Hotelling - Stability in Competition

Fixed Points
The fixed point problem is this:

Given a set S R n and a function f : S S , is there an x S such that f ( x ) = x ?
FIXED POINT THEOREMS

Banach Fixed Point Theorem
The simplest of all fixed point theorems is ascribed to Stefan Banach (18921945).
Definition
A function
f : S S is called a contraction mapping if
( f ( x) , f ( y))
C ( x, y )
for all
x, y S , where 0 C < 1 is a fixed constant.
In the one-dimensional case, the contraction mapping condition is
f ( x) f ( y) C x y .
Banachs Contraction Mapping Theorem
Let f be a contraction with domain R p and range contained in R p .
point.
Then f has exactly one fixed
The Banach Theorem is quite weak. Consider f : [ 0,1] [0,1] , where f ( x ) = x. This function
barely misses being a contraction since f ( x ) f ( y ) = x y for all x , y [ 0,1] .
However, every
point in [0,1] is a fixed point of this function.
Brouwer Fixed Point Theorem

If S R n is compact and convex and f : S S is continuous there exists x S such that
f ( x ) = x.
Theorem (a baby version of Brouwer fixed-point theorem)

Let f be a continuous function mapping [0,1] into [0,1].
In other words, domain of f is [0,1] and
f ( x ) [0,1] for all x [0,1] . The function f has a fixed pint, i.e., a point x0 [0,1] such that
1
f ( x0 ) = x0 .
The graph of f lies in the unit square.
Our
assertion is equivalent to the assertion that the

graph of f crosses the y=x line, which is almost
obvious.
Correspondences
Let S and T be nonempty sets. A correspondence F from S into T is a rule that assigns to each
element x of S a nonempty subset F ( x ) of T . We write F : S T .
(Correspondence is an
alternative term for a relation between two sets.)

It should be noted that the notation is no difference than that for functions. But of course, functions
may be treated as a special kind of correspondence where the subset F ( x ) of T associated with each
element x of S consists of exactly one element. So, we may say that a correspondence is in general
multi-valued (each element of F ( x ) is a value) whereas a function is single-valued (there being one
and only one value in F ( x ) .)
Examples.
(1) For each x [ 0, ) , let F ( x ) denote the set of real numbers whose squares are x . Thus,
F : [ 0, ) R is a correspondence, with for example,
F ( 0 ) = {0} , F (1) = {1,1} , F ( 2 ) = 2, 2 ,...
(2) For each x R define F ( x ) = ( x, ) = { y R : y > x} . Then F : R R is a correspondence.

Some examples of function values are
F ( 4 ) = ( 4, ) , F (1) = (1, ) .
2
(3) Suppose that : [ a, b ] R and : [ a, b ] R are two functions such that for all
x [ a, b ] , ( x ) ( x ) . Then we can define a correspondence F : [ a, b ] R by
F ( x ) = ( x ) , ( x ) = { y R : ( x ) y ( x )}.
(4) Now, some constructions in higher dimensions. Fix a nonzero vector u R n , and define
F : R n R n by
F ( x) = y Rn : y u x u .
Then F is a correspondence.
(5) Suppose that p = ( p1 ,..., pn ) is a given price system and w the total wealth of a consumer. The
set of all feasible consumptions is then
{( x ,..., x ) R : p x + ... + p x
= { x R : p x w}
F ( w) =
n n
1 1
w, xi 0 (1 i n )
n
+
where we write R+ for the set [ 0, ) of all non-negative real numbers, then we have a
correspondence F : R + R +n .
Graph
Suppose that F : R m R n is a correspondence.
Then the graph of F is defined to be
GF = {( x1 ,..., xm , y1 ,..., yn ) R m + n : ( y1 ,..., yn ) F ( x1 ,..., xm )}

={( x, y ) R m R n : y F ( x )}
Examples.
(1) For F : [ 0, ) R defined by F ( x ) =the set of real numbers whose squares are x.
F ( x ) = { y R : y 2 = x} . Then the graph of F is the parabola y 2 = x .

(2) For F : R R defined by F ( x) = ( x, ) , we have
GF = {( x, y ) R 2 : y ( x, )} = {( x, y ) R 2 : y > x} which is the open halfplane above the line y=x.

