• 検索結果がありません。

Well-Posedness and Fractals via Fixed Point Theory

N/A
N/A
Protected

Academic year: 2022

シェア "Well-Posedness and Fractals via Fixed Point Theory"

Copied!
9
0
0

読み込み中.... (全文を見る)

全文

(1)

Volume 2008, Article ID 645419,9pages doi:10.1155/2008/645419

Research Article

Well-Posedness and Fractals via Fixed Point Theory

Cristian Chifu and Gabriela Petrus¸el

Department of Business, Faculty of Business, Babes¸-Bolyai University Cluj-Napoca, Horea 7, 400174 Cluj-Napoca, Romania

Correspondence should be addressed to Gabriela Petrus¸el,[email protected] Received 25 August 2008; Accepted 6 October 2008

Recommended by Andrzej Szulkin

The purpose of this paper is to present existence, uniqueness, and data dependence results for the strict fixed points of a multivalued operator of Reich type, as well as, some sufficient conditions for the well-posedness of a fixed point problem for the multivalued operator.

Copyrightq2008 C. Chifu and G. Petrus¸el. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

1. Introduction

LetX, dbe a metric space. We will use the following symbolssee also1:

PX {Y ⊂X|Y /∅};

PbX {Y ∈PX|Y is bounded};

PclX {Y ∈PX|Y is closed};

PcpX {Y ∈PX|Y is compact}.

IfT : XPXis a multivalued operator, then forYPX,TY

x∈YTxwe will denote the image of the setY throughT.

Throughout the paperFT :{x∈X |xTx}resp.,SFT :{x∈X | {x}Tx}

denotes the fixed point setresp., the strict fixed point setof the multivalued operatorT. We introduce the following generalized functionals.

Theδgeneralized functional

δd:PPX−→R∪ {∞}, δdA, B sup

da, b|aA, bB .

1.1

(2)

The gap functional

Dd :PPX−→R∪ {∞}, DdA, B inf

da, b|aA, bB

. 1.2

The excess generalized functional

ρd:PPX−→R∪ {∞}, ρdA, B sup

Dda, B|aA

. 1.3

The Pompeiu-Hausdorffgeneralized functional

Hd:PX×PX−→R∪ {∞}, HdA, B max

ρdA, B, ρdB, A

. 1.4

The first purpose of this paper is to present existence, uniqueness, and data dependence results for the strict fixed point of a multivalued operator of Reich type. Since, in our approach, the strict fixed point is constructed by iterations, this generates the possibility to give some sufficient conditions for the well-posedness of a fixed point problem for the multivalued operator mentioned below.

Definition 1.1. LetX, dbe a metric space andT :XPclX. ThenT is called a multivalued δ-contraction of Reich type, if there exista, b, c∈Rwithabc <1 such that

δ

Tx, Ty

adx, y bδ

x, Tx

y, Ty

, 1.5

for allx, yX.

The notion of well-posed fixed point problem for single valued and multivalued operator was defined and studied by F.S. De Blasi and J. Myjak, S. Reich and A.J. Zaslavski, Rus and Petrus¸el2, Petrus¸el et al.3.

Definition 1.2see Petrus¸el and Rus2and3. ALetX, dbe a metric space,YPX andT :YPclXbe a multivalued operator.

Then the fixed point problem is well posed forTwith respect toDdif a1FT {x}i.e.,xTx;

b1IfxnY,n∈NandDdxn, Txn→0 asn→ ∞thenxnxasn→ ∞.

B Let X, d be a metric space, YPX and T : YPclX be a multivalued operator.

Then the fixed problem is well posed forTwith respect toHdif a2 SFT {x}i.e.,{x}Tx;

b2IfxnY,n∈NandHdTxn→0 asn→ ∞thenxnxasn→ ∞.

(3)

The second aim is to study the existence of an attractori.e., the fixed point of the multifractal operator, see 4–7 for an iterated multifunction system consisting of nonself multivalued operators.

2. Main results

We will give first another proofa constructive oneof a result given by Reich8in 1972.

For some similar results, see9,10. In our proof, the strict fixed point will be obtained by iterations.

Theorem 2.1Reich’s theorem. LetX, dbe a complete metric space and letT :XPbXbe a multivalued operator, for which there exista, b, c∈Rwithabc <1 such that

δ

Tx, Ty

adx, y bδ

x, Tx

y, Ty

, ∀x, y∈X. 2.1

ThenT has a unique strict fixed point inX, that is,SFT{x}.

