Вы находитесь на странице: 1из 49

Mathematical Foundations -1- Mappings

Mappings

© John Riley October 11, 2013


Mathematical Foundations -2- Mappings

Mappings: transformations, correspondences, functions

Linear Transformation n
 m

 a11 . . a1n 
 . . . . 
A  [aij ]   
 . . . . 
 
 am1 . . amn 

An m  n dimensional array. We call this array a matrix.

Special matrices:

 a11 
 . 
 
1 n matrix [a11 ,..., a1n ] a row vector m 1 matrix  .  a column vector
 
 . 
 am1 

It is often useful to think of an m  n matrix as either a row of n column vectors or a column of m row
vectors

© John Riley October 11, 2013


Mathematical Foundations -3- Mappings

Linear transformation
Linear combination of column vectors

 a11   a12   a1n   a11   a12   a1n 


 .   .   .   .   .   . 
           
 .  ,  .  ,…,  .  y =  .  x1 +  .  x2 +…+  .  xn
           
.
    .  .    .   .  . 
 am1   am 2   amn   am1   am 2   amn 

We write
n
y  Ax where yi   aij x j
j 1

The inner product of the i-th row and the vector x

© John Riley October 11, 2013


Mathematical Foundations -4- Mappings

Linear Transformations 2
 2

For any pair of column vectors

 a11   a12 
a  , a 
 21   22 
Let y be a linear combination of these two column vectors.

 y1   a11  x1  a12  x2  a11 x1  a12 x2 


 y   a   a   a x  a x  (*)
 2   21   22   21 1 22 2 
Graphically,
scale up the first vector by x1 ,

scale up the second by x2

and then add the two resulting vectors as shown.

© John Riley October 11, 2013


Mathematical Foundations -5- Mappings

For any set X  2


the linear combinations transform all the elements of X into a set Y  2

Linear transformation
We write the linear transformation as follows.

 y  a a12   x1 
y  Ax or  1    11
 y2   a21 a22   x2 

x 
Then yi is the inner product of the row vector  ai1 ai 2  and the column vector  1 
 x2 

 y1   a11 a12   x1   a11 x1  a12 x2 


 y   a   x   a x  a x 
 2   21 22   2   21 1 22 2 
a

© John Riley October 11, 2013


Mathematical Foundations -6- Mappings

Compound linear transformations


Consider the linear transformations y  Ax and z  By

 y   y  a x  a x 
zi  bi1 bi 2   1  ,  1    11 1 12 2  .
 y2   y2   a21 x1  a22 x2 
Then

a x  a x  x 
zi  bi1 bi 2   11 1 12 2   (bi1a11  bi 2 a21 ) x1  (bi1a12  bi 2 a22 ) x2  bi1a11  bi 2 a21 bi1a12  bi 2 a22   1 
 a21 x1  a22 x2   x2 
Therefore

 z1   b11a11  b12 a21 b11a12  b12 a22   x1 


 z   b a  b a   x  .
 2   21 11 22 21 21 12 22 22   2 
b a b a

Thus the compound mapping is the linear transformation

 a1 j  2
z  Cx where cij  bi1 bi 2      aik bkj .
 a2 j  k 1
Inner product of i-th row of B and j-th column of A

We write the compound mapping as C = BA .

© John Riley October 11, 2013


Mathematical Foundations -7- Mappings

Linear Transformations n
 n

Matrixes are square

Identity matrix n= 2

1 0 
I 
0 1 
Note that, by the rules of matrix multiplication, the identity matrix maps x onto itself.
Ix  x
Transpose of a square matrix
Flip rows and columns
The ith row becomes the ith colum

a a12 
A   11
a22 
.
 a21
Transpose

a a21 
A   11
 a12 a22 

© John Riley October 11, 2013


Mathematical Foundations -8- Mappings

Determinant of a matrix
n=2

a a12 
A   11
 a21 a22 

Delete a row and column and you get a number mij .

The matrix of these numbers is

 m11 m12   a22 a21 


m   a a11 
 21 m22   12

Convert this into a matrix of cofactors cij  where cij  (1)i j mij

 c11 c12   (1)2 m11 (1)3 m12   a22 a21 


c  
a11 
 21 22  (1) m21 (1) m22   a12
3 4
c

Then the determinant of A is defined as follows:


n n
A   aij cij   aij cij
j 1 i 1

It is a lovely theorem that it does not matter which row or column of A that we pick.

