x wheever x >> x 0 1 f is strogly decreasig if 0 1 ( ) f ( x ) f x < wheever x x ad 0 1 x x 0 1

Μαθηματικά Level Sets Defiitio: Level Sets 0 L( y ) is a level set of the real-valued fuctio f : D R if-f { } 0 0 L( y ) = x\ x D, f ( x) = y, where 0 y R Defiitio: Superior ad Iferior Sets 1. S ( y 0 ) { x \ x D, f ( x ) y 0 } is called the superior set for level 2. I ( y 0 ) { x \ x D, f ( x ) y 0 } is called the iferior set for level > is called the strictly superior set for level 3. S ( y 0 ) { x \ x D, f ( x ) y 0 } < is called the strictly iferior set for level 4. I ( y 0 ) { x \ x D, f ( x ) y 0 } 0 y 0 y 0 y 0 y

Μαθηματικά Level Sets Defiitio: Superior ad Iferior Sets 1. S ( y 0 ) { x \ x D, f ( x ) y 0 } is called the superior set for level 2. I ( y 0 ) { x \ x D, f ( x ) y 0 } is called the iferior set for level > is called the strictly superior set for level 3. S ( y 0 ) { x \ x D, f ( x ) y 0 } < is called the strictly iferior set for level 4. I ( y 0 ) { x \ x D, f ( x ) y 0 } 0 y 0 y 0 y 0 y

Μαθηματικά Cocave Fuctio Defiitio: Cocave Fuctio f : D R is a cocave fuctio if for all 1 2 x, x D f x tf x + t f x t 1 2 ( ) ( ) (1 ) ( ) [0,1] where 1 2 = + (1 ) deotes the covex combiatio of x tx t x x, x 1 2

Μαθηματικά Cocave Fuctio Theorem: Poits o ad below the Graph of a Cocave fuctio always form a Covex Set Let A {( x, y)\ x D, f ( x) y} be the set of poits o ad below the graph of f : D R, where D is a covex set ad R. The f is cocave fuctio A is a covex set

Μαθηματικά Cocave Fuctio Defiitio: Strictly Cocave Fuctios f : D R is a strictly cocave fuctio if-f for all x 1 2 1 2 x i D: f ( tx + (1 t) x ) > tf( x ) + (1 t) f ( x ) t (0,1) 1 2

Μαθηματικά Quasi - Cocave Fuctio Defiitio: Quasi-Cocave Fuctio f : D R is quasi-cocave if-f for all ( ) Quasi-Cocave Fuctios 1 2 x, x D: + (1 ) mi ( ), ( ) [0,1] 1 2 1 2 f tx t x f x f x t

Μαθηματικά Quasi - Cocave Fuctio Theorem: Quasi-Cocavity ad the Superior Sets f : D is a quasi-cocave fuctio if-f S( y ) is a covex set for all y

Μαθηματικά Quasi - Cocave Fuctio Quasi-Cocave Fuctios Defiitio: Strictly Quasi-Cocave Fuctio f : D R is strictly quasi-cocave if-f for all 1 2 x x D: ( ) + (1 ) > mi ( ), ( ) (0,1) 1 2 1 2 f tx t x f x f x t Theorem: Strictly Quasi-Cocavity ad the Superior Sets f : D is a strictly quasi-cocave fuctio if-f S( y ) is a strictly covex set for all y

Μαθηματικά Quasi - Cocave Fuctio

Μαθηματικά Quasi - Cocave Fuctio Theorem: Cocavity implies Quasi-cocavity A (strictly) cocave fuctio is always (strictly) quasi-cocave Theorem: Cobb-Douglas Fuctio a b Every Cobb-Douglas fuctio f ( x1, x2) = Ax1x2 with A, ab>0, is quasi-cocave.

