235
APPROXIMATING COMMON ELEMENTS OF THE SET OF AMENABLE SEMIGROUP AND ZERO POINT SETS AND THE SOLUTIONS SETS OF SYSTEMS OF EQUILIBRIUM PROBLEMS
Qiang Li* & Dapeng Gao
Department of Mathematics and Information, China West Normal University, Sichuan 637009, P. R. China
*Email: [email protected]
ABSTRACT
In this paper, we introduce a new iterative scheme for finding a common element of the set of fixed point of left amenable semigroup, the zero point set of the operator which is the sum of inverse strongly monotone operators and maximal monotone operators, and the set of solutions for systems of equilibrium problems in Hilbert spaces by using a hybrid steepest descent methods. Then strong convergence of the scheme to a common element of the three sets is proved. Our results improve and generalize some well-known results in the literature.
Keywords: Common fixed point; non-expansive mapping; amenable semigroup; system of equilibrium problem;
strictly pseudo-contractive mapping; hybrid steepest descent method.
1. INTRODUCTION
Let
H
be a real Hilbert space whose inner product and norm are denoted by ,
and
, respectively. When}
{ x
n is a sequence inH
, we denote strong convergence of{ x
n}
tox H
byx
n x
and weak convergence byx
n x
. LetC
be a nonempty closed convex subset ofH
and letF : C C R
be a bi-function, whereR
is the set of real numbers. The equilibrium problem for
F : C C R
is to findx
C
such that. 0,
= ) ,
( x y y C
F
(1.1)The set of solutions of (1.1) is denoted by
EP (F )
. Let{ F
i, i = 1,2, , N }
be a finite family of bi-functions fromC C
toR
, whereR
is the set of real numbers. The system of equilibrium problems for{ F
1, F
2, , F
N}
is to find a common element
x
C
such that
. 0,
= ) , (
, 0,
= ) , (
, 0,
= ) , (
C y y
x F
C y y
x F
C y y
x F
N 2 1
(1.2) We denote the set of solutions of (1.2) by
Mj=1SEP ( F
j)
, whereSEP ( F
j)
is the set of solutions to the equilibrium problems, that is,. 0,
= ) ,
( x y y C
F
i
(1.3)If
N = 1
, then the problem (1.2) is reduced to the equilibrium problems.If
N = 1
andF ( x
, y ) = Bx
, y x
, then the problem (1.2) is reduced to the variational inequality problem of findingx
C
such that. 0,
, y x y C
Bx
(1.4)The set of solutions of (1.4) is denoted by
VI ( C , B )
. Many problems in applied sciences, such as monotone inclusion problems, saddle point problems, optimization problems, variational inequality problems, Nash equilibrium problems, and equilibrium problems as special cases. Some methods have been proposed to solve) , ( C B
VI
,EP (F )
andSEP ( F
i)
; see, for example [1,2] and references therein. The above formulations (1.2) extends this formulism to such problems, covering in particular various forms of feasibility problems.236
Definition 1.1. Let
A : C H
be nonlinear mappings. ThenA
is called (1)monotone if Ax Ay , x y 0, x , y C
;(2)
inverse strongly monotone if there exists a constant > 0
such that; , ,
, x y Ax Ay
2x y C
Ay
Ax
(3)
-strictly pseudo-contractive if there exists a constant (0,1)
such that; , , ) ( )
(
22
2
x y I A x I A y x y C
Ay
Ax
(4)For the variational inequality, the following is true:
0.
>
), (
= )
,
( C A u P u A
VI
u
C
(5)Let
M : H 2
H be a set-valued mapping. The setD (M )
defined byD ( M ) = { x H : Mx }
is said to be the domain ofM
. The setR (M )
defined byR ( M ) =
xHMx
is said to be the range ofM
. The set) (M
G
defined byG ( M ) = {( x , y ) H H : x D ( M ), y R ( M )}
is said to be the graph ofM
. Recall thatM
is said to be monotone if).
( ) , ( ), , ( 0,
, f g x f y g G M
y
x
M
is said to be maximal monotone if it is not properly contained in any other mono-tone operator. Equivalently,M
is maximal monotone ifR ( I rM ) = H
for allr > 0
. The class of monotone mappings is one of the most important classes of mappings. With in the past several decades, many authors have been devoting to the studies on the existence and convergence of zero points for maximal monotone mappings, and the references therein. For a maximal monotone operatorM
onH
andr > 0
, we may define the single-valued resolvent) ( :
) (
= I rM
1H D M
J
r
. It is known thatJ
r is firmly non-expansive and( M )
1(0) = F ( J
r)
, where) ( J
rF
denotes the fixed point set ofJ
r.In 2003, Takahashi and Toyoda [3] proved the following weak convergence theorem.
