Banach Sabit Nokta Teoremi (Büzülme Teoremi)

Banach's fixed point theorem has important applications to iteration

methods for solving systems of linear algebraic equations and yields
sufficient conditions for convergence and error bounds.
To understand the situation, we first remember that for solving
such a system there are various direct methods (methods that would
yield the exact solution after finitely many arithmetical operations if
the precision-the word length of our computer-were unlimited); a
familiar example is Gauss' elimination method (roughly, a systematic
version of the elimination taught in school). However, an iteration, or
indirect method, may be more efficient if the system is special, for
instance, if it is sparse, that is, if it consists of many equations but has
only a small number of nonzero coefficients. (Vibrational problems,
networks and difference approximations of partial differential equations
often lead to sparse systems.) Moreover, the usual direct methods
require about n3/3 arithmetical operations (n = number of equations =
number of unknowns), and for large n, rounding errors may become
quite large, whereas in an iteration, errors due to roundoff (or even
blunders) may be damped out eventually. In fact, iteration methods are
frequently used to improve "solutions" obtained by direct methods.

E. Kreyszig, “Introductory Functional Analysis with Applications”,John Wiley &Sons, 1978.

 Banach Sabit Nokta Teoremi (Büzülme Teoremi)
Sabit Nokta T:X X x ,T dönüşümünün bir sabit noktasıdır.
x X 
Tx  x
Büzülme X  ( X , d )
T : X  X , T , X ‘de bir büzülmedir

 0    1  d (Tx,Ty)   (x, y) x, y  X
Teorem Banach Sabit Nokta Teoremi
BST1
X  ( X , d ), X  , X tam
T : X  X , T , X ‘de bir büzülmedir
 T ‘nin tek bir sabit noktası vardır.
Tanıt Herhangi bir x0  X belirleyip “ardışıl dizi” oluşturalım
x0 x1  Tx 0 x2  Tx1  T 2 x0 .... xn  Tx n1  ...  T n x0
( xn )