© John Riley October 11, 2013


Mathematical Foundations -9- Mappings

Example:

6 2 
A ,
 8 3 

 m11 m12   3 8 
m 
 21 m22   2 6 

 c11 c12   (1)2 m11 (1)3 m12   3 8


C   
   2 6 
3 4
 21 22  
c c ( 1) m21 ( 1) m22 

n n
A   aij cij   aij cij  2
j 1 i 1

© John Riley October 11, 2013


Mathematical Foundations -10- Mappings

n=2

a a12 
A   11
 a21 a22 

Eliminating a row and column yields a number so the minors are easily computed.
The matrix of minors is

 m11 m12   a22 a21 


m   a a11 
 21 m22   12
The matrix of cofactors is

c c  a a21 
C   11 12    22
 a11 
.
c21 c22   a12
Try any row or column and you will be able to confirm that

A  a11a22  a21a12

© John Riley October 11, 2013


Mathematical Foundations -11- Mappings

n=3
Consider the ij-th minor matrix created by deleting the i-th row and j-th column of A. This is a 2  2
matrix. Let mij be the determinant of this minor matrix.

Define the cofactor cij  (1)mij .

Then the determinant of A is defined as follows:


n n
A   aij cij   aij cij
j 1 i 1

Again it does not matter which row or column of A that we pick.

A second feature of the matrix of cofactors is that if you compute the sum product of the k-th row
where k  i and the ith row of cofactors, then the product is zero. The same is true for columns.
n n

j 1
akj cij  0, k  i , 
i 1
aik cij  0, k  j

© John Riley October 11, 2013


Mathematical Foundations -12- Mappings

Inverse of a matrix
C = matrix of cofactors

c c  a a21 
C   11 12    22
 a11 
.
c21 c22   a12
Then the transposed matrix of cofactors is

c c21    a22 a11 


C   11 
c12 c22   a21  a11 

Now consider the product of A and C

a a  c c21   a11 a12   a22 a11   A 0


AC   11 12   11   
 a21 a22  c12 c22   a21 a22   a21 a11   0 A

c c21   a11 a12    a22 a11   a11 a12   A 0


CA   11   
c12 c22   a21 a22   a21 a11   a21 a22   0 A

1 1   a22 a11 
Finally, if A  0 define A 1  C 
A  a21  a11 
.
A

© John Riley October 11, 2013


Mathematical Foundations -13- Mappings

1 1   a22 a11 
A 1  C 
A  a21  a11 
.
A

1 1   a22 a11 
A 1  C 
A  a21  a11 
.
A

Then
1 1
A 1A  CA  I and AA 1  AC  I
A A

Mapping
y  Ax

Inverse mapping

A1 y  A1Ax  Ix  x

1
We call A 1  C the inverse matrix.
A

© John Riley October 11, 2013


Mathematical Foundations -14- Mappings

Linear independence of the columns of an invertible matrix


Proof by contradiction.

 a11 . . a1n   x1   a11   a1n 


 . . . .  .   .   . 
Ax         x1  ...    xn
 . . . .  .   .   . 
      
 an1 . . ann   xn   an1   ann 

If the columns are linearly dependent then, for some x  0 , Ax  0 .

Since A is invertible, x  A1Ax  A1 0  0 . A contradiction.

Exercise:

0 1 1 
A  1 0 1 
 
1 1 0 

(a) Compute the nine minor determinants of A and hence compute A .


(b) Use your results to obtain C and hence C and hence obtain the inverse matrix.

© John Riley October 11, 2013


Mathematical Foundations -15- Mappings

Computing the inverse matrix


Example:

1 1 1 
A  1 2 0  ,
 
0 1 3

2 0 1 0 1 2
m11  , m12  , m13 
1 3 0 3 0 1

1 1 1 1 1 1
m21  , m22  , m23 
1 3 0 3 0 1

1 1 1 1 1 1
m31  , m32  , m33 
2 0 1 0 1 2

 6 3 1  6 3 1   6 2 2 
[mij ]   2 3 1 C  [cij ]   2 3 1 C   3 3 1 
     
 2 1 1  2 1 1   1 1 1 

A  sumproduct ([ith row of [aij ],[ith row of [cij ])  4

© John Riley October 11, 2013


Mathematical Foundations -16- Mappings

Inverse of A

1 1 1   6 2 2 
A  1 2 0  C   3 3 1 
1 1
A 1 
  A 4 
0 1 3  1 1 1 

As may readily be checked,

A-1 A = I

© John Riley October 11, 2013


Mathematical Foundations -17- Mappings

Function f ( x)  f ( x1 ,..., xn )

single valued mapping from S  n


Increasing Function
The function f ( x) is increasing over the set X if for any x and x  X

x  x  f ( x)  f ( x)

Strictly increasing function


The function f ( x) is increasing over the set X if for any x and x  X

x  x  f ( x)  f ( x)

© John Riley October 11, 2013


Mathematical Foundations -18- Mappings

Class Exercise:

(a) If a function f :   is differentiable and strictly increasing at x 0 , does it follow that the slope
of the function is strictly positive at x 0 ?

