Fixed Point Theory and Applications
Volume 2010 (2010), Article ID 590278, 33 pages
doi:10.1155/2010/590278
Research Article

Strong and Weak Convergence Theorems for Common Solutions of Generalized Equilibrium Problems and Zeros of Maximal Monotone Operators

1Department of Mathematics, Shanghai Normal University, Shanghai 200234, China
2Scientific Computing Key Laboratory of Shanghai Universities, Shanghai 200234, China
3Department of Mathematics, Aligarh Muslim University, Aligarh 202 002, India
4Department of Finance, National Sun Yat-Sen University, Kaohsiung 80424, Taiwan
5Department of Applied Mathematics, National Sun Yat-Sen University, Kaohsiung 80424, Taiwan

Received 27 October 2009; Accepted 12 January 2010

Academic Editor: Tomonari Suzuki

Copyright © 2010 L.-C. Zeng et al. 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.

Abstract

The purpose of this paper is to introduce and study two modified hybrid proximal-point algorithms for finding a common element of the solution set EP of a generalized equilibrium problem and the set 𝑇 1 𝑇 0 1 0 for two maximal monotone operators 𝑇 and 𝑇 defined on a Banach space 𝑋 . Strong and weak convergence theorems for these two modified hybrid proximal-point algorithms are established.

1. Introduction

Let 𝑋 be a real Banach space with its dual 𝑋 . The mapping 𝐽 𝑋 2 𝑋 defined by

𝑥 𝐽 ( 𝑥 ) = 𝑋 𝑥 , 𝑥 = 𝑥 2 = 𝑥 2 , 𝑥 𝑋 , ( 1 . 1 ) is called the normalized duality mapping. From the Hahn-Banach theorem, it follows that 𝐽 ( 𝑥 ) for each 𝑥 𝑋 .

A Banach space 𝑋 is said to be strictly convex, if 𝑥 + 𝑦 / 2 < 1 for all 𝑥 , 𝑦 𝑈 = { 𝑧 𝑋 𝑧 = 1 } with 𝑥 𝑦 . 𝑋 is said to be uniformly convex if for each 𝜖 ( 0 , 2 ] , there exists 𝛿 > 0 such that 𝑥 + 𝑦 / 2 1 𝛿 for all 𝑥 , 𝑦 𝑈 with 𝑥 𝑦 𝜖 . Recall that each uniformly convex Banach space has the Kadec-Klee property, that is,

𝑥 𝑛 𝑥 𝑥 𝑛 𝑥 𝑥 𝑛 𝑥 . ( 1 . 2 )

It is well known that if 𝑋 is strictly convex, then 𝐽 is single-valued. In the sequel, we shall still denote the single-valued normalized duality mapping by 𝐽 . Let 𝐶 be a nonempty closed convex subset of 𝑋 , 𝑓 𝐶 × 𝐶 a bifunction, and 𝐴 𝐶 𝑋 a nonlinear mapping. Very recently, Zhang [1] considered and studied the generalized equilibrium problem of finding ̂ 𝑥 𝐶 such that

𝑓 ( ̂ 𝑥 , 𝑦 ) + 𝐴 ̂ 𝑥 , 𝑦 ̂ 𝑥 0 , 𝑦 𝐶 . ( 1 . 3 )

The set of solutions of (1.3) is denoted by 𝐸 𝑃 . Problem (1.3) and related problems have been studied and investigated extensively in the literature; See, for example, [212] and references therein. Whenever 𝐴 0 , problem (1.3) reduces to the equilibrium problem of finding ̂ 𝑥 𝐶 such that

𝑓 ( ̂ 𝑥 , 𝑦 ) 0 , 𝑦 𝐶 . ( 1 . 4 ) The set of solutions of (1.4) is denoted by 𝐸 𝑃 ( 𝑓 ) . Whenever 𝑓 0 , problem (1.3) reduces to the variational inequality problem of finding ̂ 𝑥 𝐶 such that

𝐴 ̂ 𝑥 , 𝑦 ̂ 𝑥 0 , 𝑦 𝐶 . ( 1 . 5 ) The set of solutions of (1.5) is denoted by 𝑉 𝐼 ( 𝐶 , 𝐴 ) .

