• No results found

Strong Convergence Theorem by Monotone Hybrid Algorithm for Equilibrium Problems, Hemirelatively Nonexpansive Mappings, and Maximal Monotone Operators

N/A
N/A
Protected

Academic year: 2020

Share "Strong Convergence Theorem by Monotone Hybrid Algorithm for Equilibrium Problems, Hemirelatively Nonexpansive Mappings, and Maximal Monotone Operators"

Copied!
12
0
0

Loading.... (view fulltext now)

Full text

(1)

Volume 2008, Article ID 617248,12pages doi:10.1155/2008/617248

Research Article

Strong Convergence Theorem by Monotone

Hybrid Algorithm for Equilibrium Problems,

Hemirelatively Nonexpansive Mappings, and

Maximal Monotone Operators

Yun Cheng and Ming Tian

College of Science, Civil Aviation University of China, Tianjin 300300, China

Correspondence should be addressed to Yun Cheng,[email protected]

Received 17 June 2008; Accepted 11 November 2008

Recommended by Wataru Takahashi

We introduce a new hybrid iterative algorithm for finding a common element of the set of fixed points of hemirelatively nonexpansive mappings and the set of solutions of an equilibrium problem and for finding a common element of the set of zero points of maximal monotone operators and the set of solutions of an equilibrium problem in a Banach space. Using this theorem, we obtain three new results for finding a solution of an equilibrium problem, a fixed point of a hemirelatively nonexpnasive mapping, and a zero point of maximal monotone operators in a Banach space.

Copyrightq2008 Y. Cheng and M. Tian. 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

LetEbe a Banach space, letCbe a closed convex subset ofE, and letf be a bifunction from

C×CtoR, whereRis the set of real numbers. The equilibrium problem is to find

xC such thatfx, y0 yC. 1.1

The set of such solutionsx∗is denoted by EPf.

(2)

Banach space. Very recently, Su et al.3proved the following theorem by a monotone hybrid method.

Theorem 1.1see Su et al.3. LetEbe a uniformly convex and uniformly smooth real Banach space, letCbe a nonempty closed convex subset ofE, and letT :CCbe a closed hemirelatively nonexpansive mapping such that FT/. Assume that αn is a sequence in 0,1 such that lim supn→ ∞αn <1. Define a sequencexninCby the following:

x0∈C, chosen arbitrarily,

ynJ−1αnJxn 1−αnJTxn,

Cn{zCn−1∩Qn−1:φz, ynφz, xn}, C0 {zC:φz, y0≤φz, x0}, Qn{zCn−1∩Qn−1:xnz, Jx0−Jxn ≥0},

Q0C, xn1 ΠCn∩Qnx0,

1.2

where J is the duality mapping on E. Then,xn converges strongly to ΠFTx0, where ΠFT is the generalized projection fromContoFT.

In this paper, motivated by Su et al. 3, we prove a strong convergence theorem

for finding a common element of the set of solutions of an equilibrium problem and the set of fixed points of a hemirelatively nonexpansive mapping and for finding a common element of the set of zero points of maximal monotone operators and the set of solutions of an equilibrium problem in a Banach space by using the monotone hybrid method. Using this theorem, we obtain three new strong convergence results for finding a solution of an equilibrium problem, a fixed point of a hemirelatively nonexpnasive mapping, and a zero point of maximal monotone operators in a Banach space.

2. Preliminaries

LetEbe a real Banach space with dualE∗. We denote byJthe normalized duality mapping fromEto 2E∗defined by

JxfE:x, fx2 f2, 2.1

where·,·denotes the generalized duality pairing. It is well known that ifE∗ is uniformly convex, thenJ is uniformly continuous on bounded subsets ofE. In this case, J is single valued and also one to one.

LetEbe a smooth, strictly convex, and reflexive Banach space and letCbe a nonempty closed convex subset ofE. Throughout this paper, we denote byφthe function defined by

(3)

Following Alber4, the generalized projectionΠC :ECfromEontoCis defined by

ΠCx arg min

y∈Cφy, xxE. 2.3

The generalized projectionΠCfromEontoCis well defined and single valued, and it satisfies

xy2φy, xxy2 x, yE. 2.4

IfEis a Hilbert space, thenφy, x yx2andΠCis the metric projection ofEontoC.