Theorem 1.1. Let
C
be a nonempty closed convex subset of a real Hilbert space spaceH
. LetA
be an
inverse strongly monotone mapping fromC
intoH
andS
be a non-expansive mapping fromC
into itself such thatF ( S ) VI ( C , A )
. Let{ x
n}
be a sequence generated by0, ),
( ) (1
= ,
10
C x
x SP x Ax n
x
n
n n
n C n
n nwhere
n [ b a , ]
for somea , b (0,2 )
and
n [ d c , ]
for somec , d (0,1)
. Then{ x
n}
converges weakly toz F ( S ) VI ( C , A )
, wherez = lim
nP
F(S)VI(C,A)x
n.In 2007, Tada and Takahashi [4] obtained the following weak convergence theorem.
Theorem 1.2. Let
C
be a nonempty closed convex subset of a real Hilbert space spaceH
. LetF
be a bi- function fromC C
toR
satisfyingA (1) A (4)
andS
be a non-expansive mapping fromC
intoH
such thatF ( S ) EP ( F )
. Let{ x
n}
and{ u
n}
be sequences generated byx
1= x H
and let
n n n n
n
n n n n
n n
Su x
x
C u x
u u r u u u F that such C
u
) (1
=
, 0,
1 , ) , (
1
(1.5) for each
n 1
, where
n [ b a , ]
for somea , b (0,1)
andr
n (0, )
satisfiesliminf
nr
n> 0
. Then}
{ x
n converges weakly tow F ( S ) EP ( F )
, wherez = lim
nP
F(S)EP(F))x
n.Very recently,Jitpeera et al.[5],introduced the iterative scheme based on viscosity and Cesro mean
237
, 0,
1 ) 1 )
((1 )
(
=
), (
) (1
=
, 0,
1 , ) ( ) ( ) , (
0
=
1
T y n
A n I
x y
rf x
Bu u
P u
y
C y x
u u r y
u y y
u
n i n
i n
n n
n n n n
n n n C n n
n n
n n n n
n n
(1.6) where
B : C H
a
inverse strongly monotone, : C R { }
is a proper lower semi-continuous and convex function,T
i: C C
is a non-expansive mapping for alli = 1,2,3, , n
,{
n}
,{
n}
,{
n} (0,1)
,) (0,2 }
{
n
, and{ r
n} (0, )
satisfy the following conditions (i)lim
n
n= 0
and
n=1
n=
;(ii)
lim
n
n= 0
;(iii)
0 < liminf
n
n limsup
n
n< 1
; (iv)
n [ e , g ] (0,2 )
,lim
n|
n1
n|= 0
; (v)liminf
nr
n> 0
andlim
n| r
n1 r
n|= 0
.They show that if
:=
ni=1Fix ( T
i) VI ( C , B ) MEP ( , )
is nonempty, then the sequencex
n converges strongly to thez = P
( I A rf )( z )
which is the unique solution of the variational inequality. 0,
, )
(
rf A z x z y
In this paper, motivated and inspired by Takahashi and Toyoda
[3]
, Tada and Takahashi[4]
, and Jitpeera et. al [5], we introduce a new iterative scheme for finding a common element of the set of fixed point of left amenable semigroup, the zero point set of the operator which is the sum of inverse strongly monotone operators and maximal monotone operators, and the set of solutions for systems of equilibrium problems in Hilbert spaces by using a hybrid steepest descent methods. Then strong convergence of the scheme to a common element of the three sets is proved.Our results improve and generalize some well-known results in the literature.
2. PRELIMINARIES
Lemma 2.1 ([6]) Let
S
be a semigroup andC
be a nonempty closed convex subset of a reflexive Banach spaceE
. Let = { T
tx : t S }
be a non-expansive semigroup onH
such that{ T
tx : t S }
is bounded for someC
x
, letX
be a subspace ofB (S )
such that1 X
and the mappingt T
tx , y
is an element ofX
for eachx C
andy
E
, and
is a mean onX
. If we writeT
x
instead of Ttxd (t )
, then the followings
hold .
