• Nem Talált Eredményt

Periodic fixed points of random operators

N/A
N/A
Protected

Academic year: 2022

Ossza meg "Periodic fixed points of random operators"

Copied!
11
0
0

Teljes szövegt

(1)

(2010) pp. 39–49

http://ami.ektf.hu

Periodic fixed points of random operators

Ismat Beg

a

, Mujahid Abbas

a

, Akbar Azam

b

aCenter for Advanced Studies in Mathematics Lahore University of Management Sciences

bDepartment of Mathematics

COMSATS Institute of Information Technology

Submitted 10 February 2010; Accepted 19 April 2010

Abstract

Sufficient conditions for existence of random fixed point of a nonexpan- sive rotative random operator are obtained and existence of random periodic points of a random operator is proved. We also derive random periodic point theorem forǫ- expansive random operator.

Keywords:Random periodic point; random fixed point;ǫ- contractive random operator; ǫ- expansive random operator; rotative random operator; metric space; Banach space; measurable space.

MSC:47H09, 47H10, 47H40, 54H25, 60H25

1. Introduction

Random nonlinear analysis has grown into an active research area closely associ- ated with the study of random nonlinear operators and their properties needed in solving nonlinear random operator equations (see [7, 18, 21]). The study of random fixed point theory was initiated by the Prague school of probabilists in the 1950’s ([15, 24]). Random fixed point theorems are of tremendous importance in proba- bilistic functional analysis as they provide a convenient way of modelling many real life problems and random methods have also revolutionized the financial markets.

The survey article by Bharucha -Reid [8] in 1976 attracted the attention of several mathematician and gave wings to this theory. Itoh [17] extended Spacek’s and Hans’s theorems to random multivalued contraction mappings. In recent years, a lot of efforts have been made ([2, 3, 4, 5, 6, 16, 22, 23], and references therein) to show the existence of random fixed points of certain random single valued and mul- tivalued operators and various applications in diverse area from pure mathematics

39

(2)

to applied sciences have been explored. The aim of this paper is to establish the existence of random fixed point of nonexpansive rotative random operator in the setting of Banach spaces. A random analogue of Edelstein theorem to establish the existence of random periodic points for random single valuedǫ- contractive op- erator is proved. These results are then used to obtain the random periodic point of ǫ- expansive random operators. The results proved in this paper improve and generalize several well known results in the literature [9, 12, 17].

2. Preliminaries

We begin with some definitions and state the notations used throughout this paper.

Let(Ω,Σ)be a measurable space (Σ- sigma algebra) andFbe a nonempty subset of a separable metric space(X, d).A single valued mappingT: Ω→Xismeasurableif T1(U)∈Σfor each open subsetU ofX,whereT1(U) ={ω∈Ω :T(ω)∩U 6=∅}.

A mapping T: Ω×X → X is a random operator if and only if for each fixed x ∈ X, the mapping T(., x) : Ω → X is measurable and it is continuous if for eachω ∈Ω, the mappingT(ω, .) :X →X is continuous. A measurable mapping ξ: Ω→X is arandom fixed pointof a random operatorT: Ω×X →X if and only ifξ(ω) =T(ω, ξ(ω)))for eachω∈Ω.We denote the set of random fixed points of a random operatorT byRF(T)and the set of all measurable mappings fromΩinto X byM(Ω, X).For the random operatorf: Ω×X →X , the mapfω−1:X →X is defined by fω1(y) =xif and only if f(ω, x) =y.

We denote thenthiterateT(ω, T(ω, T(ω, . . . , T(ω, x)· · ·)))of random operator T: Ω×X→XbyTn(ω, x).The letterIdenotes the random operatorI: Ω×X→ X defined by I(ω, x) = xand T0 =I. The random operator T: Ω×X → X is called random periodic operator with periodp∈N, if for eachx∈X andω ∈Ω we obtainTp(ω, x) =I(ω, x).LetB(x0, r)denotes the spherical ball centred atx0

with radiusr, defined as the set{x∈X :d(x, x0)6r}.

