Important Mathematical Tools Learning objectives: . . . . Ability to apply important mathematical tools for economic analysis Ability to give interpretations of mathematical properties in terms of economic concepts • Topics reviewed – Sets: definition, notation – Binary relations, orderings – Functions, sequences – Sets: properties – Brower’s fixed point theorem – Separation theorems (Minkowski) c Ronald Wendner GE-Math-1 v2.1 1 Sets: Definition, Notation • set notation: upper case letter use elements to define set: X ≡ {x | P (x)} X ≡ {x ∈ students at KFUG | x study economics} • elements of a set: x ∈ X, x 6∈ X most sets we use are sets of vectors (w/ an interpretation) elements: points or vectors N ≡ {x | x is a natural number}, -1.3 6∈ N R ≡ {x | x is a real number}, -1.3 ∈ R R+ ≡ {x | x is a nonnegative real number}, -1.3 6∈ R+ R++ ≡ {x | x is a strictly positive real number}, 0 6∈ R++ ∅ empty set (null set) notice that ±∞ 6∈ R • subsets Xi ⊂ X iff for all x ∈ Xi: x ∈ X masters students in econ are a subset of students in econ ∅ ∈ X (!) • set operations/relationships & logical connectives – – – – – subset relation; set equality set union, intersection, complement set subtraction (∀x ∈ A)(x ∈ B) , (∃x ∈ A)(x ∈ B) (Cartesian) set product c Ronald Wendner GE-Math-2 v2.1 • Cartesian product, cross product, direct product – set ordered n-tuples C ≡ A × B ≡ {(a, b) | a ∈ A , b ∈ B} (4, 5) ∈ C, (5, 4) ∈ C, (4, 5) 6= (5, 4) Example 1. L = 2, p = (p1, p2): pl ∈ R+, p = (p1, p2) ∈ R+ × R+ = R2+, (1, 2) 6= (2, 1) Example 2. A = B = [0, 1]. What is C = A × B? Example 3. Suppose we have L commodities. Then every commodity bundle is an element (a vector) of: RL+ ≡ R+ × R+ × R+ × .... × R+ (L-fold Cartesian product) • dot product notation: – inner (dot) product of two L-vectors x = (x1, x2, ..., xL), y = (y1, y2, ..., yL): y1 L X y2 x · y ≡ (x1, x2, ..., xL) .. = xl y l . . l=1 yL Example: x = (x1, x2, ..., xL), p = (p1, p2, ..., pL); cost of consumption bundle x is: p · x c Ronald Wendner GE-Math-3 v2.1 2 Binary Relations, Orderings Consider two elements of a set: a, b ∈ X. • binary relations (R): a R b – – – – – – – a is the brother of b a is to the left of b a>b a≥b a=b a%b ab • quasi-ordering – reflexivity: a R a for all a ∈ X – transitivity: a R b, and b R c ⇒ a R c. • complete ordering for all a, b ∈ X, either a R b or b R a or both how to compare Pizza with Rogan Josh? • Rational % are a complete ordering (binary relation) on consumption space c Ronald Wendner GE-Math-4 v2.1 • upper contour sets of % on X X +(y) ≡ {x ∈ X | x % y} • lower contour sets of % on X X −(y) ≡ {x ∈ X | y % x} • indifference set of % on X X −(y) ∩ X +(y) 3 Functions, Sequences • A function f : A → B is a rule that associates to every element of set A exactly one element in set B. Set A is called the domain of the function, and set B is called the codomain of the function. → onto if for every y ∈ B, ∃ x ∈ A | y = f (x) → one to one if for any two elements in the domain, x, x0 ∈ A with x 6= x0, it holds: f (x) 6= f (x0) Example 1. Consider f (x) : R+ → R+, with f (x) = x2. Then f (x) is onto and one to one. Example 2. f (x) : R → R+, with f (x) = x2. Then f (x) is onto but not one to one c Ronald Wendner GE-Math-5 v2.1 • Sequence f : N → R – elements: xn, n = 1, 2, ....; sequence: {xn}∞ n=1 – sequences in R: 1/n, or n/(1 + n), or 5 – sequences in R2: (1/n, n/(1 + n)), or (1/n, 5) – “sequence