IfEis a reflexive strict convex and smooth Banach space, then forx, yE, φx, y 0 if and only ifx y. It is sufficient to show that ifφx, y 0, thenx y. From2.4,

we havex y. This impliesx, Jy x2 Jy2. From the definition ofJ, we have JxJy, that is,xy.

Let C be a closed convex subset of E and let T be a mapping from C into itself. We denote by FT the set of fixed points ofT. T is called hemirelatively nonexpansive if

φp, Txφp, xfor allxCandpFT.

A pointpinCis said to be an asymptotic fixed point ofT5ifCcontains a sequencexn which converges weakly topsuch that the strong limn→ ∞Txnxn 0. The set of asymptotic fixed points ofT will be denoted byFT. A hemirelatively nonexpansive mappingTfromC into itself is called relatively nonexpansive1,5,6ifFT FT.

We need the following lemmas for the proof of our main results.

Lemma 2.1see Alber4. LetCbe a nonempty closed convex subset of a smooth, strictly convex, and reflexive Banach spaceE. Then,

φx,ΠCΠCy, yφx, yxC, yE. 2.5

Lemma 2.2see Alber4. LetCbe a nonempty closed convex subset of a smooth, strictly convex, and reflexive Banach space, letxE, and letzC. Then,

z ΠCx⇐⇒ yz, JxJz ≤0 ∀yC. 2.6

Lemma 2.3see Kamimura and Takahashi7. LetEbe a smooth and uniformly convex Banach space and let {xn} and {yn} be sequences in E such that either {xn} or {yn} is bounded. If limn→ ∞φxn, yn 0. Thenlimn→ ∞xnyn0.

Lemma 2.4see Xu8. LetEbe a uniformly convex Banach space and letr >0. Then, there exists a strictly increasing, continuous, and convex functiong:0,2rRsuch thatg0 0and

tx 1−ty2tx2 1ty2t1tgxy x, yB

r, t∈0,1, 2.7

(4)

Lemma 2.5see Kamimura and Takahashi7. LetEbe a smooth and uniformly convex Banach space and let r > 0. Then, there exists a strictly increasing, continuous, and convex functiong : 0,2rRsuch thatg0 0and

gxyφx, yx, yBr. 2.8

For solving the equilibrium problem, let us assume that a bifunctionf satisfies the following conditions:

A1fx, x 0 for all xC;

A2f is monotone, that is,fx, y fy, x≤0 for all x, yC; A3for all x, y, zC, lim supt0ftz 1−tx, yfx, y; A4for all xC, fx,·is convex.

Lemma 2.6see Blum and Oettli9. LetCbe a closed convex subset of a smooth, strictly convex, and reflexive Banach spaceE, letf be a bifunction fromC×CtoRsatisfying (A1)–(A4), letr >0, and letxE. Then, there existszCsuch that

fz, y 1

ryz, JzJx ≥0 ∀yC. 2.9

Lemma 2.7see Takahashi and Zembayashi10. LetCbe a closed convex subset of a uniformly smooth, strictly convex, and reflexive Banach spaceE, letfbe a bifunction fromC×CtoRsatisfying (A1)–(A4), and letxE, forr >0. Define a mappingTr :E → 2C as follows:

Trx

zC:fz, y 1

ryz, JzJx ≥0 ∀yC

xE. 2.10

Then, the following holds:

1Tr is single valued;

2Tr is a firmly nonexpansive-type mapping11, that is, for allx, yE,

TrxTry, JTrxJTryTrxTry, JxJy; 2.11

3FTr FTr Epf; 4Epfis closed and convex.

Lemma 2.8see Takahashi and Zembayashi10. LetCbe a closed convex subset of a smooth, strictly convex, and reflexive Banach spaceEand letf be a bifunction fromC×CtoR satisfying (A1)–(A4). Then, forr >0andxE, andqFTr,

(5)

Lemma 2.9see Su et al.3. LetEbe a strictly convex and smooth real Banach space, letCbe a closed convex subset ofE, and letT be a hemirelatively nonexpansive mapping fromCinto itself. Then,FTis closed and convex.

Recall that an operatorTin a Banach space is called closed, ifxnx, Txny, then

Txy.

