Open Access

An algorithm for finding common solutions of various problems in nonlinear operator theory

  • Eric U Ofoedu1,
  • Jonathan N Odumegwu1,
  • Habtu Zegeye2 and
  • Naseer Shahzad3Email author
Fixed Point Theory and Applications20142014:9

https://doi.org/10.1186/1687-1812-2014-9

Received: 27 August 2013

Accepted: 21 November 2013

Published: 9 January 2014

Abstract

In this paper, it is our aim to prove strong convergence of a new iterative algorithm to a common element of the set of solutions of a finite family of classical equilibrium problems; a common set of zeros of a finite family of inverse strongly monotone operators; the set of common fixed points of a finite family of quasi-nonexpansive mappings; and the set of common fixed points of a finite family of continuous pseudocontractive mappings in Hilbert spaces on assumption that the intersection of the aforementioned sets is not empty. Moreover, the common element is shown to be the metric projection of the initial guess on the intersection of these sets.

MSC:47H06, 47H09, 47J05, 47J25.

Keywords

classical equilibrium problemgeneralized mixed equilibrium problem η-inverse strongly monotone mappingmaximal monotone operatornonexpansive mappingsreal Hilbert spacepseudocontractive mappingsvariational inequality problem

1 Introduction

Let H be a real Hilbert space. A mapping T with domain D ( T ) and range R ( T ) in H is called an L-Lipschitzian mapping (or simply a Lipschitz mapping) if and only if there exists L 0 such that for all x , y D ( T ) ,
T x T y L x y .

If L [ 0 , 1 ) , then T is called strict contraction or simply a contraction; and if L = 1 , then T is called nonexpansive. A point x D ( T ) is called a fixed point of an operator T if and only if T x = x . The set of fixed points of an operator T is denoted by Fix ( T ) , that is, Fix ( T ) : = { x D ( T ) : T x = x } .

A mapping T with domain D ( T ) and range R ( T ) in H is called a quasi-nonexpansive mapping if and only if Fix ( T ) and for any x D ( T ) , for any u Fix ( T ) ,
T x u x u .

Every nonexpansive mapping with a nonempty fixed point set is quasi-nonexpansive. The following examples show that the converse is not true.

Example 1.1 (see [1])

Let E = [ π , π ] be a subspace of the set of real numbers , endowed with the usual topology. Define T : E E by T x = x cos x for all x E . Clearly, F ( T ) = { 0 } . Observe that
| T x 0 | = | x | × | cos x | | x | = | x 0 | .
Thus, T is quasi-nonexpansive. The mapping T is, however, not a nonexpansive mapping since for x = π 2 and y = π ,
| T x T y | = | π 2 cos ( π 2 ) π cos π | = π .
But
| x y | = | π 2 π | = π 2 .

Example 1.2 (see [1, 2])

Let E = R be endowed with usual topology. Define T : R R by
T x = { x 2 cos ( 1 x ) , x 0 , 0 , x = 0 .
(1.1)
It is easy to see that F ( T ) = { 0 } since for x 0 , T x = x implies that x 2 cos 1 x = x . Thus, for any x 0 , cos 1 x = 2 , which is not possible. So, F ( T ) = { 0 } . Next, observe that for any x R ,
| T x 0 | = | x 2 | × | cos ( 1 x ) | | x | 2 < | x | = | x 0 | .
So, the mapping T is quasi-nonexpansive. Finally, we show that T is not nonexpansive. To see this, let x = 2 3 π and y = 1 π , then
| T x T y | = | 1 3 π cos ( 3 π 2 ) 1 2 π cos π | = 1 2 π .
But,
| x y | = | 2 3 π 1 π | = 1 3 π .
So,
| T x T y | = 1 2 π > 1 3 π = | x y | .

The concept of quasi-nonexpansive mappings was essentially introduced by Diaz and Metcalf [3]. Although Examples 1.1 and 1.2 guarantee the existence of a quasi-nonexpansive mapping which is not nonexpansive, we must note that a linear quasi-nonexpansive mapping defined on a subspace of a given vector space is nonexpansive on that subspace.

Another important generalization of the class of nonexpansive mappings is the class of pseudocontractive mappings. These mappings are intimately connected with the important class of nonlinear accretive operators. This connection will be made precise in what follows.

A mapping T with domain D ( T ) and range R ( T ) in H is called pseudocontractive if and only if for all x , y D ( T ) , the following inequality holds:
x y ( 1 + r ) ( x y ) r ( T x T y )
(1.2)
for all r > 0 . As a consequence of a result of Kato [4], the pseudocontractive mappings can also be defined in terms of the normalized duality mappings as follows: the mapping T is called pseudocontractive if and only if for all x , y D ( T ) , we have that
T x T y , x y x y 2 .
(1.3)

It now follows trivially from (1.3) that every nonexpansive mapping is pseudocontractive. We note immediately that the class of pseudocontractive mappings is larger than that of nonexpansive mappings. For examples of pseudocontractive mappings which are not nonexpansive, the reader may see [5].

To see the connection between the pseudocontractive mappings and the monotone mappings, we introduce the following definition: a mapping A with domain D ( A ) and range R ( A ) in E is called monotone if and only if for all x , y D ( A ) , the following inequality is satisfied:
A x A y , x y 0 .
(1.4)
The operator A is called η-inverse strongly monotone if and only if there exists η ( 0 , 1 ) such that for all x , y D ( A ) , we have that
A x A y , x y η A x A y 2 .
(1.5)

It is easy to see from inequalities (1.3) and (1.4) that an operator A is monotone if and only if the mapping T : = ( I A ) is pseudocontractive. Consequently, the fixed point theory for pseudocontractive mappings is intimately connected with the zero of monotone mappings. For the importance of monotone mappings and their connections with evolution equations, the reader may consult any of the references [5, 6].