(i)
T
is non-expansive mapping fromC
intoC
; (ii)T
x = x
for eachx Fix ( )
;(iii)
T
x c o { T
tx : t S }
for eachx C
.Lemma 2.2 ([7]) Let
H
be a real Hilbert spaces, there hold the following identities:(i)for each
x H
andx
C
,x
= P
Cx x x
, y x
0
for ally C
;(ii)
P
C: H C
is non-expansive,that is, P
Cx P
Cy x y
for allx , y H
;(iii)
P
C is firmly non-expansive,that is, P
Cx P
Cy
2 P
Cx P
Cy , x y
for allx , y H
;(iv)
tx (1 t ) y
2= t x
2 (1 t ) y
2 t (1 t ) x y
2, t [0,1]
, for allx , y H
; (v) x y
2 x
2 2 y , x y
.Lemma 2.3 ([8]) Let
{ x
n}
and{ y
n}
be bounded sequences in a Banach spaceE
and let{
n}
be a sequence in238
[0,1]
with0 < liminf
n
n limsup
n
n< 1
. Supposex
n1=
nx
n (1
n) y
n for all integersn 0
and
0.
) (
limsup
1
1
n n n nn
x x y y
Then,
lim
n y
n x
n = 0
.Lemma 2.4 ([9]) Let
{ a
n}
be a sequence of nonnegative real numbers such that0.
, )
1
(1
b a b c n
a
n n n n nwhere
{ b
n}
and{ c
n}
are sequences of real numbers satisfying the following conditions(i)
(0,1), = ,
0
=
n n
n
b
b
(ii)either
limsup
nc
n 0
or
n=0| b
nc
n|<
. Then,lim
na
n= 0
.Lemma 2.5 ([10]) Assume
A
be a strongly positive linear bounded operator onH
with coefficient > 0
and<
10 A
. Then I A 1
.Throughout this article, we assume that a bi-function
F : C C R
satisfies the following conditions:(A1)
F ( x , x ) = 0
for allx C
;(A2)
F
is monotone, i.e.,F ( x , y ) F ( y , x ) 0
for allx , y C
;(A3)for each
x , y , z C
,lim
t0F ( tz (1 t ) x , y ) F ( x , y )
; (A4)for eachx C
,y F ( x , y )
is convex and lower semi-continuous.Lemma 2.6 ([11]) Let
C
be a nonempty closed convex subset of a real Hilbert space spaceH
and letF
be a bi-function fromC C
toR
satisfyingA (1) A (4)
. Letr > 0
andx H
. Then, there existsz C
such that, 0,
1 , ) ,
( y z z x y C
y r z
F
Lemma 2.7 ([12]) Let
C
be a nonempty closed convex subset of a real Hilbert space spaceH
,A : C H
bea mapping and
M : H 2
H be a maximal monotone mapping. Then0.
>
(0), ) (
= )) (
( J I rA A M
1r
F
r
Lemma 2.8 ([13]) Let
H
be a real Hilbert space space. Forq
which solves the variational inequality0 ,
)
(
rf F q x q
nk
,f H, p F (T )
, the following statement is true :
,
= ) (
0 ,
)
( rf F q x q P I F rf q q
nk
where
:= F ( S ) ( A M )
1(0) ( B W )
1(0)
Mj=1SEP ( F
j)
. LetI
C be the indicator function ofC
,i.e.,
. ,
,
= 0, )
( x C
C x x
I
CLemma 2.9 ([14]) Let
C
be a nonempty closed convex subset of a real Hilbert space spaceH
. LetP
C be the metric projection fromH
ontoC
, I
C be the sub-differential ofI
C, whereI
C is as defined in (2.1) and)
1(
=
C r
I r I
J
. Then. , ,
=
= J x y P x x H y C
y
r
C
3. MAIN RESULTS
Theorem 3.1 Let
C
be a nonempty closed convex subset of a real Hilbert space spaceH
,A : C H
be an239
inverse strongly monotone,B : C H
be an
inverse strongly monotone,f
be a contraction ofC
into itself with coefficient (0,1)
,F
is a strongly positive linear bounded operator onH
with coefficient > 0
and
<
<
0
. Let{ F
k, k = 1,2, , M }
be a finite family of bi-functions fromC C
toR
which satisfy(4) (1) A
A
,S
a semigroup and = { T
tx : t S }
be a non-expansive semigroup fromC
intoC
such that
( )
= )
(
t SFix T
tFix
. LetX
be a left invariant subspace ofB (S )
such that1 X