Definition 2.1. LetF be a nonempty subset of a separable metric spaceX. The random operatorT: Ω×F →F is said to be:

(a) k(ω)-contraction random operatorif for anyx, y∈F andω∈Ω,we have d(T(ω, x), T(ω, y))6k(ω)d(x, y),

wherek: Ω→[0,1)is a measurable map.Ifk(ω) = 1 for anyω∈Ω,thenT is callednonexpansive random operator.

(b) contractive random operatorif for anyx, y∈F andω∈Ω,we have d(T(ω, x), T(ω, y))< d(x, y).

(c) ǫ-contractive random operator if for ǫ > 0 and x, y ∈ F with x 6= y and d(x, y)< ǫ,we have,

d(T(ω, x), T(ω, y))< d(x, y),

(3)

for everyω∈Ω.Obviously, every contractive random operator isǫ- contrac- tive random operator for anyǫ >0.

(d) ǫ-expansive random operator if for ǫ > 0 and x, y ∈ F with x 6= y and d(x, y)< ǫ,we have

d(T(ω, x), T(ω, y))> d(x, y), (2.1) for everyω∈Ω. If inequality(2.1)holds for everyx, y∈X withx6=y then T is called an expansive random operator.

Obviously, every expansive random operator isǫ- expansive random operator for anyǫ >0.

Definition 2.2. LetT: Ω×F →F be a random operator, whereF is a nonempty subset of a separable complete metric spaceX. A measurable mappingξ: Ω→Fis called arandom periodic pointofT there existsn>1such thatTn(ω, ξ(ω)) =ξ(ω), for everyω∈Ω.That is, random periodic point is random fixed point ofnthiterate of T for some n>1. The least such positive integernis called period of random periodic point ξ.

Note that random fixed point ofT is also random periodic point ofT of period 1 but there exists a random periodic point ofT which fails to be the random fixed point ofT as shown in the examples presented below. It is also shown that there exists a random operator having random periodic point of period 5 but does not posses the random periodic point of period 3.

Example 2.3. LetΩ = [0,1]andΣbe the sigma algebra of Lebesgue’s measurable subsets of Ω. Take X = R with d(x, y) = |x−y|, for x, y ∈ R. Define random operatorT from Ω×X to X as,

T(ω, x) =

2−x, if(ω, x)∈Ω×[0,1]

ω2−x−1, otherwise.

Define the measurable mappingξ: Ω→X asξ(ω) = 12(3ω2−1),for everyω∈Ω.

Now ξ is a random periodic point ofT with period 2 but it fails to be a random fixed point ofT.

Example 2.4. LetΩ = [0,1]andΣbe the sigma algebra of Lebesgue’s measurable subsets of Ω. Take X = R with d(x, y) = |x−y|, for x, y ∈ R. Define random operatorT fromΩ×X toX as,T(ω,1) = 3, T(ω,2) = 5, T(ω,3) = 4, T(ω,4) = 2, T(ω,5) = 1andT(ω, x) =x−ω,whenx /∈ {1,2,3,4,5}.

Define measurable mappingξ: Ω→X asξ(ω) = 1,for every ω∈Ω.Note that ξ is a random periodic point of period 5. It is also noted that random operator T in this example does not posses random fixed point because for anyξ to be the random fixed point, we must have T(ω, ξ(ω)) = ξ(ω), for every ω ∈ Ω. But this random operator equation holds only forω= 0.

(4)

Remark 2.5. LetFbe a closed subset of a complete separable metric spaceXand the sequence of measurable mappings{ξn} fromΩtoF be point wise convergent, that is,ξn(ω)→q:=ξ(ω)for eachω∈Ω. Thenξbeing the limit of the sequence of measurable mappings is measurable and closedness of F impliesξis a mapping fromΩtoF.SinceF is a subset of a complete separable metric spaceX, also ifT is a continuous random operator fromΩ×F to F then by the lemma 8.2.3 of [1], the mapω→Tn(ω, f(ω))is measurable for any measurable mappingf fromΩto F.

