note: this page is still in development.

these definitions and theorems also apply to courses on calculus, linear algebra, optimization, and real analysis.

linear algebra

basic notions of linear algebra

remark

vectors are always columns, but it is customary to write them as rows to save space.

definitionhomogeneous system

if , the system is said to be homogeneous and there is at least one solution, given by .

proposition

if , there need not exist such that .

definitionlinear combination

is a linear combination of if there exists such that .

example

consider three matrices , and such that .

observe that:

  1. rows of are linear combinations of rows of , and
  2. columns of are linear combinations of columns of .

these statements are in fact equivalent since if and only if .

elementary row operations and rref

definitionelementary row operations

simple operations that allow to transform a system of linear equations into an equivalent system.

there are three elementary row operations:

  1. switching rows,
  2. multiplying a row with a non-zero constant, and
  3. replacing a row with the sum of that row and another row.
theorem

let be an elementary matrix. then

  1. is invertible.
  2. there exist a number of elementary matrices such that .
definitionrow reduced echelon form

applying the three elementary row operations sequentially to any given matrix, we can find its row reduced echelon form (rref).

the rref of a matrix has the following properties:

  1. all the zero rows are at the bottom,
  2. the first non-zero entry of each non-zero row is 1,
  3. if the first non-zero entry of a row occurs on column and if the first non-zero entry of the next row occurs on column , then , and
  4. if the first non-zero entry of a row occurs on column , then all the earlier rows have zeros on column .
theorem

every matrix has a unique rref.

rank

definitionrank

the rank of is the number of non-zero rows in its rref .

definitionpivotal column

a pivotal column is one which contains the first non-zero entry of some row.

corollary

if is , then .

also, for a given :

  1. if and only if has no zero rows.
  2. if and only if all columns of are pivotal.
theorem

let be . then if and only if, , such that . namely, there is at least one solution to the system.

theorem

let be . then if and only if, , if , then . namely, there is at most one solution to the system.

corollary

let be . for each , has a unique solution in if and only if .

subspaces

definitionsubspace

a non-empty set is a subspace of if whenever and .

proposition

the intersection of members of an arbitrary family of subspaces is a subspace. the union of subspaces need not be a subspace.

definitionrange

the range of is the set

definitionnull space

the null space of is the set

proposition

for any , the following propositions hold:

  1. ,
  2. ,
  3. ,
  4. ,
  5. , and
  6. .
definitionlinear span

vectors span if, for every , there exist such that .

namely, each is a linear combination of .

definitionlinear independence

vectors in are linearly independent if, for all ,

a linearly independent set cannot contain .

proposition

for any given matrix, the following propositions hold:

  1. the non-zero rows of any matrix in rref are linearly independent,
  2. all subsets of a linearly independent set are also linearly independent, and
  3. a set in is linearly independent if and only if , where is formed by merging members of as columns.
lemma

is linearly dependent if and only if, for some , is a linear combination of .

definitionbasis

let be a subspace of . a set is a basis for if it spans and is linearly independent.

theorem

let be a subspace of with basis . if is linearly independent, then .

corollary

any vectors in are linearly dependent.

corollary

if and are two bases for , then .

definitiondimension

the dimension of a subspace is the number of elements in any basis for .

theoremfundamental of linear algebra i

let be . then

  1. ,
  2. , and
  3. .

orthogonality

definitionorthogonal complement

let be a subspace of . the orthogonal complement of is the set

lemma

for any subspace of , .

theoremfundamental of linear algebra ii

let be . then

  1. , and
  2. .
corollary

for any subspace of , .

inverse

theorem

let and be both . if , then .

theorem

let and be both . if , then .

definitioninvertible matrix

is invertible if there exists such that .

theorem

if and if , then .

remark

if and are invertible, then is invertible as well.

remark

if is invertible, then so is and .

real analysis

basic notions of topology

definitionmetric

let be a set. a function is a metric on if, for all , the following conditions hold:

  1. ,
  2. ,
  3. , and
  4. .
definitionmetric space

a metric space is a pair where is a set and is a metric on .

