This article was downloaded by: [Chiang Mai University]

On: 18 February 2012, At: 07:17

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Optimization: A Journal of
Mathematical Programming and
Operations Research
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/gopt20

A hybrid extragradient method

extended to fixed point problems and
equilibrium problems
a
Pham Ngoc Anh
a
Department of Scientific Fundamentals, Posts and
Telecommunications Institute of Technology, Hanoi, Vietnam

Available online: 19 Aug 2011

To cite this article: Pham Ngoc Anh (2011): A hybrid extragradient method extended to fixed point
problems and equilibrium problems, Optimization: A Journal of Mathematical Programming and
Operations Research, DOI:10.1080/02331934.2011.607497

To link to this article: http://dx.doi.org/10.1080/02331934.2011.607497

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-

conditions

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation
that the contents will be complete or accurate or up to date. The accuracy of any
instructions, formulae, and drug doses should be independently verified with primary
sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.
 Optimization
2011, 1–13, iFirst

A hybrid extragradient method extended to fixed point problems

and equilibrium problems
Pham Ngoc Anh*

Department of Scientific Fundamentals, Posts and Telecommunications

Institute of Technology, Hanoi, Vietnam
(Received 11 July 2010; final version received 18 July 2011)
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

In this article, we present a new hybrid extragradient iteration method for

finding a common element of the set of fixed points of a nonexpansive
mapping and the set of solutions of equilibrium problems for a
pseudomonotone and Lipschitz-type continuous bifunction. We obtain
strongly convergent theorems for the sequences generated by these
processes in a real Hilbert space.
Keywords: equilibrium problems; pseudomonotone; Lipschitz-type
continuous; strongly convergence; extragradient method; nonexpansive
mapping
AMS Subject Classifications: 2010; 65 K10; 90 C25

1. Introduction
Let H be a real Hilbert space with inner product h, i and norm kk, respectively. Let
C be a nonempty closed convex subset of H and f be a bifunction from C  C to R
such that f (x, x) ¼ 0 for all x 2 C. We consider the following equilibrium problems
(shortly EP( f, C )):
Find x 2 C such that f ðx , yÞ  0 8y 2 C:

The set of solutions of EP( f, C ) is denoted by Sol( f, C ).

If f (x, y) :¼ hF(x), y  xi for all x, y 2 C, where F is a mapping from C to H, then
problems EP( f, C ) become the following variational inequalities (shortly VI(F, C )):
Find x 2 C such that hFðx Þ, y  x i  0 8y 2 C:

We denote Sol(F, C ) the set of solutions of VI(F, C ).

The equilibrium problems EP( f, C ) are very general in the sense that they
include, as special cases, optimization problems, variational inequalities, minimax
problems, Nash equilibrium (see [2,5,7,8,13]), . . . . These problems appear frequently
in many practical problems arising, for instance, in physics, engineering, game

*Email: [email protected]

ISSN 0233–1934 print/ISSN 1029–4945 online

ß 2011 Taylor & Francis
DOI: 10.1080/02331934.2011.607497
http://www.informaworld.com
 2 P.N. Anh

theory, transportation, economics and network [9,10] and become an attractive field
for many researchers in both theory and applications [1,3,4,14,15,17,18].
Recall that a mapping T : C ! C is said to be nonexpansive, if

kTðxÞ  Tð yÞk  kx  yk 8x, y 2 C:

We denote the set of fixed points of T by Fix(T ) which is closed and convex.
For obtaining a common element of set of solutions of EP( f, C ) and the set of
fixed points of mapping T, Takahashi and Takahashi [19] first introduced an
iterative scheme by the viscosity approximation method. The sequence {xn} is
defined by
8 0
> x 2 H,
>
< 1
f ðun , yÞ þ hy  un , un  xn i  0 8y 2 C,
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

>
> r n
: nþ1
x :¼ n gðxn Þ þ ð1  n ÞTðun Þ 8n  0,
where g : C ! C is contractive. The authors showed that under certain conditions
over {n} and {rn}, sequences {xn} and {un} strongly converge to
z ¼ PrFix(T )\Sol( f,C )((g(z)), where PrC denotes the projection on C.
Recently Peng [17] introduced a new iterative scheme for finding a common
element of the sets Sol( f’, C ), Sol(F, C ), and Fix(T ) in a real Hilbert space where,
for each x, y 2 C, f’(x, y) ¼ f (x, y) þ ’( y)  ’(x) and ’ is a convex function from C
to R. The generated sequences {xk}, {yk}, {tk} and {zk} are defined by
8 0
>
> x 2 H,
>
>
>
> 1
> n n n n
> f’ ðu , yÞ þ hy  u , u  x i  0 8y 2 C,
>
> r
>
> 
n