Whenever 𝑋 = 𝐻 a Hilbert space, problem (1.3) was very recently introduced and considered by S. Takahashi and W. Takahashi [13]. Problem (1.3) is very general in the sense that it includes, as spacial cases, optimization problems, variational inequalities, minimax problems, Nash equilibrium problem in noncooperative games, and others; See, for example, [1, 2, 4, 69, 1417] which are references therein.

A mapping 𝑆 𝐶 𝑋 is called nonexpansive if 𝑆 𝑥 𝑆 𝑦 𝑥 𝑦 for all 𝑥 , 𝑦 𝐶 . Denote by 𝐹 ( 𝑆 ) the set of fixed points of 𝑆 , that is, 𝐹 ( 𝑆 ) = { 𝑥 𝐶 𝑆 𝑥 = 𝑥 } . Very recently, W. Takahashi and K. Zembayashi [18] proposed an iterative algorithm for finding a common element of the solution set of the equilibrium problem (1.4) and the set of fixed points of a relatively nonexpansive mapping 𝑆 in a Banach space 𝑋 . They also studied the strong and weak convergence of the sequences generated by their algorithm. In particular, they proposed the following iterative algorithm:

𝑥 0 𝑦 𝐶 , 𝑛 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 𝑛 + 1 𝛼 𝑛 𝐽 𝑆 𝑥 𝑛 , 𝑢 𝑛 𝑢 𝐶 s u c h t h a t 𝑓 𝑛 + 1 , 𝑦 𝑟 𝑛 𝑦 𝑢 𝑛 , 𝐽 𝑢 𝑛 𝐽 𝑦 𝑛 𝐻 0 , 𝑦 𝐶 , 𝑛 = 𝑧 𝐶 𝜙 𝑧 , 𝑢 𝑛 𝜙 𝑧 , 𝑥 𝑛 , 𝑊 𝑛 = 𝑧 𝐶 𝑥 𝑛 𝑧 , 𝐽 𝑥 𝐽 𝑥 𝑛 , 𝑥 0 𝑛 + 1 = Π 𝐻 𝑛 𝑊 𝑛 𝑥 , 𝑛 0 , ( 1 . 6 ) where 𝜙 ( 𝑥 , 𝑦 ) = 𝑥 2 2 𝑥 , 𝐽 𝑦 + 𝑦 2 for all 𝑥 , 𝑦 𝑋 , { 𝛼 𝑛 } [ 0 , 1 ] , and { 𝑟 𝑛 } [ 𝑎 , ) for some 𝑎 > 0 . They proved that the sequence { 𝑥 𝑛 } generated by the above algorithm converges strongly to Π 𝐹 ( 𝑆 ) 𝐸 𝑃 ( 𝑓 ) 𝑥 0 , where Π 𝐹 ( 𝑆 ) 𝐸 𝑃 ( 𝑓 ) is the generalized projection of 𝑋 onto 𝐹 ( 𝑆 ) 𝐸 𝑃 ( 𝑓 ) . They have also studied the weak convergence of the sequence { 𝑥 𝑛 } generated by the following algorithm:

𝑢 0 𝑥 𝑋 , 𝑛 𝑥 𝐶 s u c h t h a t 𝑓 𝑛 + 1 , 𝑦 𝑟 𝑛 𝑦 𝑥 𝑛 , 𝐽 𝑥 𝑛 𝐽 𝑢 𝑛 𝑢 0 , 𝑦 𝐶 , 𝑛 + 1 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 𝑛 + 1 𝛼 𝑛 𝐽 𝑆 𝑥 𝑛 , 𝑛 0 , ( 1 . 7 ) to 𝑧 𝐹 ( 𝑆 ) 𝐸 𝑃 ( 𝑓 ) , where 𝑧 = l i m 𝑛 Π 𝐹 ( 𝑆 ) 𝐸 𝑃 ( 𝑓 ) 𝑥 𝑛 .