Proof. Letq >1 andx0Xbe arbitrarily chosen. Then there existsx1Tx0such that δ

x0, T x0

qd x0, x1

. 2.2

We have δ

x1, T x1

δ T

x0

, T x1

ad x0, x1

x0, T

x0

x1, T

x1

≤abqd x0, x1

x1, T

x1

.

2.3

It follows that

δ x1, T

x1

abq 1−c d

x0, x1

. 2.4

Forx1Tx0,there existsx2Tx1such that δ

x1, T x1

qd x1, x2

. 2.5

Then

δ x2, T

x2δ T

x1

, T x2

ad x1, x2

x1, T

x1

x2, T

x2

≤abqd x1, x2

x2, T

x2

.

2.6

(4)

It follows that

δ x2, T

x2

abq 1−c d

x1, x2

abq 1−c δ

x1, T x1

abq 1−c

2

d x0, x1

.

2.7

Inductively, we can construct a sequencexnn∈Nhaving the properties 1 αxnTxn−1, n∈N;

2 βdxn, xn1δxn, Txn≤abq/1cndx0, x1.

We will prove now that the sequencexnn∈Nis Cauchy.

We successively have

d

xn, xnp

d

xn, xn1 d

xn1, xn2

· · ·d

xnp−1, xnp

abq 1−c

n

abq 1−c

n1 · · ·

abq 1−c

np−1 d

x0, x1

.

2.8

Let us denoteα: abq/1c. Then

d

xn, xnp

αn

1α· · ·αp−1 d

x0, x1

αnαp−1 α−1 d

x0, x1

. 2.9

If we choseq <1−ac/b, thenα <1.

Lettingn→ ∞, sinceαn→0, it follows that

d

xn, xnp

−→0 asn−→ ∞. 2.10

Hencexnn∈Nis Cauchy.

By the completeness of the spaceX, d,we get that there exists xX such that xnxasn→ ∞.

Next, we will prove thatx∈SFT. We have

δ x, T

x

d x, xn

δ xn, T

xn

δ T

xn

, T x

d x, xn

δ xn, T

xn

ad xn, x

xn, T

xn

x, T

x . 2.11

(5)

Then

δ x, T

x

≤ 1a 1−cd

x, xn

1b 1−

xn, T xn

2.12

becauseδxn, Txnαndx0, x1δx, Tx 0⇒Tx {x}i.e.,x∈SFT. For the last part of our proof, we will show the uniqueness of the strict fixed point.

Suppose that there existx, y∈SFT. Then d

x, y δ

T x

, T y

ad x, y

x, T

x

y, T y

. 2.13

If x and y are distinct points, then we get that a ≥ 1, which contradicts our hypothesis. Thusxy. The proof is complete.

Regarding the well-posedness of a fixed point problem, we have the following result.

Theorem 2.2. LetX, dbe a complete metric space and letT :XPbXbe a multivalued operator.

Suppose there exista, b, c∈Rwithabc <1 such that δ

Tx, Ty

adx, y bδ

x, Tx

y, Ty

, ∀x, y∈X. 2.14

Then the fixed point problem is well posed forTwith respect toHd. Proof. By Reich’s theorem, we get thatSFT{x}.

LetxnX,n∈Nsuch thatHdxn, Txn→0 asn→ ∞. Then Hd

xn, T xn

δd

xn, T xn

. 2.15

We have to show thatxnxasn→ ∞. We successively have d

xn, x

δd

xn, T xn

δd

T xn

, T x

δd

xn, T xn

ad xn, x

d

xn, T xn

d

x, T x 1d

xn, T xn

ad xn, x

.

2.16

It follows that d

xn, x

≤ 1b 1−d

xn, T xn

1b 1−aHd

xn, T xn

−→0, n−→ ∞. 2.17

Hence

xn−→x, n−→ ∞. 2.18

With respect to the same multivalued operators, a data dependence result can also be established as follows.

(6)

Theorem 2.3. LetX, dbe a complete metric space and letT1, T2 :XPbXbe two multivalued operators. Suppose that

ithere exista, b, c∈Rwithabc <1 such that

δT1x, T1y≤adx, y bδx, T1x cδy, T1y, ∀x, y∈X 2.19

(denote the unique strict fixed point ofT1byx1);

ii SFT2/∅;

iiithere existsη >0 such thatδT1x, T2x≤η, for allxX.

Then

δ

x1,SFT2

≤ 1

1−a . 2.20

Proof. Letx2∈SFT2.Thenδx2, T2x2 0.