Definition 2.6. LetF be a nonempty subset of a Banach space X. The random operatorT: Ω×F →F is said to be(k, n)−rotative random operatorfork < n, if for eachω ∈Ω,

kξ(ω)−Tn(ω, ξ(ω))k6kkξ(ω)−T(ω, ξ(ω))k,

where ξ is a mapping fromΩ to F and n∈ N.The operator T is said to ben−

rotative random operator if it (k, n)− rotative random operator for some k < n and T is calledrotative random operator if it is an n- rotative random operator for some n∈N. Note that any random periodic operator with period pis (0, p)- rotative random operator.

Remark 2.7. IfT: Ω×F →F isk(ω)contraction random operator where F is a closed subset of Banach spaceX and n >1.For anyξ: Ω→F,consider,

kξ(ω)−Tn(ω, ξ(ω))k6

n

X

k=1

Tk−1(ω, ξ(ω))−Tk(ω, ξ(ω))

6(1 +k(ω) + (k(ω))2+· · ·

+ (k(ω))n−1)kξ(ω)−T(ω, ξ(ω))k

< nkξ(ω)−T(ω, ξ(ω))k, for everyω∈Ω. ThusT is a rotative random operator.

3. Periodic and fixed points of rotative random op- erators

In this section, we first show an existence of a random fixed point of a nonexpansive rotative random operator which not only provides a random analogue of theorem 17.1 of [11] (see also [12]) but also improves theorem 2.1 of [17] in the sense that it does not require the boundedness of T(ω, F)for anyω ∈Ω. Moreover we replace continuous condensing random operator by nonexpansive rotative random operator.

Periodic point problems were systematically studied since the beginning of fifties (see [9, 10, 13, 14, 19, 20]). We show some results on the existence of random periodic points of random single valuedǫ- contractive operator in the setting of a separable metric space.

(5)

Theorem 3.1. Let F be a nonempty closed and convex subset of a separable Ba- nach space X and T: Ω×F → F be a nonexpansive rotative random operator.

Then T has a random fixed point.

Proof. Letξ: Ω→F be any fixed measurable mapping. For0 < α <1and any arbitrary measurable mappingη: Ω→F,define Tα: Ω×F →F as,

Tα(ω, η(ω)) = (1−α)ξ(ω) +αT(ω, η(ω)).

Note that for eachα, the random operatorTα has Lipschitz constantα. we may apply [8] to obtain the sequence of random operators Fα: Ω×F → F such that Tα(ω, Fα(ω, ξ(ω))) =Fα(ω, ξ(ω)), for everyω∈Ω.Consequently, we have

Fα(ω, ξ(ω)) = (1−α)ξ(ω) +αT(ω, Fα(ω, ξ(ω))).

It can be verified that eachFα is nonexpansive random operator. By iteratingFα

we obtain

Fαk(ω, ξ(ω)) = (1−α)Fαk−1(ω, ξ(ω)) +αT(ω, Fαk(ω, ξ(ω))), k∈N. (3.1) Note that,

(1−α)Fα(ω, ξ(ω))

= (1−α)ξ(ω) +αT(ω, Fα(ω, ξ(ω)))−αFα(ω, ξ(ω))

= (1−α)ξ(ω) +α(T(ω, Fα(ω, ξ(ω)))−Fα(ω, ξ(ω))).