>
>
>
> yn :¼ PrC un  n Fðun Þ ,
>
< n  
t :¼ PrC un  n Fð yn Þ ,
>
>
> zn :¼ n tn þ ð1  n ÞTðtn Þ,
>
>
>
>
>
>
> Cn :¼ fz 2 C : kzn  zk2  kxn  zk2  ð1  n Þðn  Þktn  Tðtn Þkg,
>
>
>
>
>
> Qn :¼ fz 2 H : hxn  z, x  xn i  0g,
>
: nþ1
x :¼ PrCn \Qn ðx0 Þ:

The author showed that under certain appropriate conditions imposed on {n}, {n}
and , the sequences {xn}, {un}, {tn}, {yn} and {zn} strongly converge to Pr(x0),
where  :¼ Sol( f’, C ) \ Sol(F, C ) \ Fix(T ). Other iterative algorithms for finding a
common element of the set of solutions of EP( f, C ) and Fix(T ) in a real Hilbert
space have been further developed by some authors [6,19,20,23,24]. At each iteration
of these algorithms, a regularized equilibrium problem is solved for a monotone and
Lipschitz-type continuous bifunction on C.
In this article, we introduce a new iterative process for finding a common element
of the set of fixed points of a nonexpansive mapping and the set of solutions of
equilibrium problems for a pseudomonotone, Lipschitz-type continuous bifunction.
At each iteration we only solve a strongly convex optimization problem instead of
 Optimization 3

a regularized equilibrium problem. The iterative process is given by

8 0
> x 2 C,
>
>  
>
> 1
>
> k k k 2
< y ¼ arg min k f ðx , yÞ þ 2 k y  x k : y 2 C ,
 
>
> k k 1 k 2
>
> t ¼ arg min k f ð y , yÞ þ k y  x k : y 2 C ,
>
> 2
>
:
xnþ1 :¼ n x0 þ ð1  n ÞTðtn Þ:
When f (x, y) ¼ hF(x), y  xi for all x, y 2 C, we find again the extragradient-type
method proposed by Zeng and Yao in [24] for finding a common element of the sets
Sol(F, C ) and Fix(T ).
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

2. Preliminaries
Let H be a real Hilbert space with inner product h, i and norm kk, respectively.
We list some well-known definitions and the projection which will be required in our
following analysis.
Definition 2.1 Let C be a closed convex subset in H, we denote the projection on C
by PrC(), i.e.
PrC ðxÞ ¼ arg minfk y  xk : y 2 Cg 8x 2 H:

The bifunction f : C  C ! R is said to be

. -strongly monotone on C if for each x, y 2 C, we have
f ðx, yÞ þ f ð y, xÞ  kx  yk2 ;
. monotone on C if for each x, y 2 C, we have
f ðx, yÞ þ f ð y, xÞ  0;

. pseudomonotone on C if for each x, y 2 C, we have

f ðx, yÞ  0 ) f ð y, xÞ  0;
. Lipschitz-type continuous on C with constants c1 4 0 and c2 4 0, if for each
x, y 2 C, we have
f ðx, yÞ þ f ð y, zÞ  f ðx, zÞ  c1 kx  yk2  c2 k y  zk2 :

Note that if f is pseudomonotone on C and Sol( f, C ) 6¼ ; then Sol( f, C ) is

convex [7].
The projection PrC() on C has the following basic properties:
(a) kPrC(x)  PrC( y)k  kx  yk 8x, y 2 H.
(b) kPrC(x)  PrC( y)k  hPrC(x)  PrC( y), x  yi 8x, y 2 H.
(c) hx  PrC(x), y  PrC( y)i  0 8y 2 C, x 2 H.
(d) kPrC(x)  yk2  kx  yk2  kPrC(x)  xk2 8y 2 C, x 2 H.
(e) kPrC(x)  PrC( y)k2  kx  yk2  kPrC(x)  x þ y  PrC( y)k2 8x,y 2 H.
 4 P.N. Anh