3. Strong convergence theorem

Theorem 3.1. Let Ebe a uniformly convex and uniformly smooth real Banach space, let Cbe a nonempty closed convex subset ofE, letfbe a bifunction fromC×CtoRsatisfying (A1)–(A4), and letT:CCbe a closed hemirelatively nonexpansive mapping such thatFT∩EPf/. Define a sequence{xn}inCby the following:

x0∈C, chosen arbitrarily,

ynJ−1αnJxn 1−αnJTzn,

znJ−1βnJxn 1−βnJTxn,

unCsuch thatfun, y r1

nyun, JunJyn ≥0,yC,

Cn{zCn−1∩Qn−1:φz, unφz, xn}, C0 {zC:φz, u0≤φz, x0},

Qn{zCn−1∩Qn−1:xnz, Jx0−Jxn ≥0},

Q0C, xn1 ΠCn∩Qnx0,

3.1

for everynN∪ {0}, whereJis the duality mapping onE, {αn}, {βn}are sequences in0,1such thatlim infn→ ∞1−αnβn1βn > 0 and{rn} ⊂ a,for somea > 0. Then,{xn}converges strongly toΠFT∩EPfx0, whereΠFT∩EPfis the generalized projection ofEontoFT∩EPf.

Proof. First, we can easily show thatCnandQnare closed and convex for eachn≥0.

Next, we show thatFT∩EPfCn for alln ≥ 0. LetuFT∩EPf. Putting

unTrnynfor allnN, fromLemma 2.8, we haveTrn relatively nonexpansive. SinceTrnare relatively nonexpansive andT is hemirelatively nonexpansive, we have

φu, zn φu, J−1β

nJxn 1−βnJTxn u22u, β

nJxn 1−βnJTxnβnJxn 1−βnJTxn2

u22β

nu, Jxn −21−βnu, JTxnβnxn2 1−βnTxn2 βnφu, xn 1βnφu, Txn

φu, xn,

φu, un φu, Trnynφu, ynαnφu, xn 1αnφu, znφu, xn.

(6)

Hence, we have

FT∩EPfCnn≥0. 3.3

Next, we show thatFT∩EPfQnfor alln ≥0. We prove this by induction. Forn0, we have

FT∩EPfQ0C. 3.4

Suppose thatFT∩EPfQn, byLemma 2.2, we have

xn1−z, Jx0−Jxn1 ≥0 ∀zCnQn. 3.5

As FT∩EPfCnQn, by the induction assumptions, the last inequality holds, in particular, for allzFT∩EPf. This, together with the definition ofQn1, implies that FT∩EPfQn1. So,{xn}is well defined.

Sincexn1 ΠCn∩Qnx0andCnQnCn−1∩Qn−1for alln≥1, we have

φxn, x0≤φxn1, x0 ∀n≥0. 3.6

Therefore,{φxn, x0}is nondecreasing. In addition, from the definition ofQnandLemma 2.2,

xn ΠQnx0. Therefore, for eachuFT∩EPf, we have

φxn, x0 φ

ΠQnx0, x0

φu, x0−φu, xnφu, x0. 3.7

Therefore,φxn, x0and{xn}are bounded. This, together with3.6, implies that the limit of

{φxn, x0}exists. FromLemma 2.1, we have, for any positive integerm,

φxnm, xnφxnm,ΠQnx0

φxnm, x0−φ

ΠQnx0, x0

φxnm, x0−φxn, x0 ∀n≥0.

3.8

Therefore,

lim

n→ ∞φxnm, xn 0. 3.9

From3.9, we can prove that{xn}is a Cauchy sequence. Therefore, there exists a pointxC such that{xn}converges strongly tox.