Let 𝐶 be a nonempty closed convex subset of a uniformly smooth and uniformly convex Banach space 𝑋 . Let 𝐴 𝐶 𝑋 be an 𝛼 -inverse-strongly monotone mapping and 𝑓 𝐶 × 𝐶 a bifunction satisfying the following conditions:

(A1) 𝑓 ( 𝑥 , 𝑥 ) = 0 for all 𝑥 𝐶 ;(A2) 𝑓 is monotone, that is, 𝑓 ( 𝑥 , 𝑦 ) + 𝑓 ( 𝑦 , 𝑥 ) 0 , for all 𝑥 , 𝑦 𝐶 ;(A3) for all 𝑥 , 𝑦 , 𝑧 𝐶 , l i m s u p 𝑡 0 𝑓 ( 𝑡 𝑧 + ( 1 𝑡 ) 𝑥 , 𝑦 ) 𝑓 ( 𝑥 , 𝑦 ) ;(A4) for all 𝑥 𝐶 , 𝑓 ( 𝑥 , ) is convex and lower semicontinuous.

Let 𝑆 1 , 𝑆 2 𝐶 𝐶 be two relatively nonexpansive mappings such that 𝐹 ( 𝑆 1 ) 𝐹 ( 𝑆 2 ) 𝐸 𝑃 . Let { 𝑥 𝑛 } be the sequence generated by

𝑥 0 𝐶 , 𝐶 0 𝑧 = 𝐶 ; 𝑛 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 𝑛 + 1 𝛼 𝑛 𝐽 𝑆 1 𝑥 𝑛 , 𝑦 𝑛 = 𝐽 1 𝛽 𝑛 𝐽 𝑥 𝑛 + 1 𝛽 𝑛 𝐽 𝑆 2 𝑧 𝑛 , 𝑢 𝑛 𝑢 𝐶 s u c h t h a t 𝑓 𝑛 , 𝑦 + 𝐴 𝑢 𝑛 , 𝑦 𝑢 𝑛 1 + 𝑟 𝑛 𝑦 𝑢 𝑛 , 𝐽 𝑢 𝑛 𝐽 𝑦 𝑛 𝐶 0 , 𝑦 𝐶 , 𝑛 + 1 = 𝑣 𝐶 𝑛 𝜙 𝑣 , 𝑢 𝑛 𝛽 𝑛 𝜙 𝑣 , 𝑥 𝑛 + 1 𝛽 𝑛 𝜙 𝑣 , 𝑧 𝑛 𝜙 𝑣 , 𝑥 𝑛 ; 𝑥 𝑛 + 1 = Π 𝐶 𝑛 + 1 𝑥 0 , 𝑛 0 . ( 1 . 8 ) Zhang [1] proved the strong convergence of the sequence { 𝑥 𝑛 } to Π 𝐹 ( 𝑆 1 ) 𝐹 ( 𝑆 2 ) 𝐸 𝑃 𝑥 0 under appropriate conditions.

On the other hand, a classic method of solving 0 𝑇 𝑥 in a Hilbert space 𝐻 is the proximal point algorithm which generates, for any starting point 𝑥 0 𝐻 , a sequence { 𝑥 𝑛 } in 𝐻 by the iterative scheme

𝑥 𝑛 + 1 = 𝐽 𝑟 𝑛 𝑥 𝑛 , 𝑛 = 0 , 1 , 2 , , ( 1 . 9 ) where { 𝑟 𝑛 } is a sequence in ( 0 , ) , 𝐽 𝑟 = ( 𝐼 + 𝑟 𝑇 ) 1 for each 𝑟 > 0 is the resolvent operator for 𝑇 , and 𝐼 is the identity operator on 𝐻 . This algorithm was first introduced by Martinet [19] and further studied by Rockafellar [20] in the framework of a Hilbert space 𝐻 . Later several authors studied (1.9) and its variants in the setting of a Hilbert space 𝐻 or in a Banach space 𝑋 ; See, for example, [15, 2125] and references therein. Very recently, Li and Song [24] introduced and studied the following iterative scheme:

𝑥 0 𝑦 𝑋 c h o s e n a r b i t r a r i l y , 𝑛 = 𝐽 1 𝛽 𝑛 𝐽 𝑥 𝑛 + 1 𝛽 𝑛 𝐽 𝐽 𝑟 𝑛 𝑥 𝑛 , 𝑥 𝑛 + 1 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 0 + 1 𝛼 𝑛 𝐽 𝑦 𝑛 , 𝑛 = 0 , 1 , 2 , , ( 1 . 1 0 ) where 𝐽 𝑟 = ( 𝐽 + 𝑟 𝑇 ) 1 𝐽 and 𝐽 is the duality mapping on 𝑋 .

Algorithm (1.10) covers, as special cases, the algorithms introduced by Kohsaka and Takahashi [23] and Kamimura et al. [22] in a smooth and uniformly convex Banach space 𝑋 .

Let 𝑋 be a uniformly smooth and uniformly convex Banach space, and let 𝐶 be a nonempty closed convex subset of 𝑋 . Let 𝑇 𝑋 2 𝑋 be a maximal monotone operator such that:

(A5) 𝑇 1 0 𝐸 𝑃 ( 𝑓 ) .

In addition, for each 𝑟 > 0 , define a mapping 𝑇 𝑟 𝑋 𝐶 as follows:

𝑇 𝑟 1 ( 𝑥 ) = 𝑧 𝐶 𝑓 ( 𝑧 , 𝑦 ) + 𝑟 𝑦 𝑧 , 𝐽 𝑧 𝐽 𝑥 0 , 𝑦 𝐶 ( 1 . 1 1 ) for all 𝑥 𝑋 .

Very recently, utilizing the ideas of the above algorithms in [15, 16, 18, 21, 22, 24], we [17] introduced two iterative methods for finding an element of 𝑇 1 0 𝐸 𝑃 ( 𝑓 ) and established the following strong and weak convergence theorems.

Theorem 1.1 (see [17]). Suppose that conditions (A1)–(A5) are satisfied and let 𝑥 0 𝑋 be chosen arbitrarily. Consider the sequence 𝑥 𝑛 + 1 = Π 𝐻 𝑛 𝑊 𝑛 𝑥 0 , 𝑛 = 0 , 1 , 2 , , ( 1 . 1 2 ) where 𝐻 𝑛 = 𝑧 𝐶 𝜙 𝑧 , 𝑇 𝑟 𝑛 𝑦 𝑛 𝛼 𝑛 𝜙 𝑧 , 𝑥 0 + 1 𝛼 𝑛 𝜙 𝑧 , 𝑥 𝑛 , 𝑊 𝑛 = 𝑧 𝐶 𝑥 𝑛 𝑧 , 𝐽 𝑥 0 𝐽 𝑥 𝑛 , y 0 𝑛 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 0 + 1 𝛼 𝑛 𝛽 𝑛 𝐽 𝑥 𝑛 + 1 𝛽 𝑛 𝐽 𝐽 𝑟 𝑛 𝑥 𝑛 , ( 1 . 1 3 ) 𝑇 𝑟 is defined by (1.11), { 𝛼 𝑛 } , { 𝛽 𝑛 } [ 0 , 1 ] satisfy l i m 𝑛 𝛼 𝑛 = 0 , l i m i n f 𝑛 𝛽 𝑛 ( 1 𝛽 𝑛 ) > 0 , and { 𝑟 𝑛 } ( 0 , ) satisfies l i m i n f 𝑛 𝑟 𝑛 > 0 . Then, the sequence { 𝑥 𝑛 } converges strongly to Π 𝑇 1 0 𝐸 𝑃 ( 𝑓 ) 𝑥 0 , where Π 𝑇 1 0 𝐸 𝑃 ( 𝑓 ) is the generalized projection of 𝑋 onto 𝑇 1 0 𝐸 𝑃 ( 𝑓 ) .