definitionopen ball

let be a metric space. for any and , is the open ball centered at with radius .

definitioninterior

let . is an interior point of if for some . the set of all interior points of is the interior of .

remark

definitionopen set

is open if .

remark

, and are open.

definitionclosure

let . is an closure point of if for every . the set of all closure points of is the closure of .

remark

definitionclosed set

is closed if .

remark

and are closed.

theorem

is closed if and only if is open.

definitionboundary

let . is an boundary point of if and for every . the set of all boundary points of is the boundary of .

remark
  • .
  • .
  • is closed if and only if .
  • is a closed set.
definitionmetric subspace

let be a metric space and . is a metric on and therefore is a metric space as well, usually referred to as a metric subspace of .

theorem

let be a metric subspace of .

  • is open in if and only if there exists an open set in such that .
  • is closed in if and only if there exists a closed set in such that .

sequences

definitionsequence

let be a metric space. a sequence in is a map .

definitionconvergent sequence

a sequence is said to be convergent if the following property holds: there exists some such that, for every , there exists some such that, for every , . such is called the limit of .

theorem

if exists, it is unique.

definitionsubsequence

let be a sequence. a subsequence of is a sequence obtained by deleting some (possibly none, possibly infinitely many) members of . namely, let be a non-decreasing sequence of integers. then is a subsequence of .

theorem

is convergent with a limit if and only if every subsequence of is convergent with limit .

theorem

let be a metric space and . if and only if there exists a sequence such that for all , and .

theorem

is closed if and only if every convergent sequence in converges to an element of .

definitionbounded sequence

let be a metric space. a sequence in is bounded if it is contained in a ball with finite radius, i.e., if for some and , for every .

theorembolzano–weierstrass

every bounded sequence in has a convergent subsequence.

definitioncauchy sequence

let be a metric space. a sequence in is a cauchy sequence if, for every , there exists a positive integer such that, whenever , .

theorem

if is convergent, then is cauchy. if is cauchy, then is bounded.

definitioncompleteness

let be a metric space. is complete if every cauchy sequence in converges to an element of .

theorem

let be complete and . is complete if and only if is a closed subset of .

definitionopen cover

let be a metric space and . a collection of open sets , , is an open cover for if .

definitionfinite subcover

let , , be an open cover for . if is a finite subset of and , then the collection , , is a finite subcover of .

definitioncompactness

is compact if every open cover of has a finite subcover.

theorem

a compact set is bounded and closed.

theorem

is compact if and only if every sequence in has a subsequence that converges to a point in .

theoremheine-borel

a subset of is compact if and only if it is closed and bounded.

continuity

definitioncontinuity

let and be metric spaces. a function is continuous at if, for every , there exists such that, whenever , .

definitionglobal continuity

is continuous if it is continuous at for every .

theorem

is continuous at if and only if for every sequence in , whenever , .

definitionimage

let and be metric spaces and . if , then .

definitioninverse image

let and be metric spaces and . if , then .

theorem

the following are equivalent:

  • is continuous.
  • is open if is open.
  • is closed if is closed.
definitionupper semicontinuous

let be a metric space. a function is upper semicontinuous if is closed for every .

definitionlower semicontinuous

let be a metric space. a function is lower semicontinuous if is closed for every .

theorem

is continuous if and only if is upper semicontinuous and lower semicontinuous.

definitionmaximizer / minimizer

let be a metric space and . a function has a maximizer if for some , for every . similarly, has a minimizer if for some , for every .

theoremweierstrass

let be compact and be continuous. then has a maximizer and a minimizer.

theorem

let be compact and be upper semicontinuous. then has a maximizer.

theorem

let be compact and be lower semicontinuous. then has a minimizer.

theorem

let be a metric space. if and are disjoint and closed subsets of , then there exists disjoint and open sets and such that and .

differentiable functions

definitiondifferentiability (single variable)

let , and . then is differentiable at if

exists.

namely, is differentiable at if there is some such that, for every , there exists some such that

where .