Due to the above connection, fixed point theory of pseudocontractive mappings became a flourishing area of intensive research for several authors.

Let C be a closed convex nonempty subset of a real Hilbert space H with inner product , and norm . Let f : C × C R be a bifunction. The classical equilibrium problem (EP) for a bifunction f is to find u C such that
f ( u , y ) 0 , y C .
(1.6)
The set of solutions for EP (1.6) is denoted by
EP ( f ) = { u C : f ( u , y ) 0 , y C } .

The classical equilibrium problem (EP) includes as special cases the monotone inclusion problems, saddle point problems, variational inequality problems, minimization problems, optimization problems, vector equilibrium problems, Nash equilibria in noncooperative games. Furthermore, there are several other problems, for example, the complementarity problems and fixed point problems, which can also be written in the form of the classical equilibrium problem. In other words, the classical equilibrium problem is a unifying model for several problems arising from engineering, physics, statistics, computer science, optimization theory, operations research, economics and countless other fields. For the past 20 years or so, many existence results have been published for various equilibrium problems (see, e.g., [710]). Approximation methods for such problems thus become a necessity.

Iterative approximation of fixed points and zeros of nonlinear mappings has been studied extensively by many authors to solve nonlinear mapping equations as well as variational inequality problems and their generalizations (see, e.g., [1119]). Most published results on nonexpansive mappings (for example) focus on the iterative approximation of their fixed points or approximation of common fixed points of a given family of this class of mappings.

Some attempts to modify the Mann iteration method so that strong convergence is guaranteed have recently been made (we should recall that Mann iteration method only guarantees weak convergence (see, for example, Bauschke et al. [20])). Nakajo and Takahashi [16] formulated the following modification of the Mann iteration method for a nonexpansive mapping T defined on a nonempty bounded closed and convex subset C of a Hilbert space H:
{ x 0 C , y n = α n x n + ( 1 α n ) T x n , C n = { v C : y n v 2 x n v 2 } , Q n = { v C : x n v , x 0 x n 0 } , x n + 1 = P C n Q n ( x 0 ) , n N ,
(1.7)

where P C denotes the metric projection from H onto a closed convex subset C of H. They proved that if the sequence { α n } n 0 is bounded away from 1, then { x n } n 0 defined by (1.7) converges strongly to P F ( T ) ( x 0 ) .

Formulations similar to (1.7) for different classes of nonlinear problems had been presented by Kim and Xu [21], Nilsrakoo and Saejung [22], Ofoedu et al. [23], Yang and Su [24], Zegeye and Shahzad [2527].

In this paper, motivated by the results of the authors mentioned above, it is our aim to prove strong convergence of a new iterative algorithm to a common element of the set of solutions of a finite family of classical equilibrium problems; a common set of zeros of a finite family of inverse strongly monotone mappings; a set of common fixed points of a finite family of quasi-nonexpansive mappings; and a set of common fixed points of a finite family of continuous pseudocontractive mappings in Hilbert spaces on assumption that the intersection of the aforementioned sets is not empty. Moreover, the common element is shown to be the metric projection of the initial guess on the intersection of these sets. Our theorems complement the results of the authors mentioned above and those of several other authors.

2 Preliminary

In what follows, we shall make use of the following lemmas.

Lemma 2.1 (see, e.g., Kopecka and Reich [28])

Let C be a nonempty closed and convex subset of a real Hilbert space. Let x H and P C : H C be the metric projection of H onto C, then for any y C ,
y P C x 2 + P C x x 2 x y 2 .

Lemma 2.2 Let C be a closed convex nonempty subset of a real Hilbert space H; and let P C : H C be the metric projection of H onto C. Let x H , then x 0 = P C x if and only if z x 0 , x x 0 0 for all z C .

Lemma 2.3 Let H be a real Hilbert space, then for any x , y H , α [ 0 , 1 ] ,
α x + ( 1 α ) y 2 = α x 2 + ( 1 α ) y 2 α ( 1 α ) x y 2 .

Lemma 2.4 (see Zegeye [29])

Let C be a nonempty closed convex subset of a real Hilbert space H. Let T : C H be a continuous pseudocontractive mapping, then for all r > 0 and x H , there exists z C such that
y z , T z 1 r y z , ( 1 + r ) z x 0 , y C .

Lemma 2.5 (see Zegeye [29])

Let C be a nonempty closed convex subset of a real Hilbert space H. Let T : C C be a continuous pseudocontractive mapping, then for all r > 0 and x H , define a mapping F r : H C by
F r x = { z C : y z , T z 1 r y z , ( 1 + r ) z x 0 , y C } ,
then the following hold:
  1. (1)

    F r is single-valued;

     
  2. (2)
    F r is firmly nonexpansive type mapping, i.e., for all x , y H ,
    F r x F r y 2 F r x F r y , x y ;
     
  3. (3)

    Fix ( F r ) is closed and convex; and Fix ( F r ) = Fix ( T ) for all r > 0 .

     

In the sequel, we shall require that the bifunction f : C × C R satisfies the following conditions:

(A1) f ( x , x ) = 0 , x C ;

(A2) f is monotone in the sense that f ( x , y ) + f ( y , x ) 0 for all x , y C ;

(A3) lim sup t 0 + f ( t z + ( 1 t ) x , y ) f ( x , y ) for all x , y , z C ;

(A4) the function y f ( x , y ) is convex and lower semicontinuous for all x C .