Sincexn1 ∈Cn, we have

φxn1, unφxn1, xn. 3.10

Therefore, we have

(7)

FromLemma 2.3, we have

lim

n→ ∞xn1−unnlim→ ∞xn1−xn0. 3.12

So, we have

lim

n→ ∞xnun0. 3.13

SinceJis uniformly norm-to-norm continuous on bounded sets, we have

lim

n→ ∞JxnJun0. 3.14

Letr supn∈N{xn,Txn}.SinceEis a uniformly smooth Banach space, we know thatE∗ is a uniformly convex Banach space. Therefore, fromLemma 2.4, there exists a continuous, strictly increasing, and convex functiongwithg0 0, such that

αx 1αy2 αx2 1αy2α1αgxy 3.15

forx, y∗∈Br, andα∈0,1. So, we have that foruFT∩EPf,

φu, zn φu, J−1β

nJxn 1−βnJTxn u22u, β

nJxn 1−βnJTxnβnJxn 1−βnJTxn2

φu, xnβn1βngJxnJTxn,

φu, unαnφu, xn 1αnφu, zn

φu, xn−1−αnβn1βngJxnJTxn.

3.16

Therefore, we have

1−αnβn1βngJxnJTxnφu, xnφu, un. 3.17

Since

φu, xnφu, unxn2−un2−2u, JxnJunxnunxnun2uJxnJun, 3.18

we have

lim

n→ ∞φu, xnφu, un 0. 3.19

From lim infn→ ∞1−αnβn1βn>0, we have

lim

(8)

Therefore, from the property ofg, we have

lim

n→ ∞JxnJTxn0. 3.21

SinceJ−1is uniformly norm-to-norm continuous on bounded sets, we have

lim

n→ ∞xnTxn0. 3.22

SinceT is a closed operator andxnx, thenxis a fixed point ofT. On the other hand,

φun, yn φTrnyn, ynφu, ynφu, Trnynφu, xnφu, Trnyn φu, xnφu, un. 3.23

So, we have from3.19that

lim

n→ ∞φun, yn 0. 3.24

FromLemma 2.3, we have that

lim

n→ ∞unyn0. 3.25

Fromxnxandxnun → 0, we haveynx. From3.25, we have

lim

n→ ∞JunJyn0. 3.26

Fromrna, we have

lim n→ ∞

JunJyn

rn 0. 3.27

ByunTrnyn, we have

fun, y r1

nyun, JunJyn ≥0 ∀yC. 3.28

FromA2, we have that

1

(9)

From3.27andA4, we have

fy,x≤0 ∀yC. 3.30

Fortwith 0< t≤1 andyC, letytty 1−tx. We havefyt,x≤0. So, fromA1, we have

0fyt, yttfyt, y 1−tfyt,xtfyt, y. 3.31

Dividing byt, we have

fyt, y≥0 ∀yC. 3.32

Lettingt → 0, fromA3, we have

fx, y ≥0 ∀yC. 3.33

Therefore,x∈EPf. Finally, we prove thatx ΠFT∩EPfx0. FromLemma 2.1, we have

φx, ΠFT∩EPfx0

φΠFT∩EPfx0, x0

φx, x 0. 3.34

Sincexn1 ΠCn∩Qnx0andxFT∩EPfCnQn, for alln≥0, we get fromLemma 2.1 that

φΠFT∩EPfx0, xn1

φxn1, x0≤φ

ΠFT∩EPfx0, x0

. 3.35

By the definition of φx, y, it follows that φx, x 0 ≤ φΠFT∩EPfx0, x0 and φx, x 0 ≥ φΠFT∩EPfx0, x0, whence φx, x 0 φΠFT∩EPfx0, x0. Therefore, it follows from the

uniqueness ofΠFT∩EPfx0thatx ΠFT∩EPfx0. This completes the proof.

Corollary 3.2. LetE be a uniformly convex and uniformly smooth real Banach space, let C be a nonempty closed convex subset ofE, and letfbe a bifunction fromC×CtoRsatisfying (A1)–(A4). Define a sequence{xn}inCby the following:

x0∈C, chosen arbitrarily,

unCsuch thatfun, y r1

nyun, JunJxn ≥0,yC,

Cn{zCn−1∩Qn−1:φz, unφz, xn}, C0 {zC:φz, u0≤φz, x0}, Qn{zCn−1∩Qn−1:xnz, Jx0−Jxn ≥0},

Q0C, xn1 ΠCn∩Qnx0,

(10)

for everynN∪ {0}, whereJis the duality mapping onEand{rn} ⊂a,for somea >0. Then,

{xn}converges strongly toΠEPfx0.

Proof. PuttingT IinTheorem 3.1, we obtainCorollary 3.2.