In this article, for finding a point of the set Sol( f, C ) \ Fix(T ), we assume that the
bifunction f satisfies the following conditions:
(i) f is pseudomonotone on C;
(ii) f is Lipschitz-type continuous on C;
(iii) for each x 2 C, y ° f (x, y) is convex and subdifferentiable;
(iv) Sol( f, C ) \ Fix(T ) 6¼ ; and Sol( f, C ) is closed.
Now we are in a position to describe the hybrid extragradient algorithm for
finding a common of two sets Sol( f, C ) and Fix(T ).
Algorithm 1 Initialization Choose x0 2 C, positive sequences {n} and {n} satisfy
the conditions:
8
> f2c1 n g  ð0, 1  Þ, for some  2 ð0, 1Þ,
>  
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

>
>
< n  min 1 , 1 ,
>
2c1 2c2
>
> X
1
>
>
>
: fn g  ð0, 1Þ, n ¼ 1, lim n ¼ 0:
n!1
n¼0

Step 1 Solve the strongly convex problems:

8  
>
> n 1 n 2 n
>
< y :¼ argmin 2 k y  x k þ n f ðx , yÞ : y 2 C ,
 
>
> 1
> t :¼ argmin kt  x k þ n f ð y , tÞ : t 2 C :
:
n n 2 n
2
Step 2 Set xnþ1 :¼ nx0 þ (1  n)T(tn). Increase k by 1 and go to Step 1.
In order to prove the main result in Section 3, we shall use the following lemmas
in the sequel.
LEMMA 2.2 [22] Let {an} be a sequence of nonnegative numbers satisfying the
conditions:
anþ1  ð1  n Þan þ n n 8n  0,
where {n} and {n} are sequences of real numbers such that
P Q
{n}  [0, 1] and 1n¼0 n ¼P1, or 1 n¼0 ð1  n Þ ¼ 0.
ðbÞ lim supn!1 n  0, or 1 n¼0  n  n is convergent.
Then limn!1 an ¼ 0.
Using the well-known necessary and sufficient condition for optimality of convex
programming, we have the following result.
LEMMA 2.3 [10] Let C be a convex subset of a real Hilbert space H and g : C ! R be
subdifferentiable on C. Then x is a solution to the following convex problem:
minfgðxÞ : x 2 Cg
 
if and only if 0 2 @g(x ) þ NC(x ), where @g() denotes the subdifferential of g and
NC(x) is the (outward ) normal cone of C at x 2 C.
 Optimization 5

LEMMA 2.4 [11] Assume that T is a nonexpansive self-mapping of a nonempty closed

convex subset C of a real Hilbert space H. If Fix(T ) 6¼ ;, then I  T is demiclosed; that
is, whenever {xn} is a sequence in C weakly converging to some x 2 C and the sequence
{(I  T )(xn)} strongly converges to some y,
 it follows that ðI  T ÞðxÞ
 ¼ y. Here I is the
identity operator of H.

3. Main results
In this section, we show a strong convergence theorem of sequences {xn}, {yn} and
{tn} defined by Algorithm 1 based on extragradient method (and hybrid method)
which solves the problem of finding a common element of two sets Sol( f, C ) and
Fix(T ) for a pseudomonotone, Lipschitz-type continuous bifunction f in a real
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

Hilbert space H.
LEMMA 3.1 Suppose that x 2 Sol( f, C ). Then,

ktn  x k2  kxn  x k2  ð1  2n c2 Þktn  yn k2  ð1  2n c1 Þkxn  yn k2 8n  0:

Proof It follows from Lemma 2.3 and f (x, ) is convex on C for each x 2 C that
 
n 1 n 2 n
t ¼ arg min kt  x k þ n f ð y , tÞ : t 2 C
2
if and only if
 
1 n 2
0 2 @2 n f ð y , yÞ þ k y  x k ðtn Þ þ NC ðtn Þ:
n
2
There exists w 2 @2f ( yn, tn) and w 2 NC ðtn Þ such that
0 ¼ n w þ tn  yn þ w:

After that the definition of @2f ( yn, tn) is used to get

f ð yn , tÞ  f ð yn , tn Þ  hw, t  tn i 8t 2 C:

With t ¼ x, this inequality becomes

f ð yn , x Þ  f ð yn , tn Þ  hw, x  tn i: ð1Þ

Then the definition of the normal cone NC is used to get

htn  xn , t  tn i  n hw, tn  ti 8t 2 C: ð2Þ

Substituting t ¼ x 2 C into (2), we obtain

htn  xn , x  tn i  n hw, tn  x i: ð3Þ

Using (1) and (3), we have

htn  xn , x  tn i  n f f ð yn , tn Þ  f ð yn , x Þg: ð4Þ
 6 P.N. Anh

Since x 2 Sol( f, C ), f (x, y)  0 for all y 2 C, and f is pseudomonotone on C, we have

f ( yn, x)  0. Then, (4) implies that
htn  xn , x  tn i  n f ð yn , tn Þ: ð5Þ
Now applying the Lipschitz property of f with x ¼ xn, y ¼ yn and z ¼ tn, we get

f ð yn , tn Þ  f ðxn , tn Þ  f ðxn , yn Þ  c1 k yn  xn k2  c2 ktn  yn k2 : ð6Þ

Combining (5) and (6), we have

htn  xn , x  tn i  n f f ðxn , tn Þ  f ðxn , yn Þ  c1 k yn  xn k2  c2 ktn  yn k2 g: ð7Þ

Similarly, since yn is the unique solution to the strongly convex problem
 
1
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

min k y  xn k2 þ n f ðxn , yÞ : y 2 C ,
2
we have
n f f ðxn , yÞ  f ðxn , yn Þg  hyn  xn , yn  yi 8y 2 C: ð8Þ
n
Substituting y ¼ t 2 C, we obtain
n f f ðxn , tn Þ  f ðxn , yn Þg  hyn  xn , yn  tn i: ð9Þ
From (7), (9) and

2htn  xn , x  tn i ¼ kxn  x k2  ktn  xn k2  ktn  x k2 ,

it implies that

kxn  x k2  ktn  xn k2  ktn  x k2 2hyn  xn , yn  tn i  2n c1 kxn  yn k2

 2n c2 ktn  yn k2 :
Hence

ktn  x k2  kxn  x k2  ktn  xn k2  2hyn  xn , yn  tn i þ 2n c1 kxn  yn k2

þ 2n c2 ktn  yn k2
¼ kxn  x k2  kðtn  yn Þ þ ð yn  xn Þk2  2hyn  xn , yn  tn i
þ 2n c1 kxn  yn k2 þ 2n c2 ktn  yn k2
 kxn  x k2  ktn  yn k2  kxn  yn k2 þ 2n c1 kxn  yn k2
þ 2n c2 ktn  yn k2
¼ kxn  x k2  ð1  2n c1 Þkxn  yn k2  ð1  2n c2 Þk yn  tn k2 :

Thus the lemma is proved. g

n 0
LEMMA 3.2 For each x 2 Sol( f, C ) \ Fix(T ), we obtain that kx  x k  kx  xk
 

for all n  0 and

kxn  yn k2  n kx0  x k2 þ kxn  xnþ1 kðkxn  x k þ kxnþ1  x kÞ:

 Optimization 7

Proof Now by induction, we show that

kxn  x k  kx0  x k 8n  0: ð10Þ

Indeed when n ¼ 0 the inequality is obvious. Suppose that (10) holds for n  0. Then
we have kxn  xk  kx0  xk. Combining this, Lemma 3.1, 0 5 n  minf2c11 , 2c12 g
and x 2 Fix(T ), we get

kxnþ1  x k ¼ kn x0 þ ð1  n ÞTðtn Þ  x k

 
¼ kn ðx0  x Þ þ ð1  n Þ Tðtn Þ  x k
 n kx0  x k þ ð1  n ÞkTðtn Þ  Tðx Þk
 n kx0  x k þ ð1  n Þktn  x k
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

 n kx0  x k
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
þ ð1  n Þ kxn  x k2  ð1  2n c2 Þktn  yn k2  ð1  2n c1 Þkxn  yn k2
 n kx0  x k þ ð1  n Þkxn  x k
 n kx0  x k þ ð1  n Þkx0  x k
 kx0  x k:

This shows that (10) holds for n þ 1. Therefore (10) holds for all n  0.
Using again Lemma 3.1, 0 5 n  minf2c11 , 2c12 g and x 2 Fix(T ), we get

kxnþ1  x k2 ¼ kn x0 þ ð1  n ÞTðtn Þ  x k2

 
¼ kn ðx0  x Þ þ ð1  n Þ Tðtn Þ  x k2
 n kx0  x k2 þ ð1  n ÞkTðtn Þ  Tðx Þk2
 n kx0  x k2 þ ð1  n Þktn  x k2
 n kx0  x k2 þ ð1  n Þfkxn  x k2  ð1  2n c1 Þkxn  yn k2
 ð1  2n c2 Þk yn  tn k2 g
 n kx0  x k2 þ kxn  x k2  ð1  2n c1 Þkxn  yn k2 :

Then, since {2c1n}  (0, 1  ) for some  2 (0, 1), we have

kxn  yn k2  ð1  2n c1 Þkxn  yn k2

 n kx0  x k2 þ kxn  x k2  kxnþ1  x k2
¼ n kx0  x k2
þ ðkxn  x k  kxnþ1  x kÞðkxn  x k þ kxnþ1  x kÞ
 n kx0  x k2 þ kxnþ1  xn kðkxn  x k þ kxnþ1  x kÞ:

The lemma is proved. g

THEOREM 3.3 Suppose that assumptions (i)–(iv) hold and T is nonexpansive on C.
Then the sequences {xn}, {yn} and {tn} generated by Algorithm 1 strongly converge to
 8 P.N. Anh

the same point PrSol( f,C )\Fix(T )(x0) provided

lim kxnþ1  xn k ¼ 0:
n!1

Proof For each n, (2) shows that there exists w 2 @2 f ( yn, tn) such that

htn  xn , t  tn i  n hw, tn  ti 8t 2 C:

With t ¼ yn, we have

htn  xn , yn  tn i  n hw, tn  yn i:
Combining f (x, x) ¼ 0 for all x 2 C, the latter inequality and the definition of w,
namely
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

f ð yn , tÞ  f ð yn , tn Þ  hw, t  tn i 8t 2 C,

we have
htn  xn , yn  tn i  n hw, yn  tn i
 n f f ð yn , tn Þ  f ð yn , yn Þg
¼ n f ð yn , tn Þ: ð11Þ
From (9), it follows
hyn  xn , tn  yn i  n f f ðxn , yn Þ  f ðxn , tn Þg: ð12Þ
Adding two inequalities (11) and (12), we obtain
htn  yn , yn  xn  tn þ xn i  n f f ðxn , yn Þ þ f ð yn , tn Þ  f ðxn , tn Þg:
Then, since f is Lipschitz-type continuous on C, we have
ktn  yn k2  n fc1 kxn  yn k2  c2 k yn  tn k2 g,
that is
ð1  n c2 Þktn  yn k2  n c1 kxn  yn k2 : ð13Þ
Lemma 3.2 shows that the sequence {kxn  xk} is bounded and
kxn  yn k2  n kx0  x k2 þ kxn  xnþ1 kðkxn  x k þ kxnþ1  x kÞ:
From this, limn!1 n ¼ 0 and limn!1 kxnþ1  xn k ¼ 0, it follows that
lim kxn  yn k ¼ 0: ð14Þ
n!1

Combining (13) and (14), we obtain

lim ktn  yn k ¼ 0: ð15Þ
n!1

Lemma 3.1 implies

ktn  x k  kxn  x k:
 Optimization 9

Then,

kTðxn Þ  xn k  kTðxn Þ  Tðtn Þk þ kTðtn Þ  Tð yn Þk þ kTð yn Þ  xnþ1 k þ kxnþ1  xn k

 kxn  tn k þ 2ktn  yn k þ n kTðtn Þ  x0 k þ kxnþ1  xn k
 kxn  tn k þ 2ktn  yn k þ n fkTðtn Þ  x k þ kx  x0 kg þ kxnþ1  xn k
 kxn  yn k þ 3ktn  yn k þ n fktn  x k þ kx  x0 kg þ kxnþ1  xn k
 kxn  yn k þ 3ktn  yn k þ n fkxn  x k þ kx  x0 kg þ kxnþ1  xn k:
ð16Þ
Thus, combining limn!1 n ¼ 0, limn!1 kxnþ1  xn k ¼ 0, (14)–(16), we obtain
lim kTðxn Þ  xn k ¼ 0: ð17Þ
n!1
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