Lemma 2.6 (see, e.g., [7, 30])

Let C be a closed convex nonempty subset of a real Hilbert space H. Let f : C × C R be a bifunction satisfying conditions (A1)-(A4), then for all r > 0 and x H , there exists u C such that
f ( u , y ) + 1 r y u , u x 0 , y C .
(2.1)
Moreover, if for all x H we define a mapping G r : H 2 C by
G r ( x ) = { u C : f ( u , y ) + 1 r y u , u x 0 , y C } ,
(2.2)
then the following hold:
  1. (1)

    G r is single-valued for all r > 0 ;

     
  2. (2)
    G r is firmly nonexpansive, that is, for all x , z H ,
    G r x G r z 2 G r x G r z , x z ;
     
  3. (3)

    Fix ( G r ) = EP ( f ) for all r > 0 ;

     
  4. (4)

    EP ( f ) is closed and convex.

     

Lemma 2.7 (see Ofoedu [31])

Let C be a nonempty closed convex subset of a real Hilbert space H. Let T : C C be a continuous pseudocontractive mapping. For r > 0 , let F r : H C be the mapping in Lemma  2.5, then for any x H and for any p , q > 0 ,
F p x F q x | p q | p ( F p x + x ) .

Lemma 2.8 (Compare with Lemma 13 of Ofoedu [31])

Let C be a closed convex nonempty subset of a real Hilbert space H. Let f : C × C R be a bifunction satisfying conditions (A1)-(A4). Let r > 0 and let G r be the mapping in Lemma  2.6, then for all p , q > 0 and for all x H , we have that
G p x G q x | p q | p ( G p x + x ) .

3 Main results

Let C be a nonempty closed convex subset of a real Hilbert space H. Let T 1 , T 2 , , T m : C C be m continuous pseudocontractive mappings; let S 1 , S 2 , , S l : C C be l continuous quasi-nonexpansive mappings; let A 1 , A 2 , , A d : C H be d γ j -inverse strongly monotone mappings with constants γ j ( 0 , 1 ) , j = 1 , 2 , , d ; let f 1 , f 2 , , f t : C × C R be t bifunctions satisfying conditions (A1)-(A4). For all x E , i = 1 , 2 , , m , let
F i , r x : = { z C : y z , T i z 1 r y z , ( 1 + r ) z x 0 , y C }
and for all x E , h = 1 , 2 , , t , let
G h , r ( x ) = { u C : f h ( u , y ) + 1 r y u , u x 0 , y C } ,
then in what follows we shall study the following iteration process:
{ x 0 C 0 = C chosen arbitrarily , z n = P C ( x n λ n A n + 1 x n ) , y n = α n x n + ( 1 α n ) S n + 1 z n , w n = η i = 1 m β i F i , r n y n + ( 1 η ) h = 1 t ξ h G h , r n y n , C n + 1 = { z C : w n z x n z } , x n + 1 = Π C n + 1 ( x 0 ) , n 0 ,
(3.1)

where A n = A n ( mod d ) , S n = S n ( mod l ) ; { r n } ( 0 , ) such that lim n r n = r 0 > 0 ; { α n } n 1 a sequence in ( 0 , 1 ) such that lim inf n α n ( 1 α n ) > 0 ; { β i } i = 1 m , { ξ h } h = 1 t ( 0 , 1 ) such that i = 1 m β i = 1 = h = 1 t ξ h ; η ( 0 , 1 ) and { λ n } is a sequence in [ a , b ] for some a , b R such that 0 < a < b < 2 γ , γ = min 1 j d { γ j } .

Lemma 3.1 Let C be a nonempty closed convex subset of a real Hilbert space H. Let T 1 , T 2 , , T m : C C be m continuous pseudocontractive mappings; let S 1 , S 2 , , S l : C C be l continuous quasi-nonexpansive mappings; let A 1 , A 2 , , A d : C H be d γ j -inverse strongly monotone mappings with constants γ j ( 0 , 1 ) , j = 1 , 2 , , d ; let f 1 , f 2 , , f t : C × C R be t bifunctions satisfying conditions (A1)-(A4). Let F : = i = 1 m Fix ( T i ) j = 1 d A j 1 ( 0 ) k = 1 l Fix ( S k ) h = 1 t EP ( f h ) . Let { x n } be a sequence defined by (3.1), then the sequence { x n } is well defined for each n 0 .

Proof We first show that C n is closed and convex for each n N { 0 } . From the definitions of C n it is obvious that C n is closed. Moreover, since w n z x n z is equivalent to 2 z , x n w n x n 2 + w n 2 0 , it follows that C n is convex for each n N { 0 } . Next, we prove that F C n for each n N { 0 } . From the assumption, we see that F C 0 = C . Suppose that F C k for some k 1 , then for p F , we obtain that
w k p = η i = 1 m β i F i , r k y k + ( 1 η ) h = 1 m ξ h G h , r k y k p y k p = α k x k + ( 1 α k ) S k + 1 z k p α k x k p + ( 1 α k ) S k + 1 z k p α k x k p + ( 1 α k ) z k p .
(3.2)
Furthermore,
z k p 2 = P C ( x k λ k A k + 1 x k ) p 2 x k λ k A k + 1 x k p 2 = x k p λ k ( A k + 1 x k A k + 1 p ) 2 = x k p 2 2 λ k x k p , A k + 1 x k A k + 1 p + λ k 2 A k + 1 x k A k + 1 p 2 x k p 2 + λ k ( λ k 2 γ ) A k + 1 x k A k + 1 p 2 x k p 2 ( since  λ k < 2 γ ) .
Thus,
z k p x k p .
(3.3)
Using (3.3) in (3.2) gives
w k p x k p .