Corollary 3.3. LetE be a uniformly convex and uniformly smooth real Banach space, let C be a nonempty closed convex subset ofE, and letT : CCbe a closed hemirelatively nonexpansive mapping. Define a sequence{xn}inCby the following:

x0∈C, chosen arbitrarily, ynJ−1αnJxn 1−αnJTzn,

znJ−1βnJxn 1−βnJTxn,

un ΠCyn,

Cn{zCn−1∩Qn−1:Φz, unφz, xn}, C0 {zC:φz, u0≤φz, x0}, Qn{zCn−1∩Qn−1:xnz, Jx0−Jxn ≥0},

Q0C, xn1 ΠCn∩Qnx0,

3.37

for everynN∪ {0}, whereJis the duality mapping onE, {αn}, {βn}are sequences in0,1such thatlim infn→ ∞1−αnβn1βn>0. Then,{xn}converges strongly toΠFTx0.

Proof. Putting fx, y 0 for all x, yC and rn 1 for all nin Theorem 3.1, we obtain

Corollary 3.3.

Corollary 3.4. LetE be a uniformly convex and uniformly smooth real Banach space, let C be a nonempty closed convex subset ofE, letfbe a bifunction fromC×CtoRsatisfying (A1)–(A4), and letT :CCbe a closed relatively nonexpansive mapping such thatFT∩EPf/. Define a sequence{xn}inCby the following:

x0∈C, chosen arbitrarily,

ynJ−1αnJxn 1−αnJTzn,

znJ−1βnJxn 1−βnJTxn,

unCsuch thatfun, y r1

nyun, JunJyn ≥0,yC,

Cn{zCn−1∩Qn−1:φz, unφz, xn}, C0 {zC:φz, u0≤φz, x0}, Qn{zCn−1∩Qn−1:xnz, Jx0−Jxn ≥0},

Q0C, xn1 ΠCn∩Qnx0,

(11)

for everynN∪ {0}, whereJis the duality mapping onE, {αn}, {βn}are sequences in0,1such thatlim infn→ ∞1−αnβn1βn > 0 and{rn} ⊂ a,for somea > 0. Then,{xn}converges strongly toΠFT∩EPfx0.

Proof. Since every relatively nonexpansive mapping is a hemirelatively one,Corollary 3.4is implied byTheorem 3.1.

Remark 3.5see Rockafellar 12. LetEbe a reflexive, strictly convex, and smooth Banach space and letA be a monotone operator fromE to E∗. Then,A is maximal if and only if

RJrA Efor allr >0.

LetEbe a reflexive, strictly convex, and smooth Banach space and letAbe a maximal monotone operator fromEtoE∗. UsingRemark 3.5and strict convexity ofE, we obtain that for everyr > 0 andxE, there exists a uniquexrDAsuch thatJxJxr rAxr.If

Jrx xr, then we can define a single-valued mappingJr :EDAbyJr JrA−1J, and such a Jr is called the resolvent ofA. We know thatA−10 FJrfor allr > 0 andJr is relatively nonexpansive mappingsee2for more details. UsingTheorem 3.1, we can consider the problem of strong convergence concerning maximal monotone operators in a Banach space.

Theorem 3.6. Let Ebe a uniformly convex and uniformly smooth real Banach space, let Cbe a nonempty closed convex subset ofE, letfbe a bifunction fromC×CtoRsatisfying (A1)–(A4), and letJr be a resolvent ofAand a closed mapping such thatA−10∩EPf/, wherer >0. Define a sequence{xn}inCby the following:

x0∈C, chosen arbitrarily,

ynJ−1αnJxn 1−αnJJrxn,

znJ−1βnJxn 1−βnJJrxn,

unCsuch thatfun, y r1

nyun, JunJyn ≥0,yC,

Cn{zCn−1∩Qn−1:φz, unφz, xn}, C0 {zC:φz, u0≤φz, x0}, Qn{zCn−1∩Qn−1:xnz, Jx0−Jxn ≥0},

Q0C, xn1 ΠCn∩Qnx0,

3.39

for everynN∪ {0}, whereJis the duality mapping onE,{αn}is a sequences in0,1such that lim infn→ ∞1−αnβn1βn>0and{rn} ⊂a,for somea >0, Then,{xn}converges strongly toΠA−10EPfx0.