Since {xn} is bounded, there exists a subsequence fxnj g of {xn} so that

lim suphx0  x , xn  x i ¼ lim hx0  x , xnj  x i, ð18Þ

n!1 j!1

where x :¼ PrSol( f,C )\Fix(T )(x0). Without loss of generality, we may further assume
that fxnj g weakly converges to x 2 H. Hence, (18) reduces to

lim suphx0  x , xn  x i ¼ hx0  x , x  x i: ð19Þ

n!1

Using Lemma 2.4, (17) and xnj * x as j ! 1, we have

 ¼ x:
TðxÞ  ð20Þ
nj
From (14), (15) and x * x as j ! 1, it follows

ynj * x,
 tnj * x as j ! 1:

Then, from (8) and assumptions of f, it follows

nj f f ðxnj , yÞ  f ðxnj , ynj Þg  hynj  xnj , ynj  yi 8y 2 C,

and when j ! 1, we have f ðx,  yÞ  0 for all y 2 C. It means that x 2 Sol ð f, CÞ.
Combining this and (20), we have
x 2 Solð f, CÞ \ FixðTÞ:
Using this and x ¼ PrSol( f,C )\Fix(T )(x0), we have

hx0  x , x  x i  0:
Thus, combining with (19), we have

lim suphx0  x , xn  x i  0: ð21Þ

n!1

Now, with x ¼ PrSol( f,C ) \ Fix(T )(x0), from Lemma 3.1 and the following
inequality

kx þ yk2  kxk2 þ 2hy, x þ yi 8x, y 2 H,

 10 P.N. Anh

it implies
kxnþ1  x k2 ¼ kn x0 þ ð1  n ÞTðtn Þ  x k2
 
¼ kð1  n Þ Tðtn Þ  x þ n ðx0  x Þk2
 ð1  n Þ2 kTðtn Þ  x k2 þ 2n hxnþ1  x , x0  x i
 ð1  n Þ2 ktn  x k2 þ 2n hxnþ1  x , x0  x i
 ð1  n Þktn  x k2 þ 2n hxnþ1  x , x0  x i
 ð1  n Þkxn  x k2 þ 2n hxnþ1  x , x0  x i:
Thus, we get
kxnþ1  x k2  ð1  n Þkxn  x k2 þ n n ,
where n :¼ 2hxnþ1  x, x0  xi. Then, from an application of Lemma 2.2 and (21),
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

it yields that limn!1 kxn  x k ¼ 0. From (14) and (15), it follows

limn!1 k yn  x k ¼ 0 and limn!1 ktn  x k ¼ 0. g

4. Applications
In this section, we discuss about two applications of Theorem 3.3 to finding a
common point of the set of fixed point of a nonexpansive mapping and the set of
solutions of a variational inequality problems (VIPs) for a monotone, Lipschitz-type
continuous mapping.
Let C be a nonempty closed convex subset of a Hilbert space H. For each pair x,
y 2 C, we define
f ðx, yÞ :¼ hFðxÞ, y  xi, ð22Þ
where F : C ! H. Then the equilibrium problems EP( f, C ) become the following
VIPs:
Find x 2 C such that hFðx Þ, x  x i  0 8x 2 C:
We denote Sol(F, C ) the set of solutions of (VIP).
LEMMA 4.1 If F is L-Lipschitz continuous on C, that is
kFðxÞ  Fð yÞk  Lkx  yk 8x, y 2 C,
then f defined by (22) is Lipschitz-type continuous with constants c1 ¼ c2 ¼ L2 on C.
Proof For each x, y, z 2 C, from (22), it follows that
f ðx, yÞ þ f ð y, zÞ  f ðx, zÞ ¼ hFðxÞ, y  xi þ hFð yÞ, z  yi  hFðxÞ, z  xi
¼ hFð yÞ  FðxÞ, y  zi
 kFðxÞ  Fð yÞkk y  zk
 Lkx  ykk y  zk
L L
  kx  yk2  k y  zk2
2 2
¼ c1 kx  yk2  c2 k y  zk2 :
Thus f is Lipschitz-type continuous on C. g
 Optimization 11

Now we use Algorithm 1 to find a common point of Sol(F, C ) \ Fix(T ). In this

case, the subproblems needed to solve at Step 1 are of the form
8  
>
> n 1 n 2 n n
>
< y :¼ arg min k y  x k þ  n hFðx Þ, y  x i : y 2 C ,
2
 