Μαθηματικά Covex ad Quasi - Covex Fuctio Defiitio: Covex ad Strictly Covex Fuctios 1. f : D R is a covex fuctio if for all 1 2 x, x D ( ) 1 2 1 2 f tx t x tf x t f x t + (1 ) ( ) + (1 ) ( ) [0,1] 2. f : D R is a strictly covex fuctio if for all 1 2 x x D ( ) 1 2 1 2 f tx t x tf x t f x t + (1 ) < ( ) + (1 ) ( ) (0,1)

Μαθηματικά Covex ad Quasi - Covex Fuctio Theorem: Poits o ad Above the Graph of a Cocave fuctio always form a Covex Set Let A {( x, y)\ x D, f ( x) y} be the set of poits o ad above the graph of f : D R, where D is a covex set ad R. The f is covex fuctio A is a covex set

Μαθηματικά Covex ad Quasi - Covex Fuctio Defiitio: Quasi-Covex ad Strictly q-covex Fuctio 1. f : D R is quasi-covex if-f for all 1 2 x, x D : ( ) + (1 ) max ( ), ( ) [0,1] 1 2 1 2 f tx t x f x f x t 2. f : D R is strictly quasi-covex if-f for all 1 2 x x D: ( ) + (1 ) < max ( ), ( ) (0,1) 1 2 1 2 f tx t x f x f x t

Μαθηματικά Covex ad Quasi - Covex Fuctio Theorem: (Strictly) Quasi-Covex ad the Iferior Sets f : D is a (strictly) quasi-covex fuctio if-f I ( y ) is a (strictly) covex set for all y Theorem: (Strictly) Quasi-Covex ad (Strictly) Quasi-Cocave fuctios f ( x ) is a (strictly) quasi-cocave fuctio if-f f ( x) fuctio is a (strictly) quasi-covex

Μαθηματικά Covex ad Quasi - Covex Fuctio

Μαθηματικά Calculus Calculus ad Optimizatio Fuctios of a sigle variable

Μαθηματικά Calculus

Μαθηματικά Calculus Theorem: Calculus ad Optimizatio Fuctios of a sigle variable Suppose f : D R, D, R is twice cotiuously differetiable 1. f is cocave f ( x) 0, x D 2. f is covex f ( x) 0, x D Moreover, 1. if f ( x) < 0, x D the f is strictly cocave 2. if f ( x) > 0, x D the f is strictly covex

Μαθηματικά Calculus Calculus ad Optimizatio Fuctios of several variables

Μαθηματικά Calculus Calculus ad Optimizatio Fuctios of several variables Theorem: Youg s Theorem For ay twice cotiuously differetiable fuctio f ( x), x : 2 2 f ( x) f ( x) =, i, j x x x x i j j i

Μαθηματικά Liear Algebra (Leadig) Pricipal Miors of a Matrix Defiitio: Let A be a matrix. A k k submatrix of A formed by deletig k colums ad the same k rows from A is called a k th order pricipal submatrix of A. The determiat of that pricipal submatrix is called a k th order pricipal mior of A. Defiitio: Let A be a A obtaied by deletig the last matrix. The k th order pricipal submatrix of k rows ad the last k colums from A is called the k th order leadig pricipal submatrix of A. Its determiat is called the k th order leadig pricipal mior of A

Μαθηματικά Liear Algebra Defiiteess of a Matrix Theorem: Defiiteess of a matrix Let A be a symmetric matrix (a) A is positive defiite if-f all its leadig pricipal miors are strictly positive (b) A is egative defiite if-f its leadig pricipal miors alterate i sig as follows: A < 0, A > 0, A < 0... 1 2 3

Μαθηματικά Liear Algebra Semi-Defiiteess of a Matrix Theorem: Semi-Defiiteess of a matrix Let A be a symmetric matrix (a) A is positive semi-defiite if-f every pricipal mior is 0 (b) A is egative semi-defiite if-f every pricipal mior of odd order 0 ad every pricipal mior of eve order 0

Μαθηματικά Liear Algebra Border Matrices H 0 f1 f2... f f1 f11 f12... f 1 = f2 f21 f22... f 2... f f 1 f2... f

Μαθηματικά Liear Algebra Border Matrices Theorem: Defiiteess of a bordered matrix Let H be a symmetric bordered matrix (a) H is positive defiite if-f all its bordered pricipal miors are strictly egative i.e 0 f f 0 f H = < 0 H = f f f < 0... H < 0 1 2 1 1 2 1 11 12 f1 f11 f2 f21 f22 (b) H is egative defiite if-f its bordered pricipal miors alterate i sig as follows: 0 f f 0 f H = < 0 H = f f f > 0 H < 0... 1 2 1 1 2 1 11 12 3 f1 f11 f2 f21 f22