definitionpartial derivative

let , and . the th partial derivative of at is

if this limit exists.

definitiongradient

if exists for all , then the gradient of at is the vector

definitionjacobian

let , and . suppose that exists for all and , then the jacobian of at is the matrix

definitiondifferentiability (multivariate)

let . then is differentiable at if the following two conditions hold:

  1. exists for all and ,
  2. we have

the second condition holds if and only if, for every , there is some such that, whenever and , we have

theorem

suppose , and . suppose that, for each and , exists and that is continuous on . then is differentiable at .

theoremimplicit function

let . suppose that and exist and are continuous in an open containing . suppose that and that . then there exists and such that , for all and

theoremimplicit function (general)

suppose is differentiable in an open set containing and suppose that the entries of are continuous at each . let and be defined by as

and suppose that and is non-singular. then there exists and such that , for all and .

theoreminverse function

suppose that is differentiable in an open set containing . suppose that the partials of are continuous on . suppose that and is non-singular. then there exists some and some such that , for all , and is differentiable at with .

theoremweierstrass

let be compact and let be continuous. then there exists and in such that, for all , .

theorem

suppose attains its maximum or minimum on an open set at some . if exists, then .

theoremrolle's lemma

suppose is continuous at each and is differentiable at each . furthermore, suppose that . then there exists some such that .

theoremmean value

suppose that is continuous at each and that is differentiable at each . then there exists such that

theoremmean value (multivariate)

let be an open set in containing and . suppose that whenever . suppose that is differentiable at each . then there exists such that where .

definitionhessian matrix

let be open set in and let . the hessian of at is the matrix

theorem

suppose is open in , and is continuous at each for every and . then is symmetric for each .

theorem

suppose is an open set containing and such that . furthermore, suppose that contains for each . suppose that both and are differentiable on . then there is some such that

where .

definitionconvex function

let be convex. a function is convex if, for all and for all , we have . a function is strictly convex if, for all and for all such that , we have .

theorem

suppose that is open and convex and that is differentiable. then

  1. is convex if and only if for each .
  2. is strictly convex if and only if for each such that .

correspondences

definitioncorrespondence

a correspondence from to (both euclidean) is a map which associates with every a non-empty set . namely, a correspondence from to is a function from to .

definitionclosed-valued / open-valued / compact-valued

a correspondence is closed-valued is is a closed subset of for every . we similarly define correspondences which are open-valued, compact-valued, etcetera.

definitiongraph

the graph of is the set

proposition

is closed / open if is a closed / open subset of .

definitioninverses

let . the strong and weak inverses of under are defined respectively as

proposition
  • .
  • if is such that for every , then .
  • in general, .
definitionupper hemi-continuity

a correspondence is upper hemi-continuous (uhc) at if, whenever is an open subset of and , for some . is uhc if it is uhc at every .

proposition

is uhc if and only if the strong inverse of any open set in under is open in .

definitionlower hemi-continuity

a correspondence is lower hemi-continuous (lhc) at if, whenever is an open subset of and , for some . is lhc if it is lhc at every .

proposition

is lhc if and only if the weak inverse of any open set in under is open in .

definitioncontinuity

a correspondence is continuous if it is lhc and uhc.

theorem

let be a function and let be defined as for every . then the following are equivalent:

  1. is uhc.
  2. is lhc.
  3. is continuous.
proposition

if the function is upper / lower semicontinuous, the correspondence defined by need not be upper / lower hemicontinuous.

remark

closedness and uhc are not nested.

theorem

if is uhc and closed-valued, then it is also closed.

theorem

if is closed and is compact, then is uhc.

theorem

if is open, then it is lhc.

proposition

for every , every sequence converging to and every sequence such that for every , there exists a convergent subsequence of with limit in .

proposition

for every , every sequence converging to and every , there exists a sequence converging to such that for every .

theorem
  1. if satisfies proposition 1 (sequential characterization), then is uhc.
  2. if is uhc and compact-valued, then satisfies proposition 1.
theorem

satisfies proposition 2 if and only if satisfies lhc.