So, p C k + 1 . This implies, by induction, that F C n so that the sequence generated by (3.1) is well defined for all n 0 . □

Theorem 3.2 Let C be a nonempty closed convex subset of a real Hilbert space H. Let T 1 , T 2 , , T m : C C be m continuous pseudocontractive mappings; let S 1 , S 2 , , S l : C C be l continuous quasi-nonexpansive mappings; let A 1 , A 2 , , A d : C H be d γ j -inverse strongly monotone mappings with constants γ j ( 0 , 1 ) , j = 1 , 2 , , d ; let f 1 , f 2 , , f t : C × C R be t bifunctions satisfying conditions (A1)-(A4). Let F : = i = 1 m Fix ( T i ) j = 1 d A j 1 ( 0 ) k = 1 l Fix ( S k ) h = 1 t EP ( f h ) . Let { x n } be a sequence defined by (3.1). Then the sequence { x n } n 0 converges strongly to the element of F nearest to x 0 .

Proof From Lemma 3.1, we obtain that F C n , n 0 and x n is well defined for each n 0 . From x n = P C n ( x 0 ) and x n + 1 = P C n + 1 ( x 0 ) C n + 1 C n , we obtain that
x n + 1 x n , x n x 0 0 and x n x 0 x n + 1 x 0 .
Besides, by Lemma 2.1,
x n p 2 = P C n x 0 x 0 x 0 p 2 x 0 x n 2 x 0 p 2 .
Thus, the sequence { x n x 0 } n 0 is a bounded nondecreasing sequence of real numbers. So, lim n x n x 0 exists. Similarly, by Lemma 2.1, we have that for any positive integer μ,
x n + μ x n 2 = x n + μ P C n x 0 2 x n + μ x 0 2 P C n x 0 x 0 2 = x n + μ x 0 2 x n x 0 2 for all  n 0 .
Since lim n x n x 0 exists, we have that lim n x n + μ x n = 0 and hence, { x n } n 1 is a Cauchy sequence in C. Therefore, there exists x C such that lim n x n = x . Since x n + 1 C n + 1 , we have that
w n x n + 1 x n x n + 1 .
Thus,
lim n x n + 1 w n = 0
(3.4)

and hence x n w n x n x n + 1 + x n + 1 w n 0 as n , which implies that w n x as n .

Next, we observe that for p F and using Lemma 2.3,
y n p 2 = α n x n + ( 1 α n ) S n + 1 z n p 2 = α n ( x n p ) + ( 1 α n ) ( S n + 1 z n p ) 2 = α n x n p + ( 1 α n ) S n + 1 z n p 2 α n ( 1 α n ) x n S n + 1 z n 2
(3.5)
α n x n p 2 + ( 1 α n ) z n p 2 α n ( 1 α n ) x n S n + 1 z n 2 .
(3.6)
But
z n p 2 x n p 2 + λ n ( λ n 2 γ ) A n + 1 x n A n + 1 p 2 = x n p 2 + λ n ( λ n 2 γ ) A n + 1 x n 2 .
(3.7)
So, using (3.7) in (3.6), we obtain that
y n p 2 α n x n p 2 + ( 1 α n ) [ x n p 2 + λ n ( λ n 2 γ ) A n + 1 x n 2 ] α n ( 1 α n ) x n S n + 1 z n 2 = x n p 2 + ( 1 α n ) λ n ( λ n 2 γ ) A n + 1 x n 2 α n ( 1 α n ) x n S n + 1 z n 2 .
(3.8)
Moreover, we obtain that
w n p 2 = η i = 1 m β i F i , r n y n + ( 1 η ) h = 1 m ξ h G h , r n y n p 2 y n p 2 .
(3.9)
Using (3.8) in (3.9) we get that
w n p 2 x n p 2 + ( 1 α n ) λ n ( λ n 2 γ ) A n + 1 x n 2 α n ( 1 α n ) x n S n + 1 z n 2 .
(3.10)
Now, using the fact that λ n < 2 γ , inequality (3.10) gives (for some constant M 0 > 0 ) that
α n ( 1 α n ) x n S n + 1 z n x n p 2 w n p 2 M 0 x n w n .
(3.11)
Hence, we obtain from inequality (3.11) that
x n S n + 1 z n 0 as  n .
(3.12)
Moreover, from (3.10) we obtain that
( 1 α n ) λ n ( 2 γ λ n ) A n + 1 x n 2 x n p 2 w n p 2 M 0 x n w n ,
which yields that
lim n A n + 1 x n = 0 .
(3.13)
Now,
x n z n = x n P C ( x n λ n A n + 1 x n ) = P C x n P C ( x n λ n A n + 1 x n ) x n x n + λ n A n + 1 x n = λ n A n + 1 x n b A n + 1 x n .
(3.14)
It follows from (3.13) and (3.14) that
lim n x n z n = 0 ;
(3.15)

and hence z n x as n .

We now show that x k = 1 l Fix ( S k ) . Observe that from (3.12) and (3.15) we obtain that
S n + 1 z n z n S n + 1 z n x n + z n x n 0 as  n ,
(3.16)
so that
lim n S n + 1 z n = x .
(3.17)
Let { n σ } σ 1 N be such that S n σ + 1 = S 1 for all σ N , then since z n σ x as σ , we obtain from (3.17), using the continuity of S 1 , that
x = lim σ S n σ + 1 z n σ = lim σ S 1 z n σ = S 1 x .
Similarly, if { n j } j 1 N is such that S n j + 1 = S 2 for all j N , then we have again that
x = lim j S n j + 1 z n j = lim j S 2 z n j = S 2 x .