in set X”: xn ∈ X for all n = 1, 2, ... – convergence: limn→∞ xn = x ∗ for all > 0, ∃ n0 ∈ N : ∀n > n0, d(xn, x) < ∗ “convergence in set X”: limit x ∈ X {xn}∞ n=1 = 1/n, A = [0, 1], B = (0, 1] • subsequence m(n) = strictly increasing function, m(n) : N → N e.g., m(n) = 3 n n ∞ {xm(n)}∞ n=1 is subsequence of {x }n=1 Theorem 1 (Bolzano-Weierstrass) . Every bounded sequence in RL contains a convergent subsequence. c Ronald Wendner GE-Math-6 v2.1 • Continuity of a function Let f (x) : X → R, X ⊆ RL. f (x) is continuous at x if for every sequence {xn} → x, {f (xn)} → f (x). f (x) is continuous if function is continuous at all x ∈ X. → if f (.) is continuous: image of compact set is compact Theorem 2 (Extreme value theorem) . If f (.) is continuous and X compact then f (.) has a maximum/minimum. → compactness? ... 4 Sets in Euclidean space RL: Properties • Euclidean distance d(x, x0) 0 L 0 – x, x ∈ X ⊆ R , d(x, x ) ≡ qP L l=1 (xl − x0l )2 • -neighborhood (nbhd) about x: N(x) = {x0 ∈ X | d(x, x0) < } • open set X ⊂ RL open if for every x ∈ X there exists an > 0 such that N(x) ⊂ X. c Ronald Wendner GE-Math-7 v2.1 • closed set X ⊂ RL closed if every converging sequence in X converges in X consider X = (0, 1], {xn}∞ n=1 = 1/n example: closed interval: [a, b] ⊂ R ≡ {x ∈ R | a ≤ x ≤ b} → closed sets and maximizing bahavior • bounded set X bounded if for all x, x0 ∈ X, d(x, x0) < ∞ • compact set X ⊂ RL compact if X closed and bounded • boundary of X boundary point x ∈ X: every N(x) contains x0 ∈ X and x0 6∈ X boundary of X = set of all boundary points example: X = (0, 1], boundary points = {0, 1} • closure of set X: X X ≡ X∪ boundary of X →X⊆X → X = X iff X closed c Ronald Wendner GE-Math-8 v2.1 • interior X = biggest open set contained in X → interior X ⊆ X → interior X = X iff X open → boundary of X = X\ interior X • convex set x, x0 ∈ X and λ ∈ [0, 1] convex combination x00 ≡ λ x + (1 − λ)x0 convex set: x00 ∈ X for all x, x0 ∈ X and λ ∈ [0, 1] examples (→ class) if X, X 0 convex, then: X, X, X ∩ X 0, X + X 0 convex 5 Brower’s Fixed-Point Theorem Theorem 3 (Brouwer) Suppose that X ⊂ RL is nonempty, compact, and convex. If f : X → X is a continuous function from X to itself, then f (.) has a fixed point; i.e., there is an x ∈ X such that: x = f (x). in which way may the theorem fail to hold if: → → → → f (x) is not continuous X is not closed X is not bounded X is not convex c Ronald Wendner GE-Math-9 v2.1 6 Separation Theorems Consider p ∈ RL • hyperplane H(p, k) ≡ {x ∈ RL | p · x = k} → budget “line” (p = prices, x = consumption bundle, k = wealth) → isoprofit “line” (k = profit) Lemma 1 Consider K ⊂ RL nonempty, closed, convex, and z 6∈ K, z ∈ RL. Then ∃y ∈ K and p ∈ RL\{0}: p · z < k = p · y ≤ p · x for all x ∈ K. → exists a hyperplane separating z from K and bounding for K (K on one side of H) Theorem 4 (Bounding H Theorem, Minkowski) . Consider K ⊂ RL, convex, and z ∈ boundary K. Then there exists a bounding hyperplane through z. → there exists a bounding hyperplane for K → illustrate theorem by figure c Ronald Wendner GE-Math-10 v2.1 Theorem 5 (Separating H Theorem) . Consider A, B ⊂ RL, nonempty, convex, disjoint: A ∩ B = ∅. Then, there exists p ∈ RL\{0}: p·a ≥ p·b, for all a ∈ A, b ∈ B. → there exists a separating hyperplane for A and B → illustrate theorem by figure → results needed for welfare analysis (e.g., 2nd FUN theorem) c Ronald Wendner GE-Math-11 v2.1