Thus for each ω∈Ω

(1−α)(ξ(ω)−Fα(ω, ξ(ω)))

=α(Fα(ω, ξ(ω))−T(ω, Fα(ω, ξ(ω)))). (3.2) Now supposeT is a (a, n)-rotative random operator, that is

kξ(ω)−Tn(ω, ξ(ω))k6akξ(ω)−T(ω, ξ(ω))k, for everyω∈Ω. Now,

Fα(ω, ξ(ω))−Fα2(ω, ξ(ω))

=

(1−α)ξ(ω) +αT(ω, Fα(ω, ξ(ω)))−(1−α)Fα(ω, ξ(ω))

−αT(ω, Fα2(ω, ξ(ω)))

=

(1−α)(ξ(ω)−Fα(ω, ξ(ω))) +α(T(ω, Fα(ω, ξ(ω)))

−αT(ω, Fα2(ω, ξ(ω)))

=

α(Fα(ω, ξ(ω))−T(ω, Fα(ω, ξ(ω)))) +α(T(ω, Fα(ω, ξ(ω)))

−αT(ω, Fα2(ω, ξ(ω)))

Fα(ω, ξ(ω))−T(ω, Fα2(ω, ξ(ω)))

(6)

6αkFα(ω, ξ(ω))−Tn(ω, Fα(ω, ξ(ω)))k +α

Tn(ω, Fα(ω, ξ(ω)))−T(ω, Fα2(ω, ξ(ω))) 6αakFα(ω, ξ(ω))−T(ω, Fα(ω, ξ(ω)))k

Tn−1(ω, Fα(ω, ξ(ω)))−Fα2(ω, ξ(ω))

= (1−α)akFα(ω, ξ(ω))−ξ(ω)k +α

Tn1(ω, Fα(ω, ξ(ω)))−Fα2(ω, ξ(ω)) ,

for everyω∈Ω.Now we claim that the following inequality holds for everyω∈Ω andm>2.

α

Tm−1(ω, Fα(ω, ξ(ω)))−Fα2(ω, ξ(ω)) 6(m−1)−mα+αmkξ(ω)−Fα(ω, ξ(ω))k

m

Fα(ω, ξ(ω))−Fα2(ω, ξ(ω))

. (3.3)

For this consider, α

T(ω, Fα(ω, ξ(ω)))−Fα2(ω, ξ(ω))

T(ω, Fα(ω, ξ(ω)))−(1−α)Fα(ω, ξ(ω))−αT(ω, Fα2(ω, ξ(ω)))

(1−α)(T(ω, Fα(ω, ξ(ω)))−Fα(ω, ξ(ω)))−α(T(ω, Fα2(ω, ξ(ω)))

−T(ω, Fα(ω, ξ(ω))))

6(1−α)kα(T(ω, Fα(ω, ξ(ω)))−Fα(ω, ξ(ω)))k

2

T(ω, Fα2(ω, ξ(ω)))−T(ω, Fα(ω, ξ(ω)))

= (1−α)2kξ(ω)−Fα(ω, ξ(ω))k+α2

T(ω, Fα2(ω, ξ(ω)))−T(ω, Fα(ω, ξ(ω))) 6(1−α)2kξ(ω)−Fα(ω, ξ(ω))k+α2

Fα2(ω, ξ(ω))−Fα(ω, ξ(ω)) . So (3.3) is valid form= 2and for anyω∈Ω.

Assuming the validity of (3.3) form=j and for anyω∈Ω,consider α

Tj(ω, Fα(ω, ξ(ω)))−Fα2(ω, ξ(ω))

Tj(ω, Fα(ω, ξ(ω)))−(1−α)Fα(ω, ξ(ω))−αT(ω, Fα2(ω, ξ(ω)))

(1−α)(Tj(ω, Fα(ω, ξ(ω)))−Fα(ω, ξ(ω))) +α(Tj(ω, Fα(ω, ξ(ω)))

−T(ω, Fα2(ω, ξ(ω))))

6α(1−α)