Continuing, we obtain that S k x = x , k = 3 , , l . Hence, x k = 1 l F ( S k ) .

Next, we show that x j = 1 d A j 1 ( 0 ) . Since A j is γ-inverse strongly monotone for j = 1 , 2 , , d , we have that A j is 1 γ -Lipschitz continuous. Thus,
A n + 1 x n A n + 1 x 1 γ x n x 0 as  n .
(3.18)
Hence, from (3.18) and (3.13), we obtain that
A n + 1 x A n + 1 x n A n + 1 x + A n + 1 x n 0 as  n .
As a result, we get that
lim n A n + 1 x = 0 .
Let { n s } s 1 N be such that A n s + 1 = A 1 for all s N . Then
A 1 x = lim s A n s + 1 x = 0 .

Similarly, we have that A j x = 0 for j = 2 , , d . Thus, x j = 1 d A i 1 ( 0 ) .

Furthermore, we show that x i = 1 m Fix ( T i ) = i = 1 m Fix ( F i , r ) , r > 0 . Using the fact that x n x , z n x as n , we obtain that
F 1 , r n y n x y n x α n x n x + ( 1 α n ) z n x x n x + z n x 0 as  n .
(3.19)
Thus, we obtain from (3.19) that
lim n F 1 , r n y n = x = lim n y n .
This implies that lim n F 1 , r n y n y n = 0 . But by Lemma 2.7,
F 1 , r n y n F 1 , r 0 y n | r n r 0 | r n ( F 1 , r n y n + y n ) 0 as  n .
Thus,
lim n F 1 , r 0 y n = lim n F 1 , r n y n = x .
So, the continuity of F 1 , r 0 and the fact that y n x as n give
x = lim n F 1 , r 0 y n = F 1 , r 0 x .
A similar argument gives
x = lim n F i , r 0 y n = F i , r 0 x , i = 2 , 3 , , m .
Hence,
x i = 1 m Fix ( F i , r 0 ) = i = 1 m Fix ( T i ) .
Moreover, we show that x h = 1 t EP ( f h ) = h = 1 t Fix ( G h , r 0 ) . Observe that
G 1 , r n y n x y n x α n x n x + ( 1 α n ) z n x x n x + z n x 0 as  n .
(3.20)
Thus, we obtain from (3.20) that
lim n G 1 , r n y n = x = lim n y n .
This implies that lim n G 1 , r n y n y n = 0 . But by Lemma 2.8,
G 1 , r n y n G 1 , r 0 y n | r n r 0 | r n ( G 1 , r n y n + y n ) 0 as  n .
Thus,
lim n G 1 , r 0 y n = lim n G 1 , r n y n = x .
So, the continuity of G 1 , r 0 and the fact that y n x as n give
x = lim n G 1 , r 0 y n = G 1 , r 0 x .
A similar argument gives
x = lim n G h , r 0 y n = G i , r 0 x , h = 2 , 3 , , t .
Hence,
x h = 1 t Fix ( G h , r 0 ) = h = 1 t EP ( f h ) .
Finally, we prove that x = P F ( x 0 ) . From x n = P C n ( x 0 ) n 0 , we obtain that
x 0 x n , x n z 0 , z C n .
Since F C n , we also have that
x 0 x n , x n p 0 , p F .
(3.21)
So,
0 x 0 x n , x n p = x 0 x + x x n , x n x + x p = x 0 x , x n x + x 0 x , x p + x x n , x n x + x x n , x p x 0 x y , x p + x 0 x x n x + x n x x p x n x 2 .
(3.22)
Inequality (3.22) implies that
0 x 0 x , x p + ( x 0 x + x p ) x n x .
(3.23)
By taking limit as n in (3.23), we obtain that
x 0 x , x p 0 , p F .

Now, by Lemma 2.2 we have that x = P F ( x 0 ) . This completes the proof. □

Remark 3.3 We note that x = P F ( x 0 ) makes sense since it could be easily shown that F is closed and convex. In fact, it is enough to show that the set of zeros of γ-inverse monotone mappings and a fixed point set of continuous quasi-nonexpansive mappings are convex sets. Closure of the two sets simply follows from the continuity of the mappings involved.