(b) If the function f : n


 is increasing at x 0 in all of its arguments is it an increasing function?

**

© John Riley October 11, 2013


Mathematical Foundations -19- Mappings

Class Exercise:

(a) If a function f :   is differentiable and strictly increasing at x 0 , does it follow that the slope
of the function is strictly positive at x 0 ?

(b) If the function f : n


 is increasing at x 0 in all of its arguments is it an increasing function?

(a) Consider the following function

f ( x)  x 3 , x 0  0

© John Riley October 11, 2013


Mathematical Foundations -20- Mappings

Class Exercise:

(a) If a function f :   is differentiable and strictly increasing at x 0 , does it follow that the slope
of the function is strictly positive at x 0 ?

(b) If the function f : n


 is increasing at x 0 in all of its arguments is it an increasing function?

(a) Consider the following function

f ( x)  x 3 , x 0  0

(b) Consider the following function f : 2


 

f ( x)  x14  4 x12 x2 2  x2 4

Suppose x0  (0,0) . Both f ( x1 ,0)  x14 and f (0, x2 )  x2 4 are strictly increasing functions on 2
 .
However if ( x1 , x2 )  ( z, z ) then f ( z, z )  2 z 4 . Thus the function f is definitely not an increasing
function.

© John Riley October 11, 2013


Mathematical Foundations -21- Mappings

© John Riley October 11, 2013


Mathematical Foundations -22- Mappings

Special functions

Linear Function l ( x)  a  x

Quadratic form q( x)  xAx where A is a square matrix

WOLOG we may assume that A is symmetric aij  a ji

a a12   x1  a x  a x 
y  [ x1 x2 ]  11  [ x1 x2 ]  11 1 12 2   a11 x12  (a21  a21 ) x1 x2  a22 x2 2  a11 x12  2a12 x1 x2  a22 x2 2
 a21 a22   x2   a21 x1  a22 x2 

Quadratic Function f ( x)  k  a  x  xAx

Equivalently, f ( x)  k  ax  xAx

© John Riley October 11, 2013


Mathematical Foundations -23- Mappings

Special functions

Linear Function l ( x)  a  x

Quadratic form q( x)  xAx where A is a square matrix

WOLOG we may assume that A is symmetric aij  a ji

a a12   x1  a x  a x 
y  [ x1 x2 ]  11  [ x1 x2 ]  11 1 12 2   a11 x12  (a21  a21 ) x1 x2  a22 x2 2  a11 x12  2a12 x1 x2  a22 x2 2
 a21 a22   x2   a21 x1  a22 x2 

Quadratic Function f ( x)  k  a  x  xAx

Equivalently, f ( x)  k  ax  xAx

© John Riley October 11, 2013


Mathematical Foundations -24- Mappings

Negative definite matrix


A symmetric matrix A with the property that q( x)  x ' Ax is strictly negative for all x  0.

Consider n=2 q( x)  xAx  a11 x12  2a12 x1 x2  a22 x2 2

**

© John Riley October 11, 2013


Mathematical Foundations -25- Mappings

Negative definite matrix


A symmetric matrix A with the property that q( x)  x ' Ax is strictly negative for all x  0.

Consider n=2 q( x)  xAx  a11 x12  2a12 x1 x2  a22 x2 2

Note first that a11  0 (set x2  0 )

© John Riley October 11, 2013


Mathematical Foundations -26- Mappings

Negative definite matrix


A symmetric matrix A with the property that q( x)  x ' Ax is strictly negative for all x  0.