Tj(ω, Fα(ω, ξ(ω)))−Fα(ω, ξ(ω)) +α2

Tj(ω, Fα(ω, ξ(ω)))−T(ω, Fα2(ω, ξ(ω))) 6jα(1−α)kFα(ω, ξ(ω))−T(ω, Fα(ω, ξ(ω)))k

2

Tj1(ω, Fα(ω, ξ(ω)))−Fα2(ω, ξ(ω)) 6jα(1−α)kFα(ω, ξ(ω))−T(ω, Fα(ω, ξ(ω)))k

+α[(j−1)−jα+αj]kξ(ω)−Fα(ω, ξ(ω))k +αj+1

Fα(ω, ξ(ω))−Fα2(ω, ξ(ω))

6j(1−α)22[(j−1)−jα+αj]kξ(ω)−Fα(ω, ξ(ω))k

(7)

j+1

Fα(ω, ξ(ω))−Fα2(ω, ξ(ω)) 6[j−(j+ 1)α+αj+1]kξ(ω)−Fα(ω, ξ(ω))k

j+1

Fα(ω, ξ(ω))−Fα2(ω, ξ(ω)) .

So by induction inequality (3.3)is valid for everyω∈Ωandm>2.

Now consider, forω∈Ω

Fα(ω, ξ(ω))−Fα2(ω, ξ(ω)) 6(1−α)akFα(ω, ξ(ω))−ξ(ω)k

Tn−1(ω, Fα(ω, ξ(ω)))−Fα2(ω, ξ(ω)) 6(1−α)akFα(ω, ξ(ω))−ξ(ω)k

+ [(n−1)−nα+αn]kξ(ω)−Fα(ω, ξ(ω))k +αn

Fα(ω, ξ(ω))−Fα2(ω, ξ(ω)) . It further implies that

(1−αn)

Fα(ω, ξ(ω))−Fα2(ω, ξ(ω))

6[(1−α)a+ (n−1)−nα+αn]kξ(ω)−Fα(ω, ξ(ω))k, for everyω∈Ω. Now we arrive at

Fα(ω, ξ(ω))−Fα2(ω, ξ(ω))

6(1−αn)1[(1−α)a+ (n−1)−nα+αn]kξ(ω)−Fα(ω, ξ(ω))k 6(a+n)(1−α)(1−αn)1−1kξ(ω)−Fα(ω, ξ(ω))k

= [(a+n)(

n−1

X

i=0

αi)1−1]kξ(ω)−Fα(ω, ξ(ω))k

=g(α)kξ(ω)−Fα(ω, ξ(ω))k,

for every ω ∈ Ω, where g(α) = [(a+n)(Pn−1

i=0 αi)1−1]. Since g is continuous and decreasing for α∈(0,1]with g(1) = na <1, there exists b ∈(0,1]such that g(1)<1 forα∈(b,1].For such α,the sequence of measurable mappings defined byηn(ω) =Fαn(ω, ξ(ω))→η(ω), for eachω∈Ω, η: Ω→F,being the limit of the sequence of measurable functions, is also measurable (see remark 2.6).From (3.1)

it follows that η is a random fixed point ofT.

Example 3.2. LetΩ = [0,1]andΣbe the sigma algebra of Lebesgue’s measurable subsets of Ω. Take X = R with d(x, y) = |x−y|, for x, y ∈ R. Define random operatorT from Ω×X to X as,T(ω, x) =ω−x.

Define a fixed measurable mapping ξ: Ω→ X as ξ(ω) = ω3, for everyω ∈Ω.

Note that T is nonexpansive random operator. Since random operator equation T2(ω, ξ(ω)) = ξ(ω)holds for every ω ∈ Ω, therefore it is (2,1)−rotative random operator. Thus the conditions of Theorem 3.1 are satisfied. Moreover a measurable mappingη: Ω→X defined asη(ω) =ω2,for everyω∈Ω,serve as a unique random fixed point ofT.

(8)

Theorem 3.3. Let X be a separable metric space and T: Ω×X → X be a ǫ- contractive random operator. Let ξ0: Ω→X be any measurable mapping such that a sequence {Tn(ω, ξ0(ω))} has a point wise convergent subsequence of measurable mappings. Then T has a random periodic point.