corollary

if is uhc and compact-valued and is compact, then is compact.

remark

a subset of a metric space is compact if and only if every sequence has a subsequence that converges to a point in .

theoremthe maximum

let and be euclidean sets. let and . for every , consider the problem

define . suppose that for every .

we can define a function as for some .

if is compact-valued, uhc and lhc and if is continuous, then

  1. is compact-valued and uhc, and
  2. is continuous.

convexity

definitionconvex set

a non-empty set is convex if, for every and every , .

definitionconvex combination

a convex combination of vectors is a vector of the form where are non-negative numbers which add up to 1.

theorem

is convex if and only if where

i.e., if and only if it contains all convex combinations of its elements.

proposition
  • arbitrary intersections of convex sets are convex.
  • is convex if and are convex.
  • for every scalar , is convex if is convex.
  • the closure and interior of a convex set are convex using the euclidean metric.
definitionconvex hull

the convex hull of a set is the "smallest" convex superset of . namely,

proposition

is convex and .

theorem

proposition

is convex if and only if .

theoremcarathéodory

let be non-empty. if , then can be written as a convex combination of no more than members of , i.e., there exists and with such that .

proposition
  • .
  • if is open, then is open.
  • the convex hull of a closed set in need not be closed. but if is compact, then is compact.
theoremshapley-folkman

let for every , and let . then there exists such that

  1. for every ,
  2. , and
  3. .
remark

if can be written as where , and, if , then there exist with such that .

scalars and do not need to add up to 1.

definitionhyperplane

fix and . the hyperplane formed by and is

where is called the normal vector of .

theoremminkowski

let be non-empty, convex and closed, and let . there exists and such that for every .

theorem

suppose that is non-empty and convex, and that is a sequence in . if , then there exists such that, for every , .

theorem

suppose that is non-empty and convex. if , then there exists such that for every .

theorem

suppose that is a non-empty and convex set, and let . then there exists such that for every .

theorem

let and be disjoint and convex subsets of . there exists such that for every .

definitionconvex function

let be convex. a function is convex if , , . similarly, is said to be strictly convex if , , .

definitionconcave function

let be convex. a function is concave if , , . similarly, is said to be strictly concave if , , .

definitionsubgradient

suppose is convex and . a vector is a subgradient for at if

for all .

definitionsupergradient

suppose is convex and . a vector is a supergradient for at if

for all .

theorem

suppose that is convex and . if has a subgradient at every , then is convex.

theorem

suppose that is convex, is convex, and . then is continuous at .

theorem

suppose that is convex, is convex, and . then has a subgradient at .

theorem

suppose that is convex, is convex, , and is differentiable at . then is the unique subgradient of at .

theorem

suppose that is convex and . if is a subgradient of at and is a subgradient of at , then .

theorem

suppose that is open and convex, and is differentiable and convex. then .

support functions

maximization

definitionbarrier cone

for any non-empty , the barrier cone of is defined as

definitionsupport function

for any non-empty , the support function of is a function such that

remark

if and , then as well. hence is a cone and, in particular, .

proposition

need not be convex even if is.

theorem

let be non-empty.

  • if solves and solves , then . this is known as monotonicity of solutions.
  • for every and . namely, homogeneity of degree 1.
  • if is convex, then is convex.
  • suppose is closed and convex, and . solves if and only if is a subgradient of at .
  • suppose is closed and convex, and is differentiable at . then solves if and only if . also known as hotelling's lemma in profit maximization.

minimization

definitionbarrier cone

for any non-empty , the barrier cone of is defined as

definitionsupport function

for any non-empty , the support function of is a function such that

remark

note that . furthermore, .

theorem

let be non-empty.