Consider n=2 q( x)  xAx  a11 x12  2a12 x1 x2  a22 x2 2

Note first that a11  0 (set x2  0 )

Complete the square ( x1   )2  x12  2 x1   2

a12 a12 2 2 a12 2 2


q( x)  a11 ( x  2
1
2
x1 x2  ( ) x2 )  (a22  ) x2
a11 a11 a11

a12 1
 a11 ( x1  x2 )2  (a11a22  a12 2 ) x2 2
a11 a11

Then q( x)  0 for all x  0 if and only if a11  0 and a11a22  a12 2  0

© John Riley October 11, 2013


Mathematical Foundations -27- Mappings

Negative semi-definite matrix


A symmetric matrix A with the property that q( x)  x ' Ax is negative for all x .

n=2: q( x)  0 for all x if and only if a11  0, a22  0 and a11a22  a12 2  0

© John Riley October 11, 2013


Mathematical Foundations -28- Mappings

Negative semi-definite matrix


A symmetric matrix A with the property that q( x)  x ' Ax is negative for all x .

n=2: q( x)  0 for all x if and only if a11  0, a22  0 and a11a22  a12 2  0

Case (i) a11  0


a 1
q( x)  a11 ( x1  12 x2 ) 2  (a11a22  a12 2 ) x2 2
a11 a11

Then a necessary and sufficient condition is a11  0 and a11a22  a12 2  0

© John Riley October 11, 2013


Mathematical Foundations -29- Mappings

Negative semi-definite matrix


A symmetric matrix A with the property that q( x)  x ' Ax is negative for all x .

n=2: q( x)  0 for all x if and only if a11  0, a22  0 and a11a22  a12 2  0

Case (i) a11  0


a 1
q( x)  a11 ( x1  12 x2 ) 2  (a11a22  a12 2 ) x2 2
a11 a11

Then a necessary and sufficient condition is a11  0 and a11a22  a12 2  0

Case (ii) a11  0


Exercise: Complete the proof.

© John Riley October 11, 2013


Mathematical Foundations -30- Mappings

Limit of a function
The function f : S  n
 has a limit L at x0  S if, for any   0 , there exists a deleted

neighborhood N D ( x0 ,  ) such that if x  N D ( x0 ,  ) S then f ( x)  L   .

1  x 2 , x  1
Exercise: Suppose f ( x)   .
 2  x 2
, x  1

(a) Does the function have a limit at x0  1 if S  [0,2] ?


(b) What if S  [0,1] ?

Continuous function
Let f be a function defined on S  n
and suppose x0  S . Then f is continuous at x 0 if it has a

limit at x 0 equal to f ( x0 ) .

© John Riley October 11, 2013


Mathematical Foundations -31- Mappings

Example 1: A function f :  with no limit at x0  2


4

Consider some small  .


2
For all x in the interval (2   ,2) , f ( x) is close to 2.

For all x in the interval (2,2   ) , f ( x) is close to 4.

Thus the function does not approach a limit L as

x approaches x0  2 .

© John Riley October 11, 2013


Mathematical Foundations -32- Mappings

Example 2: A function f :[0,2)  (2,4]  with a limit point at x0  2

Note that even though the function is not defined 4

at x0  2 , for all x close to 2, f ( x) is close to 4.

Thus the limit of the function is L  4 .

KEY POINT: The function itself need not be defined at x 0 to have a limit point at x 0 .

© John Riley October 11, 2013


Mathematical Foundations -33- Mappings

Example 3: A discontinuous function f :  with a limit point

a
Arguing as in Example 2, the function has a

limit L  2 at x0  2 . Also f ( x0 )  a . 2

Therefore the function is discontinuous at x 0 unless a  2 .

© John Riley October 11, 2013


Mathematical Foundations -34- Mappings

Rules of limits: f : S  n
 , f :S  n

Rule 1: Limit of sum = sum of limits

Rule 2: Limit of product = product of limits

We will prove Rule 1 this assuming that L1 , L2  0 .

We know that for any 1 ,  2 there exist deleted delta neighborhoods N D ( x0 ,1 ), N D ( x0 ,  2 ) such that

L1  1  f ( x)  L1  1 , x  N D ( x0 ,1 )  S , L2   2  g ( x)  L2   2 , x  N D ( x0 ,  2 )  S

We wish to prove that for any  there exists a deleted delta neighborhood N D ( x0 ,  ) such that

L1  L2    f ( x)  g ( x)  L1  L2   , x  N D ( x 0 ,  )  S

© John Riley October 11, 2013


Mathematical Foundations -35- Mappings

We will prove Rule 1 this assuming that L1 , L2  0 .