Remark 3.4 Several authors (see, e.g., [8, 31] and references therein) have studied the following problem: Let C be a closed convex nonempty subset of a real Hilbert space H with inner product , and norm . Let f : C × C R be a bifunction and Φ : C R { + } be a proper extended real-valued function, where denotes the set of real numbers. Let Θ : C H be a nonlinear monotone mapping. The generalized mixed equilibrium problem (abbreviated GMEP) for f, Φ and Θ is to find u C such that
f ( u , y ) + Φ ( y ) Φ ( u ) + Θ u , y u 0 , y C .
(3.24)
The set of solutions for GMEP (3.24) is denoted by
GMEP ( f , Φ , Θ ) = { u C : f ( u , y ) + Φ ( y ) Φ ( u ) + Θ u , y u 0 , y C } .
These authors always claim that if Φ 0 Θ in (3.24), then (3.24) reduces to the classical equilibrium problem (abbreviated EP), that is, the problem of finding u C such that
f ( u , y ) 0 , y C
(3.25)
and GMEP ( f , 0 , 0 ) is denoted by EP ( f ) , where
EP ( f ) = { u C : f ( u , y ) 0 , y C } .
If f 0 Φ in (3.24), then GMEP (1.6) reduces to the classical variational inequality problem and GMEP ( 0 , 0 , Θ ) is denoted by VI ( Θ , C ) , where
VI ( Θ , C ) = { u C : Θ u , y u 0 , y C } .
If f 0 Θ , then GMEP (3.24) reduces to the following minimization problem:
find  u C  such that  Φ ( y ) Φ ( u ) , y C ;
and GMEP ( 0 , Φ , 0 ) is denoted by Argmin ( Φ ) , where
Argmin ( Φ ) = { u C : Φ ( u ) Φ ( y ) , y C } .
If Θ 0 , then (3.24) becomes the mixed equilibrium problem (abbreviated MEP) and GMEP ( f , Φ , 0 ) is denoted by MEP ( f , Φ ) , where
MEP ( f , Φ ) = { u C : f ( u , y ) + Φ ( y ) Φ ( u ) 0 , y C } .
If Φ 0 , then (1.6) reduces to the generalized equilibrium problem (abbreviated GEP) and GMEP ( f , 0 , Θ ) is denoted by GEP ( f , Θ ) , where
GEP ( f , Θ ) = { u C : f ( u , y ) + Θ u , y u 0 , y C } .
If f 0 , then GMEP (3.24) reduces to the generalized variational inequality problem (abbreviated GVIP) and GMEP ( 0 , Φ , Θ ) is denoted by GVIP ( Φ , Θ , C ) , where
GVIP ( Φ , Θ , C ) = { u K : Φ ( y ) Φ ( u ) + Θ u , y u 0 , y C } .
It is worthy to note that if we define Γ : C × C R by
Γ ( x , y ) = f ( x , y ) + Φ ( y ) Φ ( x ) + Θ x , y x ,

then it could be easily checked that Γ is a bifunction and satisfies properties (A1)-(A4). Thus, the so-called generalized mixed equilibrium problem reduces to the classical equilibrium problem for the bifunction Γ. Thus, consideration of the so-called generalized mixed equilibrium problem in place of the classical equilibrium problem studied in this paper leads to no further generalization.

4 Application (convex differentiable optimization)

In Section 1, we defined a Lipschitz continuous mapping and an inverse strongly monotone mapping. Inverse strongly monotone mappings arise in various areas of optimization and nonlinear analysis (see, for example, [3238]). It follows from the Cauchy-Schwarz inequality that if a mapping A : D ( A ) H R ( A ) H is 1 L -inverse strongly monotone, then A is L-Lipschitz continuous. The converse of this statement, however, fails to be true. To see this, take for instance A = I , where I is the identity mapping on H, then A is L-Lipschitz continuous (with L = 1 ) but not 1 L -inverse strongly monotone (that is, not firmly nonexpansive in this case).

Baillon and Haddad [39] showed in 1977 that if D ( A ) = H and A is the gradient of a convex functional on H, then A is 1 L -inverse strongly monotone if and only if A is L-Lipschitz continuous. This remarkable result, which has important applications in optimization theory (see, for example, [4042]), has become known as the Baillon-Haddad theorem. In fact, we have the following theorem.

Theorem 4.1 (Baillon-Haddad) (see Corollary 10 of [39])

Let ϕ : H R be a convex Fréchet-differentiable functional on H such that ϕ is L-Lipschitz continuous for some L ( 0 , + ) , then ϕ is a 1 L -inverse strongly monotone mapping (where ϕ denotes the gradient of the functional ϕ).

Now, let us turn to the problem of minimizing a continuously Fréchet-differentiable convex functional with minimum norm in Hilbert spaces.

Let K be a closed convex subset of a real Hilbert space H, consider the minimization problem given by
min x K ϕ ( x ) ,
(4.1)
where ϕ is a Fréchet-differentiable convex functional. Let Ω K , the solution set of (4.1), be nonempty. It is known that a point z Ω if and only if the following optimality condition holds:
z K , ϕ ( z ) , x z 0 , x K .
(4.2)

It is easy to see that if K = H , then optimality condition (4.2) is equivalent to z Ω if and only if z ( ϕ ) 1 ( 0 ) .

Thus, we obtain the following as a corollary of Theorem 3.2.

Theorem 4.2 Let C be a nonempty closed convex subset of a real Hilbert space H. Let T 1 , T 2 , , T m : C C be m continuous pseudocontractive mappings; let S 1 , S 2 , , S l : C C be l continuous quasi-nonexpansive mappings; let ϕ 1 , ϕ 2 , , ϕ d : H H be d convex and Fréchet-differentiable functionals on H such that ( ϕ ) j is L j -Lipschitz continuous for some L j ( 0 , + ) , j = 1 , 2 , , d ; let f 1 , f 2 , , f t : C × C R be t bifunctions satisfying conditions (A1)-(A4). Let F : = i = 1 m Fix ( T i ) j = 1 d ( ϕ j ) 1 ( 0 ) k = 1 l Fix ( S k ) h = 1 t EP ( f h ) . Let { x n } n 0 be a sequence defined by
{ x 0 C 0 = C chosen arbitrarily , z n = P C ( x n λ n ( ϕ ) n + 1 x n ) , y n = α n x n + ( 1 α n ) S n + 1 z n , w n = η i = 1 m β i F i , r n y n + ( 1 η ) h = 1 t ξ h G h , r n y n , C n + 1 = { z C n : w n z x n z } , x n + 1 = Π C n + 1 ( x 0 ) , n 0 ,

where ( ϕ ) n = ( ϕ ) n ( mod d ) , S n = S n ( mod l ) ; { r n } ( 0 , ) such that lim n r n = r 0 > 0 ; { α n } n 1 a sequence in ( 0 , 1 ) such that lim inf n α n ( 1 α n ) > 0 ; { β i } i = 1 m , { ξ h } h = 1 t ( 0 , 1 ) such that i = 1 m β i = 1 = h = 1 t ξ h ; η ( 0 , 1 ) and { λ n } is a sequence in [ a , b ] for some a , b R such that 0 < a < b < 2 L , L = max 1 j d { L j } . Then the sequence { x n } n 0 converges strongly to the element of F nearest to x 0 .