  • if solves and solves , then . this is known as monotonicity of solutions.
  • for every and . namely, homogeneity of degree 1.
  • if is convex, then is concave.
  • suppose is closed and convex, and . solves if and only if is a supergradient of at .
  • suppose is closed and convex, and is differentiable at . then solves if and only if . also known as shephard's lemma in cost minimization.

nonlinear programming

minimization problems

definitionnonlinear programming problem

let , , . . . , be differentiable functions. the nlp problem is

kuhn-tucker necessity conditions

definitionkuhn-tucker pair for nlp

a pair is a kuhn-tucker pair for nlp if

  1. for all ,
  2. for all ,
  3. , and
  4. for all .

we will say that satisfies the kuhn-tucker conditions for nlp if there exists some such that is a kuhn-tucker pair for nlp.

definitionactive constraints

for any , let . hence is the set of active constraints at .

lemma

suppose that is a kuhn-tucker pair for nlp. if , then . if , then .

lemma

suppose that

  1. for all , and
  2. is a collection of non-negative numbers satisfying .

if for all and for all , then is a kuhn-tucker pair for nlp.

definitionlinear programming problem

for and , consider the following lp problem:

lp is a special case of nlp where and .

definitionkuhn-tucker pair for lp

a pair is a kuhn-tucker pair for lp if

  1. for all ,
  2. for all ,
  3. , and
  4. for all .
theorem

suppose that and for all . then the set

is a non-empty, closed and convex cone.

theorem

if solves the lp problem, then there exists some such that is a kuhn-tucker pair for lp.

definitiongeneral constraint qualification

a solution to the nlp problem satisfies the general constraint qualification (gcq) condition if it solves the following lp problem:

theorem

if solves the nlp problem and satisfies the gcq condition, then it satisfies the kt conditions for nlp.

definitioncottle constraint qualification

a solution to the nlp problem satisfies the cottle constraint qualification (ccq) condition if there exists some such that for all .

theorem

suppose solves the nlp problem. if satisfies the ccq condition, then it satisfies the gcq condition.

definitionlinear independence constraint qualification

a solution to the nlp problem satisfies the linear independence constraint qualification (licq) condition is the collection is linearly independent.

theorem

suppose solves the nlp problem. if satisfies the licq condition, then it satisfies the ccq condition.

corollary

suppose solves the nlp problem. if satisfies the licq condition, then it satisfies the kt conditions for nlp.

kuhn-tucker sufficiency conditions

theorem

let be differentiable. then

  1. is convex if and only if , and
  2. is concave if and only if ,

for every and .

theorem

suppose satisfies the kt conditions for nlp. if is convex and each is concave, then solves the nlp problem.

definitionquasi-convex / quasi-concave

a function is quasi-convex / quasi-concave if the lower-contour set / the upper-contour set is convex for every .

theorem

let be differentiable. then

  1. is quasi-concave if and only if , and
  2. is quasi-convex if and only if ,

for every and .

definitionpseudo-convex / pseudo-concave

a differentiable function is pseudo-convex if, for every and , implies . a differentiable function is pseudo-concave if, for every and , implies .

theorem

suppose is a kt pair for nlp. if is pseudo-convex and every is quasi-concave, then solves the nlp problem.

theorem

suppose solves the nlp. if each is concave and if there exists such that for all , then satisfies the ccq condition.

remark

for a nlp problem with concave constraints, the existence of an such that for all is called the slater constraint qualification condition.

comparative statics

remark

consider the problem subject to for all , where and all functions are differentiable. call this problem . let be the value of .

definitionregular solution

a vector is a regular solution to if there exist and such that

  1. satisfies the kt conditions,
  2. solves the problem for every and ,
  3. for all , and
  4. is differentiable at .
theorem

suppose that is a regular solution to with associated multiplier vector . then

remark

suppose that is a regular solution to with associated multiplier vector . if , then . if , , and , then .

maximization problems

remark

consider the problem subject to for all where all function are differentiable. let .

theorem

if solves the maximization problem above and is a linearly independent set, then there exists such that and for all .

theorem

suppose that

  1. for all ,
  2. there exists a set of positive numbers such that ,
  3. is pseudo-concave, and
  4. each is quasi-convex.

then solves the maximization problem above.

consumer theory

utility maximization

consider the following utility maximization problem

the kt conditions are

remark

is interpreted as the marginal utility of income. call the solutions the demand functions and the indirect utility function.