We know that for any 1 ,  2 there exist deleted delta neighborhoods N D ( x0 ,1 ), N D ( x0 ,  2 ) such that

L1  1  f ( x)  L1  1 , x  N D ( x0 ,1 )  S , L2   2  g ( x)  L2   2 , x  N D ( x0 ,  2 )  S

We wish to prove that for any  there exists a deleted delta neighborhood N D ( x0 ,  ) such that

L1  L2    f ( x)  g ( x)  L1  L2   , x  N D ( x 0 ,  )  S

© John Riley October 11, 2013


Mathematical Foundations -36- Mappings

Proof:

We will prove Rule 1

L1  1  f ( x)  L1  1 , x  N D ( x0 ,1 )  S , L2   2  g ( x)  L2   2 , x  N D ( x0 ,  2 )  S

Then

( L1  1 )  ( L2   2 )  f ( x)  g ( x)  ( L1  1 )  ( L2   2 )  L1  L2  (1   2 ) ,

for all x  N D ( x0 ,  )  S where   Min{1 , 2 }

© John Riley October 11, 2013


Mathematical Foundations -37- Mappings

Proof:

We will prove Rule 1

L1  1  f ( x)  L1  1 , x  N D ( x0 ,1 )  S , L2   2  g ( x)  L2   2 , x  N D ( x0 ,  2 )  S

Then

( L1  1 )  ( L2   2 )  f ( x)  g ( x)  ( L1  1 )  ( L2   2 )  L1  L2  (1   2 ) ,

for all x  N D ( x0 ,  )  S where   Min{1 , 2 }

That is,

L1  L2  (1   2 )  f ( x)  g ( x)  L1  L2  (1   2 )  L1  L2  (1   2 ) ,

for all x  N D ( x0 ,  )  S where   Min{1 , 2 }

© John Riley October 11, 2013


Mathematical Foundations -38- Mappings

Proof:

We will prove Rule 1

L1  1  f ( x)  L1  1 , x  N D ( x0 ,1 )  S , L2   2  g ( x)  L2   2 , x  N D ( x0 ,  2 )  S

Then

( L1  1 )  ( L2   2 )  f ( x)  g ( x)  ( L1  1 )  ( L2   2 )  L1  L2  (1   2 ) ,

for all x  N D ( x0 ,  )  S where   Min{1 , 2 }

That is,

L1  L2  (1   2 )  f ( x)  g ( x)  L1  L2  (1   2 )  L1  L2  (1   2 ) ,

for all x  N D ( x0 ,  )  S where   Min{1 , 2 }

Choose 1  12  and  2  12  .

Then L1  L2    f ( x)  g ( x)  L1  L2  

QED

© John Riley October 11, 2013


Mathematical Foundations -39- Mappings

Properties of functions on a set

Some properties of functions must be defined point-wise. When this is the case and the property P
holds for the function f at every point in a set S, the function f is said to have the property P on S.

Example 1: f ( x)  x1 x2 is continuous on 2
 (the set of all positive elements of 2
) because it is
2
continuous at each point in  .

Example 2: f ( x)  ln x1 x2 is continuously differentiable on 2


 (the set of strictly positive elements of
2 2
) because the partial derivatives exist and are continuous at each point in  .

Note: If the partial derivatives of a function f : S  n


 are continuous we write f  C1 .

© John Riley October 11, 2013


Mathematical Foundations -40- Mappings

However if a property need not be defined point-wise, it is best defined on a set directly.

An increasing function can defined point-wise on a set or directly on the set. To see why
mathematicians take the latter approach consider the following example.
Example 1:

 x, x  [0,1]
S  [0,1] [2,3] , f ( x)   .
 x  5, x  [2,3]

Note that the function is strictly increasing at each point in S however the function is not strictly
increasing on S.

© John Riley October 11, 2013


Mathematical Foundations -41- Mappings

Key foundational theorem for economics

Extreme Value Theorem


If the function f is continuous over a compact (i.e. closed and bounded) set S  n
, then f attains its
maximizing and minimizing value at some point in S.

Class exercise:

To see that there may be no maximizing value if S is not compact and f is not continuous,

consider the following examples.

(i) f ( x)  x, S  [0,2)
(ii) f ( x)  x, S   (the set of positive real numbers)

 x, x  1
(iii) f ( x)   , S  [0,1]
 0, x  1

© John Riley October 11, 2013


Mathematical Foundations -42- Mappings

Exercise: Suppose that f : S  is continuous and for every x0  S there exists a   0 such that for
f ( x)  f ( x 0 )
all x  N ( x ,  ) ,
0
 0 . If S is a closed interval, show that f is strictly increasing on S .
x  x0

HINT: If not, then for some x 0 and x1  x0 , (a) f ( x1 )  f ( x0 ) or (b) f ( x1 )  f ( x0 ) . First consider

case (a) and appeal to the Extreme Value theorem for the interval [ x0 , x1 ] . Use your conclusion to
obtain a contradiction.

