Fixed Point
  Theorems
   a gentle
introduction
  Annual Seminar Week
           IITB
Overview
• Some History
• What are fixed points(FP)?
• The statement of Brouwer’s Fixed point theorem
   o The hairy ball theorem

• Coffee cup , hurricanes , and maps
• An interesting construction of FP’s in a very
  restricted case
What are fixed points?
• Fixed points to a functions are the points where
  f(x)=x
• Fixed point theorem’s basically say that under
  certain conditions , f will have a fixed point
• And variations in these conditions give rise to
  various fixed points theorems.
The obvious fixed point
         theorem
• Every function that maps to itself in one dimension
  has a fixed point (a.k.a. the Intermediate-value
  theorem)

            x2



            x1


                  x1     x2
Generalization to n-dimensions
        Brouwer’s fixed point theorem
• Every continuous function from a closed ball of a
  Euclidean Space to itself has a fixed point.

• Ball => Compact , Convex (not the spherical notion)

• Euclidean Spaces -> n-dimensional spaces
  obvious examples are 2d spaces , 3d spaces
Hairy Ball theorem
• You can't comb a hairy ball flat without creating a
  cowlick!
Some implications
• Fixed point in a coffee cup!
Some implications
• There is always a hurricane somewhere on the
  earth!
• This follows from the hairy ball
   theorem and the fact that
  wind is a continuous transform .
• Brouwer’s FPT is used by
  John Nash
   (“A beautiful mind”)
  to prove the existence of
  Nash-Equilibrium
Some implications

• In computer graphics we sometimes need a
  continuous function that generates an orthogonal
  vector to a given vector.
• The hairy ball theorem implies that there is no such
  function!
A Stronger FP theorem
    Kakutani Fixed point theorem
• Constraint in Brouwer’s FP theorem is modified such
  that now the function is mapping to a subset of
  itself(closed ball).
Geometrical Construction of
fixed point, in a map overlay
• Two maps of different sizes of a country are
  arranged on a table such that one of them lies on
  top of the other and is completely inside it.
• FP in this setting would be a point on the table
  where both the maps point to the same location.
• We use the fact that line joining FP to the vertices
  makes the same angle in both the cases with a
  corresponding edge of the map
• And then some pure geometry !

Fixed Point Theorems

  • 1.
    Fixed Point Theorems a gentle introduction Annual Seminar Week IITB
  • 2.
    Overview • Some History •What are fixed points(FP)? • The statement of Brouwer’s Fixed point theorem o The hairy ball theorem • Coffee cup , hurricanes , and maps • An interesting construction of FP’s in a very restricted case
  • 3.
    What are fixedpoints? • Fixed points to a functions are the points where f(x)=x • Fixed point theorem’s basically say that under certain conditions , f will have a fixed point • And variations in these conditions give rise to various fixed points theorems.
  • 4.
    The obvious fixedpoint theorem • Every function that maps to itself in one dimension has a fixed point (a.k.a. the Intermediate-value theorem) x2 x1 x1 x2
  • 5.
    Generalization to n-dimensions Brouwer’s fixed point theorem • Every continuous function from a closed ball of a Euclidean Space to itself has a fixed point. • Ball => Compact , Convex (not the spherical notion) • Euclidean Spaces -> n-dimensional spaces obvious examples are 2d spaces , 3d spaces
  • 6.
    Hairy Ball theorem •You can't comb a hairy ball flat without creating a cowlick!
  • 7.
    Some implications • Fixedpoint in a coffee cup!
  • 8.
    Some implications • Thereis always a hurricane somewhere on the earth! • This follows from the hairy ball theorem and the fact that wind is a continuous transform . • Brouwer’s FPT is used by John Nash (“A beautiful mind”) to prove the existence of Nash-Equilibrium
  • 9.
    Some implications • Incomputer graphics we sometimes need a continuous function that generates an orthogonal vector to a given vector. • The hairy ball theorem implies that there is no such function!
  • 10.
    A Stronger FPtheorem Kakutani Fixed point theorem • Constraint in Brouwer’s FP theorem is modified such that now the function is mapping to a subset of itself(closed ball).
  • 11.
    Geometrical Construction of fixedpoint, in a map overlay • Two maps of different sizes of a country are arranged on a table such that one of them lies on top of the other and is completely inside it. • FP in this setting would be a point on the table where both the maps point to the same location. • We use the fact that line joining FP to the vertices makes the same angle in both the cases with a corresponding edge of the map • And then some pure geometry !