>
> 1
>
: t n
:¼ arg min kt  x n 2
k þ n hFð y n
Þ, t  xn
i : y 2 C :
2
Hence, we have
8  
>
> n 1  n n
 2  n n

>
< y ¼ arg min 2 k y  x  n Fðx Þ k : y 2 C ¼ PrC x  n Fðx Þ ,
 
>
> 1  n  2  n 
> n n n
: t ¼ arg min kt  x  n Fð y Þ k : y 2 C ¼ PrC x  n Fð y Þ :
2
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

Thus, Algorithm 1 and its convergence become the following results proposed by
Zeng and Yao [24].
THEOREM 4.2 [24] Let C be a nonempty closed convex subset of a real Hilbert space
H. Let F : C ! H be a monotone, L-Lipschitz continuous mapping and T : C ! C be a
nonexpansive mapping such that Fix(T ) \ Sol(F, C ) 6¼ ;. Let {xn}, {yn} be the
sequences generated by
8 0
< x ¼ x 2 H, 
yn ¼ PrC xn  n Fðxn Þ ,   ð23Þ
: nþ1
x ¼ n x0 þ ð1  n ÞTPrC xn  n Fð yn Þ ,
where {n} and {n} satisfy the conditions:
(a) {Ln}  (0, 1 P) for some  2 (0, 1).
(b) fn g  ð0, 1Þ, 1n¼0 n ¼ 1, limn!1 n ¼ 0:

Then the sequences {xn}, {yn} strongly converge to the same point PrFix(T )\Sol(F,C )(x0)
provided
lim kxnþ1  xn k ¼ 0:
n!1

As in Nadezhkina and Takahashi [16] when C ¼ H we give the following

application of Theorem 3.3.
THEOREM 4.3 Let H be a real Hilbert space. Let F : H ! H be a monotone,
L-Lipschitz continuous mapping and T : H ! H be a nonexpansive mapping such that
Fix(T ) \ F1(0) 6¼ ;. Let {xn} be a sequence generated by
8 0
< x ¼ x 2 H,
yn ¼ xn  n Fðxn Þ,   ð24Þ
: nþ1
x ¼ n x0 þ ð1  n ÞT xn  n Fð yn Þ 8n  0,
where {n} and {n} satisfy the conditions:
(a) {Ln}  (0, 1 P) for some  2 (0, 1).
(b) fn g  ð0, 1Þ, 1n¼0 n ¼ 1, limn!1 n ¼ 0:

Then the sequence {xn} strongly converges to PrFixðFÞ\F1 ð0Þ ðx0 Þ provided
lim kxnþ1  xn k ¼ 0:
n!1
 12 P.N. Anh

Proof For f(x, y) ¼ hF(x), y  xi, we have F1(0) ¼ Sol(F, H) and PrH ¼ I.
By Theorem 3.3, we have the desired results. g
Remark 1 If f ¼ 0 then Sol( f, C ) \ Fix(T ) ¼ Fix(T ) and the sequence {xn} of
Algorithm 1 becomes a iteration sequence first introduced by Halpern in [12] as
follows:
x0 2 H, xnþ1 ¼ n x0 þ ð1  n ÞTðxn Þ:

Under assumptions on parameters n, Wittmann in [21] proved that {xn} strongly
converges to a fixed point PrFix(T )(x0) of a nonexpansive mapping T. By using this
fixed point technique and the extragradient method for finding a common point of a
VIP and a nonexpansive mapping, Zeng and Yao proposed the iteration scheme (23),
Nadezhkina and Takahashi also proposed the iteration scheme (24) and other
schemes [19,23]. At each iteration n in all of these algorithms, it requires the
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

assumption that xnþ1  xn ! 0 or solving approximation auxiliary equilibrium

problems for finding a common solution of an equilibrium problem and a fixed point
problem. At each main iteration n of Algorithm 1, we only solve strongly convex
problems on C, but the proof of convergence has been done still under the
assumptions that xnþ1  xn ! 0.

Acknowledgements
The author thank the referee for his/her useful comments, remarks and important suggestions
on this article. This work was supported by the Vietnam National Foundation for Science
Technology Development (NAFOSTED).