Proof. Since Jr is a closed relatively nonexpansive mapping and A−10 FJr, from

(12)

Acknowledgment

This work is supported by Tianjin Natural Science Foundation in China Grant no. 06YFJMJC12500.

References

1 C. Martinez-Yanes and H.-K. Xu, “Strong convergence of the CQ method for fixed point iteration processes,”Nonlinear Analysis: Theory, Methods & Applications, vol. 64, no. 11, pp. 2400–2411, 2006.

2 W. Takahashi and K. Zembayashi, “Strong convergence theorem by a new hybrid method for equilibrium problems and relatively nonexpansive mappings,”Fixed Point Theory and Applications, vol. 2008, Article ID 528476, 11 pages, 2008.

3 Y. Su, D. Wang, and M. Shang, “Strong convergence of monotone hybrid algorithm for hemi-relatively nonexpansive mappings,”Fixed Point Theory and Applications, vol. 2008, Article ID 284613, 8 pages, 2008.

4 Y. I. Alber, “Metric and generalized projection operators in Banach spaces: properties and applications,” inTheory and Applications of Nonlinear Operators of Accretive and Monotone Type, A. G. Kartsatos, Ed., vol. 178 ofLecture Notes in Pure and Applied Mathematics, pp. 15–50, Marcel Dekker, New York, NY, USA, 1996.

5 D. Butnariu, S. Reich, and A. J. Zaslavski, “Asymptotic behavior of relatively nonexpansive operators in Banach spaces,”Journal of Applied Analysis, vol. 7, no. 2, pp. 151–174, 2001.

6 S.-Y. Matsushita and W. Takahashi, “A strong convergence theorem for relatively nonexpansive mappings in a Banach space,”Journal of Approximation Theory, vol. 134, no. 2, pp. 257–266, 2005.

7 S. Kamimura and W. Takahashi, “Strong convergence of a proximal-type algorithm in a Banach space,”SIAM Journal on Optimization, vol. 13, no. 3, pp. 938–945, 2002.

8 H. K. Xu, “Inequalities in Banach spaces with applications,”Nonlinear Analysis: Theory, Methods & Applications, vol. 16, no. 12, pp. 1127–1138, 1991.

9 E. Blum and W. Oettli, “From optimization and variational inequalities to equilibrium problems,”The Mathematics Student, vol. 63, no. 1–4, pp. 123–145, 1994.

10 W. Takahashi and K. Zembayashi, “Strong and weak convergence theorems for equilibrium problems and relatively nonexpansive mappings in Banach spaces,” Nonlinear Analysis: Theory, Methods & Applications, vol. 70, no. 1, pp. 45–57, 2009.

11 F. Kohsaka and W. Takahashi, “Existence and approximation of fixed points of firmly nonexpansive type mappings in Banach spaces,” to appear inSIAM Journal on Optimization.

References

Related documents

The present study also shows that MH84 reversed mitochondrial dysfunction and PGC-1 α levels, de- creased β -secretase processing of A β PP, and reduced levels of soluble A β 40

In addition to the sampling properties, one result which becomes obvious from the present analysis is that subnets of the same size can differ quite considerably; and, in par-

Conclusions: We show that α -synuclein aggregates induce microglial activation and demonstrate for the first time that galectin-3 plays a significant role in microglia

In conclusion, we successfully applied NGS to the genetic analysis of drug resistance in MTB, as well as in suscepti- ble strains. The Ion AmpliSeq Tb Panel showed satisfac-

This preliminary clinical study aimed to examine whether next‑generation sequencing (NGS) targeting 16S rDNA, which was PCR‑ amplified directly from the tip of a central

The EQUATOR website provides an up to date centralized resource for: researchers writing up their studies (notably guidance on reporting, scientific writing, ethical conduct in

Using data obtained from the Alzheimer ’ s Disease Neuroimaging Initiative (ADNI), the purpose of this study is to determine whether prediction of development of clinical dementia

Five core-top samples were examined for CellTracker Green-labeled nematodes: one L ’ Atalante upper halocline pushcore, one Discovery mid-halocline pushcore, two Ur- ania