Μαθηματικά Liear Algebra Border Matrices Theorem: Semi-Defiiteess of a bordered matrix Let H be a symmetric bordered matrix (a) H is positive semi-defiite if-f every bordered pricipal mior is 0 (b) H is egative semi-defiite if-f every bordered pricipal mior of odd order 0 ad every bordered pricipal mior of eve order 0

Μαθηματικά Liear Algebra (Border) Matrices ad (Quasi) Cocavity/Covexity Theorem: Cocavity Covexity i May Variables Let D be a covex subset of o which f is twice cotiuously differetiable f is cocave (covex) H ( x ) is egative (positive) semi-defiite, x D Moreover If H ( x ) is egative (positive) defiite x D the f is strictly cocave (covex)

Μαθηματικά Liear Algebra Theorem: Cocavity Covexity ad Secod-Order Ow Partial Derivatives Let f : D R be a twice cotiuously differetiable fuctio 1. If f cocave fii ( x) 0 i = 1, 2,... x 2. If f covex fii ( x) 0 i = 1, 2,... x

Μαθηματικά Liear Algebra (Border) Matrices ad (Quasi) Cocavity/Covexity Theorem: Quasi-Cocavity (Covexity) i may variables Let D be a covex subset of o which f is twice cotiuously differetiable f is quasi cocave (covex) H ( x ) is egative (positive) semi-defiite, x D Moreover If ( ) H x is egative (positive) defiite x D the f is strictly quasi -cocave (quasi-covex)

Μαθηματικά Homogeeous Fuctio Homogeeous Fuctios Defiitio: Homogeeous Fuctios A real-valued fuctio f ( x ) is called homogeeous of degree k, if f ( tx) = t k f ( x) t > 0

Μαθηματικά Homogeeous Fuctio Homogeeous Fuctios Theorem: Partial Derivatives of Homogeeous Fuctios If f ( x ) is h.o.d. k, its partial derivatives are h.o.d. k-1 Theorem: Euler s Theorem Let f ( x ) be a cotiuously differetiable homogeeous fuctio of degree k o The for all x + f ( x) f ( x) f ( x) x + x +... + x = kf ( x) x x x 1 2 1 2

Μαθηματικά Optimizatio Optimizatio Maxima ad miima for sigle-variable fuctios Cosider the fuctio of a sigle-variable f ( x) = y ad assume it is differetiable whe we say the fuctio achieves a local maximum at x, we mea that f x f x x B x ( ) ( ), ε ( ) whe we say the fuctio achieves a global maximum at x, we mea that f ( x ) f ( x), x D uique local maximum at x if f ( x ) > f ( x), x x B ( x ) ε uique global maximum at x if f ( x ) > f ( x), x x D

Μαθηματικά Optimizatio Optimizatio Maxima ad miima for sigle-variable fuctios Theorem: (a) If f ( x0) = 0 ad f ( x0) < 0 the x 0 is local max of f (b) If f ( x0) = 0 ad f ( x0) > 0 the x 0 is local mi of f (c) If f ( x0) = 0 ad f ( x0) = 0 the x 0 ca be max, mi, or either

Μαθηματικά Optimizatio Optimizatio Maxima ad miima for sigle-variable fuctios If f ( x ) is a twice cotiuously differetiable fuctio whose domai is a iterval I, the (a) If f ( x0 ) = 0 ad f ( x) < 0, x I the x 0 is a global max of f (b) If f ( x0 ) = 0 ad f ( x) > 0, x I the x 0 is a global mi of f

Μαθηματικά Optimizatio Optimizatio Real-valued fuctios of -variables Defiitio: Let f : D, D (1) A poit x D is a global max if f ( x ) f ( x), x D (1) A poit x D is a uique global max if f ( x ) > f ( x), x D ad x x (2) A poit x D is a local max if f ( x ) f ( x), x Bε ( x ) D (2) A poit x D is a uique local max if f ( x ) > f ( x), x Bε ( x ) D ad x x

Μαθηματικά Optimizatio Theorem: Optimizatio Real-valued fuctios of -variables Let f : D, D be a twice cotiuously differetiable fuctio If x D is a local max or mi of f ad if x is a iterior poit of D, the x solves the system f x x... ( ) 1 ( ) f x x 2 ( ) f x x = = = 0 0 0