Proof. Let{Tni(ω, ξ0(ω))}be a subsequence of{Tn(ω, ξ0(ω))}such thatTni(ω, ξ0

(ω))→ξ(ω)for eachω∈Ωasni→ ∞where{ni}is a strictly increasing sequence of positive integers. The mapping ξ: Ω → X being point wise limit of sequence of measurable mappings is measurable. Define sequence of measurable mappings ξi: Ω→X as ξi(ω) =Tni(ω, ξ0(ω)). Givenǫ >0, there exists an integern0 such that

d(ξi(ω), ξ(ω))< ǫ

4, fori>n0and ω∈Ω.

Put k=ni+1−ni.Consider,

d(ξi+1(ω), Tk(ω, ξ(ω))) =d(Tk(ω, ξi(ω)), Tk(ω, ξ(ω)))

< d(ξi(ω), ξ(ω))< ǫ

4, for eachω∈Ω.

Now,

d(ξ(ω), Tk(ω, ξ(ω)))

6d(ξi+1(ω), Tk(ω, ξ(ω))) +d(ξi+1(ω), ξ(ω))

< ǫ 4+ ǫ

4 = ǫ

2, for everyω∈Ω.

Now we claim that ξ is a random periodic point ofT. To prove this, assume that η: Ω→X be any measurable mapping such thatη(ω) =Tk(ω, ξ(ω))but

η(ω)6=ξ(ω), for some ω∈Ω. (3.4) Which implies that0< d(η(ω), ξ(ω))< ǫ. AsT is aǫ- contractive random operator therefore forω∈Ωfor which (3.4) holds, we have

d(T(ω, ξ(ω)), T(ω, η(ω)))< d(ξ(ω), η(ω)).

Define h: Ω×X2 →R as,h(ω, x(ω), y(ω)) = d(T(ω,x(ω)),T(ω,y(ω)))

d(x(ω),y(ω)) , wherex(ω)6=

y(ω) ∈ X for each ω ∈ Ω. Now h(ω, ., .) is continuous at (ξ(ω), η(ω)) for every ω∈Ωfor which (3.4) is valid.

Take 0 < α < 1, then there exists δ > 0 such that x(ω) ∈ B(ξ(ω), δ) and y(ω)∈B(η(ω), δ)gives

d(T(ω, x(ω)), T(ω, y(ω)))< αd(x(ω), y(ω)).

As, lim

r→∞Tk(ω, ξr(ω)) = Tk(ω, ξ(ω)) = η(ω), for every ω ∈ Ω. So there exists n1>n0 such that

d(ξr(ω), ξ(ω))< δ

(9)

and

d(Tk(ω, ξr(ω)), η(ω))< δ, forr>n1 andω∈Ω.Hence we have

d(T(ω, ξr(ω)), T(ω, Tk(ω, ξr(ω))))< αd(ξr(ω), Tk(ω, ξr(ω))). (3.5) Consider,

d(ξr(ω), Tk(ω, ξr(ω)))

6d(ξr(ω), ξ(ω)) +d(ξ(ω), Tk(ω, ξ(ω))) +d(Tk(ω, ξ(ω)), Tk(ω, ξr(ω)))

< ǫ 4 +ǫ

2 + ǫ

4 =ǫ, (3.6)

for r>n1 >n0 and ω ∈Ωfor which (3.4) holds. Now using (3.5) and (3.6), we have

d(T(ω, ξr(ω)), T(ω, Tk(ω, ξr(ω))))

< αd(ξr(ω), Tk(ω, ξr(ω)))< d(ξr(ω), Tk(ω, ξr(ω)))< ǫ,

for r>n1. Since T is a ǫ- contractive random operator so forr >n1 andq >0, we have

d(Tq(ω, ξr(ω)), Tq(ω, Tk(ω, ξr(ω))))

< d(ξr(ω), Tk(ω, ξr(ω)))< ǫ α.