We have d

x1, x2 δ

T1

x1 , T2

x2

δ T1

x1 , T1

x2 δ

T1

x2 , T2

x2

ad x1, x2

x1, T1

x1

x2, T1

x2 η ad

x1, x2

T2

x2 , T1

x2

ηad x1, x2

1cη.

2.21

It follows that

d x1, x2

≤ 1c

1−aη. 2.22

By taking supx

2∈SFT2, it follows that δ

x1,SFT2

≤ 1c

1−aη. 2.23

LetX, dbe a complete metric space and letF1, . . . , Fm :XPXbe a finite family of multivalued operators.

The systemF F1, . . . , Fmis said to be an iterated multifunction system.

The operator

TF :PX−→PX, TFY m

i1

FiY, YPX 2.24

is called the multifractal operator generated by the iterated multifunction systemF F1, . . . , Fm.

(7)

Remark 2.4. iIfFi : XPcpXare multivaluedαi-contractions for eachi ∈ {1,2, . . . , m}, then the multifractal operatorTFis anα-contraction too, whereα:max{αi|i∈ {1, . . . , m}}

Nadler Jr.7.

iiIfFi :XPcpXare multivaluedϕi-contractionssee4for eachi∈ {1,2, . . . , m}, then the multifractal operatorTFis anϕ-contraction too, see Andres and Fiˇser4for the definitions and the result.

iiiIfF F1, . . . , Fmis an iterated multifunction system, such thatFi:XPcpXis upper semicontinuous for eachi∈ {1, . . . , m}, then the multifractal operator

TF :PcpX−→PcpX, TFY m

i1

FiY 2.25

is well defined. A fixed pointYPcpXofTF is called an attractor of the iterated multi- function systemF.

The following result is well known, see, for example, Granas and Dugundji11.

Lemma 2.5. LetX, dbe a complete metric space,x0X,r >0 and

B:B x0, r

xX|d x, x0

r

. 2.26

Letf:BXbe anα-contraction.

Ifdx0, fx0≤1−αr, thenfhas a unique fixed point inB.

Our next result concerns with the existence of an attractor for an iterated multifunction system.

Theorem 2.6. LetX, dbe a complete metric space,x0Xandr >0. LetFi :Bx 0, rPcpX, i∈ {1, . . . , m}a finite family of multivalued operators.

Suppose that

iFiis anαi-contraction, for eachi∈ {1, . . . , m};

iiδx0, Fix0≤1−max{αi|i∈ {1, . . . , m}}r, for alli∈ {1, . . . , m}.

Then there existsYB{x 0}, r ⊂ PcpXa unique attractor of the iterated multifunction systemF F1, . . . , Fm.

Proof. SinceFi:Bx 0, rPcpXis anαi-contraction, for eachi∈ {1, . . . , m}it follows thatFi

is upper semicontinuous, for eachi∈ {1, . . . , m}. ByRemark 2.4iii, we get that the operator TF :B{x 0}, r⊂PcpX→PcpX,TFY m

i1FiY,YB{x 0}, ris well defined.

Any fixed point YB{x 0}, r ⊂ PcpX of TF is an attractor of the iterated multifunction systemF F1, . . . , Fm.

Notice first that, ifYB{x 0}, r ⊂ PcpX, H, thenH{x0}, Y≤ r, which implies thatdx0, yr, for allyY. ThusyBx 0, r, for allyY.

(8)

We will show thatTFsatisfies the following two conditions:

iTF is anα-contraction, withα:max{αi|i∈ {1, . . . , m}}, that is, H TF

Y1

,TF

Y2

αH Y1, Y2

,Y1, Y2B x0

, r

PcpX; 2.27

iiH{x0},TF{x0}≤1−αr.

Indeed, we have

iLet Y1, Y2B{x 0}, r ⊂ PcpXs¸iuTFY1. By the definition of TF, it follows that there existsj ∈ {1, . . . , m}and there existsy1Y1such thatuFjy1. Since Y1, Y2PcpX, there existsy2Y2such thatdy1, y2HY1, Y2.

Since, for arbitraryε >0 and eachA, BPcpXwithHA, Bε,we have that for all aAthere existsbBsuch thatda, bε, by the following relations

H Fj

y1

, Fj

y2

αjd y1, y2

αjH Y1, Y2

, 2.28

we obtain that foruFjy1TFY1, there existsvFjy2TFY2such thatdu, vαjHY1, Y2αHY1, Y2.