Proof Since, by our hypothesis, ( ϕ ) j is L j -Lipschitz continuous for some L j ( 0 , + ) , j = 1 , 2 , , d , we obtain from Theorem 4.1 that ( ϕ ) j is 1 L j -inverse strongly monotone, j = 1 , 2 , , d ; and since L = max 1 j d { L j } , it is then easy to see that ( ϕ ) j is 1 L -inverse strongly monotone, j = 1 , 2 , , d . The rest, therefore, follows as in the proof of Theorem 3.2 with γ = 1 L . This completes the proof. □

Declarations

Acknowledgements

This article was funded by the Deanship of Scientific Research (DSR), King Abdulaziz University, Jeddah. N Shahzad acknowledges with thanks DSR for financial support.

Authors’ Affiliations

(1)
Department of Mathematics, Nnamdi Azikiwe University
(2)
Bahir Dar University
(3)
Department of Mathematics, King Abdulaziz University

References

  1. Kim GE: Weak and strong convergence for quasi-nonexpansive mappings in Banach spaces. Bull. Korean Math. Soc. 2012, 49(4):799–813. 10.4134/BKMS.2012.49.4.799View ArticleMathSciNetGoogle Scholar
  2. Dotson WG: On the Mann iterative process. Trans. Am. Math. Soc. 1970, 149: 63–73.View ArticleMathSciNetGoogle Scholar
  3. Diaz JB, Metcalf FT: On the set of subsequential limit points of successive approximations. Trans. Am. Math. Soc. 1969, 135: 459–485.MathSciNetGoogle Scholar
  4. Kato T: Nonlinear semigroups and evolution equations. J. Math. Soc. Jpn. 1967, 19: 508–520. 10.2969/jmsj/01940508View ArticleGoogle Scholar
  5. Chidume CE Lecture Notes in Mathematics 1965. Geometric Properties of Banach Spaces and Nonlinear Iterations 2009.Google Scholar
  6. Ofoedu EU, Zegeye H: Further investigation on iteration processes for pseudocontractive mappings with application. Nonlinear Anal. TMA 2012, 75: 153–162. 10.1016/j.na.2011.08.015View ArticleMathSciNetGoogle Scholar
  7. Blum E, Oettli W: From optimization and variational inequalities to equilibrium problems. Math. Stud. 1994, 63(1–4):123–145.MathSciNetGoogle Scholar
  8. Katchang P, Jitpeera T, Kumam P: Strong convergence theorems for solving generalized mixed equilibrium problems and general system of variational inequalities by the hybrid method. Nonlinear Anal. Hybrid Syst. 2010. 10.1016/j.nahs.2010.07.001Google Scholar
  9. Zegeye, H, Shahzad, N: Strong convergence theorems for a solution of finite families of equilibrium and variational inequality problems. Optimization, 1–17, iFirst (2012)Google Scholar
  10. Zegeye H, Shahzad N: Convergence theorems for a common point of solutions of equilibrium and fixed point of relatively nonexpansive multi-valued mapping problems. Abstr. Appl. Anal. 2012., 2012: Article ID 859598Google Scholar
  11. Censor Y, Gibali A, Reich S, Sabach S: Common solutions to variational inequalities. Set-Valued Var. Anal. 2012, 20: 229–247. 10.1007/s11228-011-0192-xView ArticleMathSciNetGoogle Scholar
  12. Censor Y, Gibali A, Reich S: A von Neumann alternating method for finding common solutions to variational inequalities. Nonlinear Anal. 2012, 75: 4596–4603. 10.1016/j.na.2012.01.021View ArticleMathSciNetGoogle Scholar
  13. Ishikawa S: Fixed point by a new iteration method. Proc. Am. Math. Soc. 1974, 44: 147–150. 10.1090/S0002-9939-1974-0336469-5View ArticleMathSciNetGoogle Scholar
  14. Mann WR: Mean value methods in iteration. Proc. Am. Math. Soc. 1953, 4: 506–510. 10.1090/S0002-9939-1953-0054846-3View ArticleGoogle Scholar
  15. Martines-Yanes C, Xu HK: Strong convergence of CQ method for fixed point iteration. Nonlinear Anal. 2006, 64: 2400–2411. 10.1016/j.na.2005.08.018View ArticleMathSciNetGoogle Scholar
  16. Nakajo K, Takahashi W: Strong convergence theorems for nonexpansive mappings. J. Math. Anal. Appl. 2003, 279: 372–379. 10.1016/S0022-247X(02)00458-4View ArticleMathSciNetGoogle Scholar
  17. Reich S: Weak convergence theorems for nonexpansive mappings in Banach spaces. J. Math. Anal. Appl. 1979, 67: 274–276. 10.1016/0022-247X(79)90024-6View ArticleMathSciNetGoogle Scholar
  18. Zegeye H, Shahzad N: Strong convergence theorems for a common zero of countably infinite family of α -inverse strongly accretive mappings. Nonlinear Anal. 2009, 71: 531–538. 10.1016/j.na.2008.10.091View ArticleMathSciNetGoogle Scholar
  19. Zegeye H: Strong convergence theorems for maximal monotone mappings in Banach spaces. J. Math. Anal. Appl. 2008, 343: 663–671. 10.1016/j.jmaa.2008.01.076View ArticleMathSciNetGoogle Scholar
  20. Bauschke HH, Matouskova E, Reich S: Projections and proximal point methods: convergence results and counterexamples. Nonlinear Anal. 2004, 56: 715–738. 10.1016/j.na.2003.10.010View ArticleMathSciNetGoogle Scholar