, and the function
t T
tx , y
is an element ofX
for eachx C
andy H
,{
n}
a left regular sequence of means onX
such that
lim
n
n1
n = 0
.M : H 2
H andN : H 2
H be maximal monotone operators such thatC M
D ( )
andD ( W ) C
. Assume that
:= F ( S ) ( A M )
1(0) ( B W )
1(0)
Mj=1SEP ( F
j)
. If the sequences{ x
n}
,{ y
n}
and{ u
n}
are generated iteratively by
x
1 C
and
n n n n n n n n
n
n n n n
s n n
n n n
n n
n F r n F r n FM
n rM n
y T F I
x y
rf x
Bu s u J Au
u J y
x T T T
u
) )
((1 )
(
=
) (
) (1 ) (
=
=
1
1 1, 2 2,
,
where
J = ( I
nM )
1n
,
J = ( I s
nW )
1sn ,
{
n}
is a sequence in(0,2 )
,{ s
n}
is a sequence in(0,2 )
,M k n
r
k,}
=1{
are a real sequence in(0, )
and{
n}
,{
n}
,{
n}
are three sequences in(0,1)
. Assume that the following restrictions are satisfied(C1)
0 < liminf
n
n limsup
n
n< 1
; (C2)lim
n
n= 0
and
n=1
n=
;(C3)
0 < a
n b < 2
,0 < c s
n d < 2
and0 < k
n e < 1
; (C4)lim
n|
n1
n|= lim
n| s
n1 s
n|= lim
n|
n1
n|= 0
;(C5)
liminf
nr
k,n> 0
andlim
n| r
k,n1 r
k,n|= 0
for eachk {1,2,3, , M }
, wherea
,b
,c
,d
,k
,e
are real numbers. Then the sequence{ x
n}
,{ y
n}
and{ u
n}
converges strongly to a pointq
, which is the unique solution of the variational inequality. 0,
, )
(
F rf q p q p
(3.2)Equivalently, we have
q = P
( I F rf )( q )
. Proof We shall divide the proof into several steps.By taking
= (
n n n)
n
J
nu Au
v
,= (
n n n)
sn
n
J u s Bu
w
and fork {1,2,3, , M }
and for alln N
, we shall equivalently write scheme(3.1)
as follows:
n n n n n n n n
n
n n n
n n
n M n n
y T F I
x y
rf x
w v
y
x u
) )
((1 )
(
=
) (1
=
=
1
T
Step 1. We show that the mapping
P
( I F rf )
has a unique fixed point.Since
f
be a contraction ofC
into itself with coefficient (0,1)
. Then, we have240
. , , ))
( (1
=
) (1
) ( ) (
) )(
( ( ) ))(
( ( ) ))(
( ( ) ))(
( (
C y x y x r
y x y
x r
y x F I y f x f r
y F I rf x F I rf y
F I rf P x F I rf P
Since
0 < 1 ( r ) < 1
, it follows thatP
( I F rf )
is a contraction ofC
into itself. Therefore, by the Banach Contraction Mapping Principle, has a unique fixed point, sayq C
that is,).
)(
(
= P I F rf q
q
Step 2. We show that
( I
nA )
,( I s
nB )
is non-expansive.For all
x , y C
, we have.
) (2
, 2
= )
( ) (
2
2 2
2 2
2 2
y x
Ay Ax y
x
Ay Ax Ay
Ax y x y
x y
A I x A I
n n
n n
n n
which implies that
( I
nA )
is a non-expansive, so is( I s
nB )
. Step 3. We show that the sequence{ x
n}
,{ y
n}
and{ u
n}
are bounded.Let 1
1, 2 2, 1 1,
=
, Fr n F r n Fk
n rk Fk
n rk k
n
T T
T T
T
fork {1,2,3, , M }
andT
n0= I
. Since for each} , {1,2,3, M k
, Fkn rk
T
, is non-expansive, x
, we note thatu
n= T
nMx
n. It follows that.3.3
=
u
n x
T
nMx
n T
nMx
x
n x
.3.4 )
( ) (
=
v x
J u Au J x
nAx
u
n x
x
n x
n n n n n
n
. )
( ) (
=
w x
J u s Bu J x
s
nBx
u
n x
x
n x
sn n n n n
s
n (3.5)
.
=
) (1
) (1
x x
x x x
x
x w x
v x
y
n
n n n
n
n n n
n n
(3.6)
which yields that
) }.
( ) , (
{
) ) (
(
) ) ) (
) (
(1
=
1 )
(
1 )
(
1 )
( )
( ) (
) (1 )
(
) )(
) ((1 ) (
) )
( (
1
=
r Fx x
x rf x max
Fx x
r rf x r
x r
x x x
x Fx
x rf x
x r
x x x
x Fx
x rf x
u r
x y x
x Fx
x rf x
f y f r
x y T F I
x x Fx
y rf
x y T F I
x x Fx
y rf x
x
n
n n
n
n n n n
n n
n n
n n
n n
n n
n n
n n n n
n n
n n
n n n
n n
n n
n
n n n
n n
n n
n n
By induction,