definitionindirect utility function properties
  1. zero degree homogeneity.
  2. is non-decreasing in , and non-increasing in .
  3. is quasi-convex in .

expenditure minimization

consider the following expenditure minimization problem

remark

let be the solution and . the expenditure function is mathematically identical to the cost function.

definitionexpenditure function properties
  1. is 1-homogeneous.
  2. is non-decreasing.
  3. is concave.
  4. by shephard's lemma, .
theoremroy's identity
definitionrelationships between utility maximization and expenditure minimization
theoremslutsky equation

preferences and utility representation

binary relations

definitionbinary relation

let be a set and . a binary relation on is any subset of . we will typically denote binary relations by , , or , , .

let be a binary relation. instead of , we write . similarly, instead of , we write .

definitionstrict preference

let be a binary relation. suppose we would like to mean is strictly better that . here are some properties which we might like to impose on .

  • irreflexivity: for no .
  • asymmetry: for any , if , then .
  • acyclicity: for every and for every , if for every , then .
  • transitivity: if and , then .
  • negative transitivity: if and , then either or .
  • connectedness: if , then either or .
remark
  • is a (strict) linear order if it is connected, asymmetric and negatively transitive.
  • is a (strict) weak order if it is asymmetric and negatively transitive.
  • is a (strict) partial order if it is irreflexive and transitive.
definitionweak preference

let be a binary relation. suppose we would like to mean is at least as good as . here are some properties which we might like to impose on .

  • reflexivity: for all .
  • completeness: for all , or .
  • transitivity: if and , then .
  • antisymmetry: if and , then .
remark
  • is a (weak) linear order if it is antisymmetric, complete and transitive.
  • is a (weak) weak order if it is complete and transitive.
  • is a (weak) partial order if it is reflexive and transitive.
definitionindifference

let be a binary relation. suppose we would like to mean that and are equally good. here are some properties which we might like to impose on .

  • reflexivity: for all .
  • symmetry: if , then .
  • transitivity: if and , then .
remark

call an equivalence class if it is reflexive, symmetric and transitive.

proposition

suppose that is a (strict) weak order and that binary relations and are defined, using , as follows

then is a (weak) weak order and is an equivalence class. if, in particular, is a (strict) linear order, then is a (weak) linear order.

proposition

suppose that is a (weak) weak order and that binary relations and are defined, using , as follows

then is a (strict) weak order and is an equivalence class. if, in particular, is a (weak) linear order, then is a (strict) linear order.

preference maximization

definitionmenu

let . any member of is a menu.

remark

for any binary relation and any menu , let

in general, these sets may be empty and they need not to coincide.

proposition

suppose is finite. is acyclic if and only if for all .

proposition

suppose is a (weak) weak order and define using as follows:

then for all . if is a (weak) linear order, then is a singleton.

theorem

let be a binary relation on a countable set . admits a utility representation if and only if is a weak order.

remark

in general, there exist weak orders (even linear orders) which do not have utility representations.

theorem

let be a linear order on an arbitrary set . admits a utility representation if and only if contains a countable -dense subset.

theorem

a binary relation on an arbitrary set admits a utility representation if and only if is a weak order on and contains a countable -dense subset.

lemma

given , say that the pair is a gap if but there exits no such that and . let is a gap for some , is a gap for some and .

suppose is a linear order on an arbitrary set . is countable if one of the following two conditions holds:

  1. contains a countable -dense subset.
  2. admits a utility representation.

a topological approach

corollary

let be a complete and transitive binary relation (i.e., a weak order) on . define, as usual, if and only if and . recall that such is asymmetric and negatively transitive. endow with the euclidean metric. we say that admits a continuous utility representation if there exists a continuous function such that if and only if .

theorem

let be convex and be a binary relation on . is a continuous weak order if and only if it admits a continuous utility representation.

proposition

let . if is a strictly monotone weak order and and are scalars, then

theorem

suppose is a strictly monotone and continuous weak order on . then admits a continuous utility representation.