By the above relation and by the similar onewhere the roles ofTFY1andTFY2are reversed, the first conclusion follows.

iiWe have to show that δ

x0

,TF

x0

≤1−αr 2.29

or equivalently for alluTF{x0},we havedx0, u≤1−αr. SinceuTF{x0} it follows that there existsj∈ {1, . . . , m}such thatuFjx0. Then

d x0, u

δ x0, Fj

x0

≤1−αr. 2.30

By Lemma 2.5, applied to TF, we get that there exists YB{x 0}, r ⊂ PcpX a unique fixed point for TF, that is, a unique attractor of the iterated multifunction system F F1, . . . , Fm. The proof is complete.

Remark 2.7. An interesting extension of the above results could be the case of a set endowed with two metrics, see12for other details.

References

1 G. Mot¸, A. Petrus¸el, and G. Petrus¸el, Topics in Multivalued Analysis and Applications to Mathematical Economics, House of the Book of Science, Cluj-Napoca, Romania, 2007.

2 A. Petrus¸el and I. A. Rus, “Well-posedness of the fixed point problem for multivalued operators,” in Applied Analysis and Differential Equations, O. Cˆarj˘a and I. I. Vrabie, Eds., pp. 295–306, World Scientific, Hackensack, NJ, USA, 2007.

(9)

3 A. Petrus¸el, I. A. Rus, and J.-C. Yao, “Well-posedness in the generalized sense of the fixed point problems for multivalued operators,” Taiwanese Journal of Mathematics, vol. 11, no. 3, pp. 903–914, 2007.

4 J. Andres and J. Fiˇser, “Metric and topological multivalued fractals,” International Journal of Bifurcation and Chaos, vol. 14, no. 4, pp. 1277–1289, 2004.

5 M. F. Barnsley, “Lecture notes on iterated function systems,” in Chaos and Fractals (Providence, RI, 1988), vol. 39 of Proceedings of Symposia in Applied Mathematics, pp. 127–144, American Mathematical Society, Providence, RI, USA, 1989.

6 J. E. Hutchinson, “Fractals and self-similarity,” Indiana University Mathematics Journal, vol. 30, no. 5, pp. 713–747, 1981.

7 S. B. Nadler Jr., “Multi-valued contraction mappings,” Pacific Journal of Mathematics, vol. 30, no. 2, pp.

475–488, 1969.

8 S. Reich, “Fixed points of contractive functions,” Bollettino dell’Unione Matematica Italiana, vol. 5, pp.

26–42, 1972.

9 C. Chifu and G. Petrus¸el, “Existence and data dependence of fixed points and strict fixed points for contractive-type multivalued operators,” Fixed Point Theory and Applications, vol. 2007, Article ID 34248, 8 pages, 2007.

10 I. A. Rus, A. Petrus¸el, and G. Petrus¸el, “Fixed point theorems for set-valuedY-contractions,” in Fixed Point Theory and Its Applications, vol. 77 of Banach Center Publications, pp. 227–237, Polish Academy of Sciences, Warsaw, Poland, 2007.

11 A. Granas and J. Dugundji, Fixed Point Theory, Springer Monographs in Mathematics, Springer, New York, NY, USA, 2003.

12 A. Petrus¸el and I. A. Rus, “Fixed point theory for multivalued operators on a set with two metrics,”

Fixed Point Theory, vol. 8, no. 1, pp. 97–104, 2007.

参照

関連したドキュメント

To this purpose, we use fixed point theory, applying results such as the well- known fixed point theorem of Tarski, presenting some results regarding the existence of extremal

C˘adariu and Radu applied the fixed point method to the investigation of Cauchy and Jensen functional equations.. In this paper, we will adopt the idea of C˘adariu and Radu to prove

The purpose of this paper is to obtain some fixed point and common fixed point results of comparable maps satisfying certain contractive conditions on partially ordered cone

We investigate the approximate fixed point theorems for Reich op- erator [7] which in turn generalize the results of Berinde [1] showing that the existence of approximate fixed

The purpose of this paper is to prove two theorems which generalize the corresponding results of Khojesteh et al [1]. Keywords: Common Fixed Points,

Since, for contraction mappings, the unique fixed point may be computed using an iteration scheme, Hutchinson’s theorem has given rise to the computation of self-similar sets

The purpose of this paper is to use a fixed point approach to obtain asymptotic stability results of a nonlinear neutral differential equation with variable delays.. An

In this paper, considering the fact that Hardy-Rogers type operator is a ´Ciri´c type operator (however the reverse need not be true), we introduce - Hardy Roger type contraction