Put q=nr+1−nr,we haved(ξr+1(ω), Tk(ω, ξr+1(ω)))<αǫ.Hence, d(ξs(ω), Tk(ω, ξs(ω)))< ǫαs−r.

Now,

d(ξ(ω), η(ω)) 6d(ξ(ω), ξs(ω)) +d(ξs(ω), Tk(ω, ξs(ω))) +d(Tk(ω, ξs(ω)), η(ω))→0, as s→ ∞.

for thoseω∈Ωfor which (3.4) holds. This contradiction concludes the result.

Corollary 3.4. If in theorem 3.2, the random periodic point ξ(say) ofT satisfies d(ξ(ω), T(ω, ξ(ω)))< ǫ, for everyω ∈Ω. (3.7) Then ξ is a random fixed point ofT.

Proof. Letkbe the positive integer such thatTk(ω, ξ(ω)) =ξ(ω),for everyω∈Ω.

Ifξis not a random fixed point ofT,thenξ(ω)6=T(ω, ξ(ω)for someω∈Ω.Since T isǫ- contractive random operator, using (3.7) we have

d(ξ(ω), T(ω, ξ(ω))) =d(Tk(ω, ξ(ω)), Tk+1(ω, ξ(ω)))

< d(ξ(ω), T(ω, ξ(ω))).

This contradiction concludes the proof.

(10)

Remark 3.5. IfX is a separable compact metric space andT: Ω×X →X is an ǫ- contractive random operator. Then applying theorem 3.3, we conclude that T has a random periodic point.

Theorem 3.6. Let X be a separable compact metric space and T: Ω×X → X be an ǫ- contractive random operator. Then T has finitely many random periodic points.

Proof. Let ξ, ζ: Ω → X be two random periodic points ofT with ξ(ω) 6=ζ(ω) and d(ξ(ω), ζ(ω)) < ǫ for some ω ∈ Ω. Let m, n > 1 be two integers such that Tm(ω, ξ(ω)) = ξ(ω) and Tn(ω, ζ(ω)) = ζ(ω) for every ω ∈ Ω. Obviously Tmn(ω, ξ(ω)) =ξ(ω)andTmn(ω, ζ(ω)) =ζ(ω)for eachω∈Ω.Now consider,

d(ξ(ω), ζ(ω)) =d(Tmn(ω, ξ(ω)), Tmn(ω, ζ(ω)))

< d(ξ(ω), ζ(ω)),

which is contradiction. Therefore any two random periodic point of T must be at leastǫ- apart. Compactness ofX prevents us defining infinitely many random

periodic points fromΩ×X to X.

Acknowledgement. The authors are thankful to referee for precise remarks to improve the presentation of the paper.

References

[1] Aubin, J. P., and Frankowska, H., Set-Valued Analysis, Birkhauser, Boston, 2009.

[2] Beg, I., Random fixed points of random operators satisfying semicontractivity con- ditions,Math. Japan., 46 (1) (1997), 151–155.

[3] Beg, I., Random Edelstein theorem,Bull.Greek Math. Soc., 45 (2001), 31–41.

[4] Beg, I., Minimal displacement of random variables under Lipschitz random maps, Topological Methods in Nonlinear Anal., 19 (2002), 391–397.

[5] Beg, I., and Abbas, M., Iterative procedures for solutions of random operator equations in Banach spaces,J. Math. Anal. Appl., 315(1)(2006), 181–201.

[6] Beg, I., and Thakur, B.S., Solution of random operator equations using general composite implicit iteration process,Int. J. Modern Math., 4(1)(2009), 19–34.

[7] Bharucha-Reid, A.T.,Random Integral Equations, Academic Press, New York and London, 1972.

[8] Bharucha- Reid, A.T., Fixed point theorems in probabilistic analysis,Bull. Amer.

Math. Soc., 82 (1976), 641–645.

[9] Edelstein, M., On fixed and periodic points under contractive mapping,J. London Math. Soc., 37 (1962), 74–79.

