• Nie Znaleziono Wyników

A short proof of the Caristi theorem

N/A
N/A
Protected

Academic year: 2021

Share "A short proof of the Caristi theorem"

Copied!
2
0
0

Pełen tekst

(1)

ANNALES SOCIETATIS MATHEMATICAE POLONAE Series I: COMMENTATIONES MATHEMATICAE X X (1978) ROCZNIKI POLSKIEQO TOWARZYSTWA MATEMATYCZNEGO

Séria I: PRACE MATEMATYCZNE X X (1978)

L.

P a s i c k i

(Lublin)

A short proof of the Caristi theorem

In 1975 Caristi [2] proved a fixed point theorem which was next used in proving some other results [2], [4]. The original proof used the transfinite induction method and was rather complicated. Then several other proofs лГеге found [1], [3], [4], but all of them are more compli­

cated than the new proof presented below, which uses Zorn’s lemma, bu t in a simple way.

T h e o r e m

1 (Wong [4]). Let f be a self map on a non-empty complete metric space (X , d) and V : _X-> [

0

, oo) a lower semicontinuous function.

Let us presume that the following condition holds:

(

1

) Lor any x

e

X , x Ф f(x) there exists у

e

X

{x} such that : d{x, y) < V { x ) - V ( y ) .

By these assumptions f has a fixed point in X .

P ro o f. In view of Zorn’s lemma there exists a maximal set А <=z X , a

e

A such th a t for all points x, у

e

A

d{x, y) < \V {x )-V { y )\.

For a = inf(F(a?): x e A) there exists a sequence of points' zi e A such th a t ( V {z{))ieN is non-increasing and lim V (£*•) = a.

I t follows from г_>0°

<*(«<, % ) < \ У ( Ъ ) - У { * j ) \

th a t there exists b

e

X , b = lim zt .

i-*CQ

For any x

e

A , if V(x) Ф a, then for sufficiently large i we have d(b, x) < d(b, z j + d fa , x) < d{b, «<) + У(®) —7(«<)

(if for x 0

e

A V {

x

q) = a, then we obtain in a similar way d(b, x 0) =

0

) and then by the lower semicontinuity of V

d { b ,x ) < V { x ) - V ( b ) .

13 — Roczniki PTM Prace Mat. XX.2

(2)

428 L. P a sick i

This means th a t b e A and th a t there is no point y e X such th a t b Ф у and

d{b, y) < V ( b ) - V ( y )

because such у would belong to A. Then it must be so th a t:

d{b,f(b)) = 0 Q.E.D.

The theorem which follows, is a consequence of Theorem 1.

Theorem 2

(Caristi

[2]).

Let all assumptions of Theorem 1 except

(1)

be satisfied. Let us assume that for x e X

M ) < V{x) — V (f(x )).

Then f has a fixed point.

We can modify Wong’s theorem as follows:

Theorem 3.

Let (X , d) be a non-empty complete metric space and V: X ->[0, oo) a lower semicontinuous function. I f a sentence formula g satisfies the following condition on X :

from ~g(x) it follows that there exists y e X such that:

d(x, y) < V ( x ) - V ( y ) , then there exists b e X such that g(b).

R e f e r e n c e s

[1] F .E . B r o w d e r , On theorem of Caristi and KirTc (to appear).

[2] J. C a r is ti, Fixed point theorems for mappings satisfying inwardness conditions, Trans. Amer. Math. Soc. (to appear).

[3] W. A. K irk , CarisWs fixed point theorem and metric convexity (to appear).

[4] C. S. W on g, On a fixed point theorem of contractive type, preprint.

Cytaty

Powiązane dokumenty

The main result of the present paper is Theorem 3, which is a generalization of the C 0 -closing lemma to the case of a not necessarily compact manifold.. Moreover, under

In this paper we give a version of the theorem on local integral invariants of systems of ordinary differential equations1. We give, as an immediate conclusion of this theorem,

To obtain a proof of Gödel’s second incompleteness theorem from this we have to translate the syntax of first-order arithmetic and computability theory into the

Central limit theorems for functionals of general state space Markov chains are of crucial importance in sensible implementation of Markov chain Monte Carlo algorithms as well as

In the present note we intend to clarify the actual p art of the algebraic and topological conditions in this important result, obtaining thus a sur­..

S is conceived as a topological space with the topology determined by the

A short proof of the classical theorem of Menger concerning the number of disjoint AB-paths of a finite graph for two subsets A and B of its vertex set is given.. The main idea of

(The Euler genus of a surface Σ is 2h if Σ is the sphere with h handles and k if Σ is the sphere with k crosscaps.) In this note, we give a simple proof of Theorem 1.1 by using