Μαθηματικά Optimizatio Optimizatio Real-valued fuctios of -variables SECOND ORDER CONDITIONS Theorem: Sufficiet Coditios Let f : D, D be a twice cotiuously differetiable fuctio Suppose that x satisfies f ( x ) = 0, i = 1, 2,... x ad that the leadig pricipal i miors of H ( x ) alterate i sig at x. The H < 0, H > 0, H < 0... 1 2 3 x is a uique local max of f

Μαθηματικά Optimizatio Optimizatio Real-valued fuctios of -variables SECOND ORDER CONDITIONS Theorem: Necessary Coditios Let f : D, D be a twice cotiuously differetiable fuctio (1) If f ( x ) reaches a local iterior maximum at x the f x x ( ) i = 0, i = 1, 2,... ad H ( x ) is egative semi-defiite (2) If f ( x ) reaches a local iterior miimum at x the ad H ( x ) is positive semi-defiite f ( x ) x i = 0, i = 1, 2,...

Μαθηματικά Optimizatio Optimizatio Real-valued fuctios of -variables Theorem: Global Theorem Let f : D, D be a twice cotiuously differetiable fuctio which is [strictly] CONCAVE (covex) o D. The followig statemets are equivalet, where x is a iterior poit of D : (1) f x x ( ) i = 0, for i = 1, 2,... (2) f achieves a [uique] GLOBAL MAXIMUM (global miimum) at x

Μαθηματικά Optimizatio Costraied Optimizatio Equality Costraits max f ( x, x ) s.t. g( x, x ) = 0 x, x 1 2 1 2 1 2

Μαθηματικά Optimizatio Costraied Optimizatio Equality Costraits max f ( x, x ) s.t. g( x, x ) = 0 x, x 1 2 1 2 1 2 Solve: 1. By substitutio

Μαθηματικά Optimizatio Costraied Optimizatio Equality Costraits max f ( x, x ) s.t. g( x, x ) = 0 x, x 1 2 1 2 1 2 Solve: 1. By substitutio 2. Lagrage s Method

Μαθηματικά Optimizatio Costraied Optimizatio Equality Costraits Theorem: Sufficiet Coditios for Local Optima with Equality Costraits Let the objective fuctio be f ( x ) ad the m costraits be j g ( x) = 0, j = 1, 2,... m Let ( x, Λ ) solve the F.O.C. The 1. x is a local maximum of f ( x ) subject to the costraits, if the bordered pricipal miors, evaluated at ( x, Λ ), alterate i sig begiig with egative 2. x is a local miimum of f ( x ) subject to the costraits, if the bordered pricipal miors, evaluated at ( x, Λ ), are all egative

Μαθηματικά Optimizatio Costraied Optimizatio Iequality Costraits Theorem: Necessary Coditios for Optima of Real-valued fuctios s.t. Noegative Costraits Let the objective fuctio f ( x ) be cotiuously differetiable 1. If x maximizes f ( x ) s.t. x 0, the x satisfies: (i) f ( x) x i 0, i = 1, 2... (ii) (iii) f( x ) xi = 0, i = 1, 2... xi x 0 i = 1, 2,... i

Μαθηματικά Optimizatio Costraied Optimizatio Iequality Costraits KUHN-TUCKER CONDITIONS

Μαθηματικά Optimizatio Value Fuctios M ( a) f ( xa ( ), a) Theorem: Theorem of the Maximum If the objective fuctio ad the costrait are cotiuous i the parameters, ad if the domai is a compact set, the M ( a ) ad x( a ) are cotiuous i a

Μαθηματικά Optimizatio THE ENVELOPE THEOREM Cosider the problem max f ( x ; a ) s.t. g( x; a) = 0 ad x 0 x ad suppose the objective fuctio ad costrait are cotiuously differetiable i a. For each a, let xa ( ) >> 0 uiquely solve the problem ad assume that it is also cotiuously differetiable i the parameters a. Let Lxaλ (,, ) be the problem s associated Lagragia fuctio ad let ( x( a), λ ( a)) solve the Kuh-Tucker coditios. Fially, let M ( a ) be the problem s associated maximum-value fuctio. The the Evelope Theorem states that M( a) L = j = 1, 2,... m a a j j x( a) λ ( a)