[10] Fournier, G., and Górniewicz, L., The Lefschetz fixed point theorem for some non compact multivalued maps,Fund. Math., 94 (1977), 245–254.

(11)

[11] Goebel, K., andKirk, W.A.,Topics in Metric Fixed Point Theory, Cambridge University Press, Cambridge 1990.

[12] Goebel, K., and Koter, M., A remark on nonexpansive mappings, Canadian Math. Bull., 24 (1981), 113–115.

[13] Górniewicz, L.,Topological Fixed Point Theory of Multivalued Mappings, Kluwer, Dordrecht, 1999.

[14] Halpern, B., Periodic points on tori,Pacific J. Math., 83 (1979), 117–133.

[15] Hanš, O., Reduzierende zulliällige transformaten,Czechoslovak Math. J., 7 (1957), 154–158.

[16] Hanš, O., Random operator equations, in: Proceedings of the Fourth Berkeley Sym- posium on Mathematical Statistics and Probability, Vol. II, Part I, University of Cal- ifornia Press, Berkeley, 1961, 185–202.

[17] Itoh, S., Random fixed point theorems with an application to random differential equations in Banach spaces,J. Math. Anal. Appl., 67 (1979), 261–273.

[18] Kalmoun, E.M., Some deterministic and random vector equilibrium problems, J.

Math. Anal. Appl., 267 (2002), 62–75.

[19] Kampen, J., On fixed points of maps and iterated maps and applications,Nonlinear Anal., 42(3) (2000), 509–532.

[20] Matsuoka, T., The number of periodic points of smooth maps, Ergod. Th. and Dynam. Sys., 8 (1989), 153–163.

[21] O’Regan, D., A continuation type result for random operators,Proc. Amer. Math.

Soc., 126, 7(1998), 1963–1971.

[22] Papageorgiou, N.S., Random fixed point theorems for measurable multifunctions in Banach spaces,Proc. Amer. Math. Soc., 97 (1986), 507–514.

[23] Papageorgiou, N.S., On measurable multifunctions with stochastic domain, J.

Austral. Math. Soc.(Ser. A), 45 (1988), 204–216.

[24] Spacek, A., Zufällige Gleichungen,Czechoslovak Math. J., 5 (1955), 462–466.

Ismat Beg, Mujahid Abbas

Center for Advanced Studies in Mathematics, Lahore University of Management Sciences, Lahore-54792, Pakistan

Phone: 0092-42-35608229 Fax: 0092-42-35722591 e-mail: ibeg@lums.edu.pk

mujahid@lums.edu.pk Akbar Azam

Department of Mathematics,

COMSATS Institute of Information Technology, Islamabad, Pakistan

Hivatkozások

KAPCSOLÓDÓ DOKUMENTUMOK

2 High-speed elevator car system dynamics model In order to solve the transverse acceleration response of a high speed elevator car system with random parameters under

Keywords: Almost sure limit theorem, multiindex, random field, α -mixing random field, strong law of large numbers.. MSC:

Tómács in [6] proved a general convergence rate theorem in the law of large numbers for arrays of Banach space valued random elements.. We shall study this theorem in case Banach

Thus it is natural to consider random sequences (n k ), and in this paper we consider the case when the gaps n k+1 − n k are independent, identically distributed (i.i.d.)

More precisely, in addition to extending the results of Heyde (1982, 1993) on the weak Donsker and strong Strassen type asymptotic path behaviour of a 2-dimensional anisotropic

The main aim of this paper is to compute the volume of general and unital qubit channels and investigate the distribution of the resulted quantum states if a general or unital

We prove the quenched version of the central limit theorem for the displacement of a random walk in doubly stochastic random environment, under the H − 1 -condition, with

4. Számoljuk ki annak a valószínűségét, hogy a 3.feladatbeli játékos először az m-dik lépésben lesz újra kiinduló pénzénél úgy, hogy addig egyszer sem kellett a kiinduló