  21. Kim TH, Xu HK: Strong convergence of modified Mann iteration for asymptotically nonexpansive mappings and semigroups. Nonlinear Anal. 2006, 64: 1140–1152. 10.1016/j.na.2005.05.059View ArticleMathSciNetGoogle Scholar
  22. Nilsrakoo W, Saejug S: Weak and strong convergence theorems for countable Lipschitzian mappings and its applications. Nonlinear Anal. TMA 2008, 69(8):2695–2708. 10.1016/j.na.2007.08.044View ArticleGoogle Scholar
  23. Ofoedu EU, Shehu Y, Ezeora JN: Solution by iteration of nonlinear variational inequalities involving finite family of asymptotically nonexpansive mappings and monotone mappings. Panam. Math. J. 2008, 18(4):61–75.MathSciNetGoogle Scholar
  24. Yang L, Su Y: Strong convergence theorems for countable Lipschitzian mappings and its applications in equilibrium and optimization problems. Fixed Point Theory Appl. 2009., 2009: Article ID 462489 10.1155/2009/462489Google Scholar
  25. Zegeye H, Shahzad N: Strong convergence theorems for a finite family of asymptotically nonexpansive mappings and semigroups. Nonlinear Anal. 2007. 10.1016/j.na.2007.11.005Google Scholar
  26. Zegeye H, Shahzad N: A hybrid approximation method for equilibrium, variational inequality and fixed point problems. Nonlinear Anal. Hybrid Syst. 2010, 4: 619–630. 10.1016/j.nahs.2010.03.005View ArticleMathSciNetGoogle Scholar
  27. Zegeye H, Shahzad N: Approximating common solution of variational inequality problems for two monotone mappings in Banach spaces. Optim. Lett. 2011, 5(4):691–704. 10.1007/s11590-010-0235-5View ArticleMathSciNetGoogle Scholar
  28. Kopecka E, Reich S: A note on alternating projections in Hilbert space. J. Fixed Point Theory Appl. 2012, 12: 41–47. 10.1007/s11784-013-0097-4View ArticleMathSciNetGoogle Scholar
  29. Zegeye H: An iterative approximation method for a common fixed point of two pseudo-contractive mappings. ISRN Math. Anal. 2011., 2011: Article ID 621901 10.5402/2011/621901Google Scholar
  30. Kassay G, Reich S, Sabach S: Iterative methods for solving systems of variational inequalities in reflexive Banach spaces. SIAM J. Optim. 2011, 21: 1319–1344. 10.1137/110820002View ArticleMathSciNetGoogle Scholar
  31. Ofoedu EU: A general approximation scheme for solutions of various problems in fixed point theory. Int. J. Anal. 2013., 2013: Article ID 762831 10.1155/2013/762831Google Scholar
  32. Byrne CL: Applied Iterative Methods. AK Peters, Wellesley; 2008.Google Scholar
  33. Combettes PL: Solving monotone inclusions via compositions of nonexpansive averaged operators. Optimization 2004, 53: 475–504. 10.1080/02331930412331327157View ArticleMathSciNetGoogle Scholar
  34. Dunn JC: Convexity, monotonicity and gradient processes in Hilbert space. J. Math. Anal. Appl. 1976, 53: 145–158. 10.1016/0022-247X(76)90152-9View ArticleMathSciNetGoogle Scholar
  35. Lions PL, Mercier B: Splitting algorithms for the sum two nonlinear operators. SIAM J. Numer. Anal. 1979, 16: 964–979. 10.1137/0716071View ArticleMathSciNetGoogle Scholar
  36. Liu F, Nashed MZ: Regularization of nonlinear ill-posed variational inequalities and convergence rates. Set-Valued Anal. 1998, 6: 313–344. 10.1023/A:1008643727926View ArticleMathSciNetGoogle Scholar
  37. Tseng P: Applications of a splitting algorithm to decomposition in convex programming and variational inequalities. SIAM J. Control Optim. 1991, 29: 119–138. 10.1137/0329006View ArticleMathSciNetGoogle Scholar
  38. Zhu DL, Marcotte P: Co-coercivity and its role in the convergence of iterative schemes for solving variational inequalities. SIAM J. Optim. 1996, 6: 714–726. 10.1137/S1052623494250415View ArticleMathSciNetGoogle Scholar
  39. Baillon JB, Haddad G: Quelques propriétés des opérateurs angle-bornés et n -cycliquement monotones. Isr. J. Math. 1977, 26: 137–150. 10.1007/BF03007664View ArticleMathSciNetGoogle Scholar
  40. Combettes PL, Wajs VR: Signal recovery by proximal forward-backward splitting. Multiscale Model. Simul. 2005, 4: 1168–1200. 10.1137/050626090View ArticleMathSciNetGoogle Scholar
  41. Yamada I, Ogura N: Hybrid steepest descent method for variational inequality problem over the fixed point set of certain quasi-nonexpansive mappings. Numer. Funct. Anal. Optim. 2004, 25: 619–655.View ArticleMathSciNetGoogle Scholar
  42. Masad E, Reich S: A note on the multiple-set split convex feasibility problem in Hilbert space. J. Nonlinear Convex Anal. 2007, 8: 367–371.MathSciNetGoogle Scholar

Copyright

© Ofoedu et al.; licensee Springer. 2014

This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.