References

[1] P.N. Anh, An LQP regularization method for equilibrium problems on polyhedral,
Vietnam J. Math. 36 (2008), pp. 209–228.
[2] P.N. Anh, A logarithmic quadratic regularization method for solving pseudomonotone
equilibrium problems, Acta Math. Vietnam. 34 (2009), pp. 183–200.
[3] P.N. Anh and J.K. Kim, Outer approximation algorithms for pseudomonotone equilibrium
problems, Comput. Math. Appl. 61 (2011), pp. 2588–2595.
[4] P.N. Anh, L.D. Muu, V.H. Nguyen, and J.J. Strodiot, Using the Banach contraction
principle to implement the proximal point method for multivalued monotone variational
inequalities, J. Optim. Theory Appl. 124 (2005), pp. 285–306.
[5] P.N. Anh, L.D. Muu, and J.J Strodiot, Generalized projection method for non-Lipschitz
multivalued monotone variational inequalities, Acta Math, Vietnam. 34 (2009), pp. 67–79.
[6] P.N. Anh and D.X. Son, A new iterative scheme for pseudomonotone equilibrium problems
and a finite family of pseudocontractions, J. Appl. Math. Inform. (2011) (accepted).
[7] E. Blum and W. Oettli, From optimization and variational inequality to equilibrium
problems, Math. Stud. 63 (1994), pp. 127–149.
[8] S. Dafermos, Exchange price equilibria and variational inequalities, Math. Program. 46
(1990), pp. 391–402.
[9] P. Daniele, F. Giannessi, and A. Maugeri, Equilibrium Problems and Variational Models,
Kluwer Academic Publisher, Dordrecht, The Netherlands, 2003.
[10] F. Facchinei and J.S. Pang, Finite-dimensional Variational Inequalities and Complementary
Problems, Springer-Verlag, New York, 2003.
 Optimization 13

[11] K. Goebel and W.A. Kirk, Topics on Metric Fixed Point Theory, Cambridge University
Press, Cambridge, England, 1990.
[12] B. Halpern, Fixed points of nonexpanding maps, Bull. of American Mathematical Society
73 (1967), pp. 957–961.
[13] I.V. Konnov, Application of the proximal point method to nonmonotone equilibrium
problems, J. Optim. Theory Appl. 119 (2003), pp. 317–333.
[14] G. Mastroeni, Gap function for equilibrium problems, J. Global Optim. 27 (2004),
pp. 411–426.
[15] A. Moudafi, Proximal point algorithm extended to equilibrium problem, J. Natur. Geom.
15 (1999), pp. 91–100.
[16] N. Nadezhkina and W. Takahashi, Weak convergence theorem by an extragradient method
for nonexpansive mappings and monotone mappings, J. Optim. Theory Appl. 128 (2006),
pp. 191–201.
[17] J.W. Peng, Iterative algorithms for mixed equilibrium problems, strict pseudocontractions
and monotone mappings, J. Optim. Theory Appl. 144 (2010), pp. 107–119.
Downloaded by [Chiang Mai University] at 07:17 18 February 2012

[18] T.D. Quoc, L.D. Muu, and N.V. Hien, Extragradient algorithms extended to equilibrium
problems, Optimization 57 (2008), pp. 749–776.
[19] S. Takahashi and W. Takahashi, Viscosity approximation methods for equilibrium
problems and fixed point problems in Hilbert spaces, J. Math. Anal. Appl. 331 (2007),
pp. 506–515.
[20] S. Wang and B. Guo, New iterative scheme with nonexpansive mappings for equilibrium
problems and variational inequality problems in Hilbert spaces, J. Comput. Appl. Math.
233 (2010), pp. 2620–2630.
[21] R. Wittmann, Approximation of fixed points of nonexpansive mappings, Arch. Math. 58
(1992), pp. 486–491.
[22] H.K. Xu and T.H. Kim, Convergence of hybrid steepest-descent methods for variational
inequalities, J. Optim. Theory Appl. 119 (2003), pp. 185–201.
[23] Y. Yao, Y.C. Liou, and Y.J Wu, An extragradient method for mixed equilibrium problems
and fixed point problems, Fixed Point Theory Appl. 2009 (2009), Article ID 632819
(15 pages, doi:10.1155/2009/632819).
[24] L.C. Zeng and J.C. Yao, Strong convergence theorem by an extragradient method for fixed
point problems and variational inequality problems, Taiwanese J. Math. 10 (2006),
pp. 1293–1303.