(3) For F ( x ) = ( x ) , ( x ) = { y R : ( x ) y ( x )} , x [a , b ] where ( x ) ( x ) for all
x [a , b ] , we have GF = {( x, y ) R 2 : ( x ) y ( x )}
That is the region on the plane bounded between the curves y = ( x ) , y = ( x ) , x=a, and x=b
SEMI-CONTINUITY
Closed and Open Sets
Definition: a set A R n is open if for each point in A, its neighborhood is contained in A.
Definition: a set A R n is closed if its complement, R n \ A is open.
Definition
A correspondence C is called upper semi-continuous (usc), if the set {( x, y) : y C ( x)} is closed.

The set {( x, y) : y C ( x)} is the graph of the correspondence.
Example 1
Here is a correspondence defined on [ 1,1] .

0.5 .
If x=0, C(x)={ 0.5, 0.5 }.
If x [1, 0) then C(x)=0.5.
If x (0,1] then C(x)=
It is easy to check that this correspondence is usc.
Example 2
The correspondence whose graph is portrayed in the

left figure is upper semi-continuous.
Definition
A correspondence C on S R n is called convex valued if C(x) is a convex set for all x S .

The correspondence in both Example 1 and Example 2 above is not convex valued.
Kakutanis fixed point theorem

Let S R n be a compact and convex set.
convex valued.
Let C be a correspondence from S into itself that is usc and
Then, there is an x* S such that x* C ( x* ) .
Example 1
Let C be a correspondence defined on the closed interval [0, 1] that maps a point x to the closed interval
[1-x/2, 1-x/4]. Then C satisfies all the assumptions of the theorem and must have fixed points.
In the diagram, any point on the 45 line (dotted
line in red) which intersects the graph of the
function (shaded in grey) is a fixed point, so in
fact there is an infinity of fixed points in this
particular case. For example, x=0.72 (dashed line
in blue) is a fixed point since 0.72[10.72/2,
10.72/4].
Example 2
The requirement that (x) be convex for all x is essential for the theorem to hold.
Consider the following function defined on [0,1]:
3 / 4,
C ( x ) = {3 / 4,1 / 4},
1 / 4,
0 x < 0.5
x = 0.5
0.5 < x 1
The correspondence has no fixed point. Though it

satisfies all other requirements of Kakutani's
theorem, its value fails to be convex at x = 0.5.

Hotelling - Stability in Competition

Uploaded by

Copyright:

Available Formats

Hotelling - Stability in Competition

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hotelling - Stability in Competition

Uploaded by

Copyright:

Available Formats

Fixed Points

The fixed point problem is this:

FIXED POINT THEOREMS

f : S S is called a contraction mapping if

x, y S , where 0 C < 1 is a fixed constant.

In the one-dimensional case, the contraction mapping condition is

Then f has exactly one fixed

point in [0,1] is a fixed point of this function.

Brouwer Fixed Point Theorem

Theorem (a baby version of Brouwer fixed-point theorem)

In other words, domain of f is [0,1] and

The graph of f lies in the unit square.

assertion is equivalent to the assertion that the

alternative term for a relation between two sets.)

F ( 0 ) = {0} , F (1) = {1,1} , F ( 2 ) = 2, 2 ,...

(2) For each x R define F ( x ) = ( x, ) = { y R : y > x} . Then F : R R is a correspondence.

Then the graph of F is defined to be

GF = {( x1 ,..., xm , y1 ,..., yn ) R m + n : ( y1 ,..., yn ) F ( x1 ,..., xm )}

F ( x ) = { y R : y 2 = x} . Then the graph of F is the parabola y 2 = x .

GF = {( x, y ) R 2 : y ( x, )} = {( x, y ) R 2 : y > x} which is the open halfplane above the line y=x.

A correspondence C is called upper semi-continuous (usc), if the set {( x, y) : y C ( x)} is closed.

Here is a correspondence defined on [ 1,1] .

If x=0, C(x)={ 0.5, 0.5 }.

If x [1, 0) then C(x)=0.5.

If x (0,1] then C(x)=

It is easy to check that this correspondence is usc.

The correspondence whose graph is portrayed in the

A correspondence C on S R n is called convex valued if C(x) is a convex set for all x S .

Kakutanis fixed point theorem

Let C be a correspondence from S into itself that is usc and

Then, there is an x* S such that x* C ( x* ) .

The correspondence has no fixed point. Though it

You might also like