© John Riley October 11, 2013


Mathematical Foundations -43- Mappings

Correspondence:
Multi valued mapping

Each point x  X  n
is mapped into a subset y( x)  m

Example 1: Supply correspondence

 6q, q  [0,10]

C (q)  60  12(q  10), q  (10,20]
 , q  (20, )

The firm solves the following maximization problem. q

Max{ pq  C (q) | q   }
q

The set of quantities that solve this problem is written as follows.


y( p)  arg Max{ pq  C (q) | q   }.
q

Note that y ( p) is not single valued for all prices.

© John Riley October 11, 2013


Mathematical Foundations -44- Mappings

Example 2: Demand correspondence

A consumer enjoys ice cream but really loves raspberry ripple.

Each scoop of raspberry ripple (commodity 2) is as good as two scoops of vanilla (commodity 3)

U ( x1 , x2 , x3 )  ln x1  ln(2 x2  x3 )

The marginal utilities of the three commodities are

U 1 U 2 U 1
 ,  , 
x1 x1 x2 2 x2  x3 x3 2 x2  x3

Hence the marginal utilities per dollar are

1 U 1 1 U 2 1 1 U 1
 ,  ( ),  .
p1 x1 p1 x1 p2 x2 p2 2 x2  x3 p3 x3 p3 x3

© John Riley October 11, 2013


Mathematical Foundations -45- Mappings

Case (i) p  P r  { p | p2  2 p3}

Then the marginal utility per dollar is higher for scoops of raspberry so x3* ( p)  0

Equating marginal utilities per dollar

1 1 U 1 U 2 1
   ( ).
p1 x1 p1 x1 p2 x3 p2 2 x2

Then spending on each of the other two commodities is half total spending so

I I
x* ( p )  ( , ,0) .
2 p1 2 p2

Case (ii) p  Pv  { p | p2  2 p3}

By an almost identical argument

I I
x* ( p )  ( ,0, ).
2 p1 2 p3

© John Riley October 11, 2013


Mathematical Foundations -46- Mappings

Case (iii) p  Pb  3
 \ Pr Pv

1 U 1 U

p2 x2 p3 x3

Consumer is indifferent between

I I I I
x 0 ( p)  ( , ,0) , x1 ( p)  ( ,0, ) and any convex combination.
2 p1 2 p2 2 p1 2 p3

Thus for p  Pb there is a set of maximizing consumption bundles

x* ( p)  (1   ) x0 ( p)   x1 ( p), 0    1.

© John Riley October 11, 2013


Mathematical Foundations -47- Mappings

Correspondence

A non-empty set valued mapping X ( ) of each value of  into a subset of n


 .

Upper hemi-continuous correspondence


The set valued mapping X ( ) is upper hemi-continuous at  0 if for any open neighborhood, V of

X ( 0 ) there exists a   neighborhood of  0 , N ( 0 ,  ) , such that X ( )  V ,for all   N ( 0 , )

Exercise: Explain why the following correspondence is upper hemi-continuous on S  [0,4] .

[2 ,3 ], 0    2
X ( )  
 [ ,4 ], 2    4
Lower hemi-continuous correspondence
The set valued mapping X ( ) is lower hemi-continuous at  0 if for any open set V that intersects

X ( 0 ) , there exists a   neighborhood of  0 , N ( 0 ,  ) , such that X ( ) intersects V for all

  N ( 0 , )

© John Riley October 11, 2013


Mathematical Foundations -48- Mappings

Consider the examples below. In both cases X ( ) is an interval. Note that the left hand mapping is

continuous from the right  0 and the second is continuous from the left but neither is continuous.
However the mapping on the left satisfies the conditions for upper hemi-continuity and the one on the
right does not.

Upper hemi-continuity Lower hemi-continuity

The right-hand diagram does, however, satisfy the conditions for lower hemi-continuity while the one
on the left does not.

© John Riley October 11, 2013


Mathematical Foundations -49- Mappings

Continuous correspondence

A mapping is continuous if it is both upper and lower hemi-continuous.

Exercise: Supply correspondence

Explain why the supply correspondence of Example 1 is upper hemi-continuous.

Exercise: Demand correspondence

For the demand correspondence of Example 2, explain why x* ( p) is upper hemi-continuous but not
lower hemi-continuous

© John Riley October 11, 2013

Вам также может понравиться