Theorem 1.2 (see [17]). Suppose that conditions (A1)–(A5) are satisfied and let 𝑥 0 𝑋 be chosen arbitrarily. Consider the sequence 𝑥 𝑛 + 1 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 0 + 1 𝛼 𝑛 𝛽 𝑛 𝐽 𝑇 𝑟 𝑛 𝑥 𝑛 + 1 𝛽 𝑛 𝐽 𝐽 𝑟 𝑛 𝑇 𝑟 𝑛 𝑥 𝑛 , 𝑛 = 0 , 1 , 2 , , ( 1 . 1 4 ) where 𝑇 𝑟 is defined by (1.11), { 𝛼 𝑛 } , { 𝛽 𝑛 } [ 0 , 1 ] satisfy the conditions 𝑛 = 0 𝛼 𝑛 < and l i m i n f 𝑛 𝛽 𝑛 ( 1 𝛽 𝑛 ) > 0 , and { 𝑟 𝑛 } ( 0 , ) satisfies l i m i n f 𝑛 𝑟 𝑛 > 0 . If 𝐽 is weakly sequentially continuous, then { 𝑥 𝑛 } converges weakly to an element 𝑧 𝑇 1 0 𝐸 𝑃 ( 𝑓 ) , where 𝑧 = l i m 𝑛 Π 𝑇 1 0 𝐸 𝑃 ( 𝑓 ) 𝑥 𝑛 .

The purpose of this paper is to introduce and study two new iterative methods for finding a common element of the solution set 𝐸 𝑃 of generalized equilibrium problem (1.3) and the set 𝑇 1 𝑇 0 1 0 for maximal monotone operators 𝑇 and 𝑇 in a uniformly smooth and uniformly convex Banach space 𝑋 . Firstly, motivated by Theorem 1.1 and a result of Zhang [1], we introduce a sequence { 𝑥 𝑛 } that converges strongly to Π 𝑇 1 0 𝑇 1 0 𝐸 𝑃 𝑥 0 under some appropriate conditions.

Secondly, inspired by Theorem 1.2 and a result of Zhang [1], we define a sequence that converges weakly to an element 𝑧 𝑇 1 𝑇 0 1 0 𝐸 𝑃 , where 𝑧 = l i m 𝑛 Π 𝑇 1 0 𝑇 1 0 𝐸 𝑃 𝑥 𝑛 (Section 4).

Our results represent a generalization of known results in the literature, including those in [1618, 24]. Our Theorems 3.1 and 4.2 are the extension and improvements of Theorems 1.1 and 1.2 in the following way:

(i)the problem of finding an element of 𝑇 1 𝑇 0 1 0 𝐸 𝑃 includes the one of finding an element of 𝑇 1 0 𝐸 𝑃 ( 𝑓 ) as a special case;(ii)the algorithms in this paper are very different from those in [17] because of considering the complexity involving the problem of finding an element of 𝑇 1 𝑇 0 1 0 𝐸 𝑃 .

2. Preliminaries

Throughout the paper, we denote the strong convergence, weak convergence, and weak convergence of a sequence { 𝑥 𝑛 } to a point 𝑥 𝑋 by 𝑥 𝑛 𝑥 , 𝑥 𝑛 𝑥 and 𝑥 𝑛 𝑥 , respectively.

Assumption 2.1. Let 𝑋 be a uniformly smooth and uniformly convex Banach space and let 𝐶 be a nonempty closed convex subset of 𝑋 . Let 𝐴 𝐶 𝑋 be an 𝛼 -inverse-strongly monotone mapping and let 𝑓 𝐶 × 𝐶 be a bifunction satisfying the conditions (A1)–(A4). Let 𝑇 , 𝑇 𝑋 2 𝑋 be two maximal monotone operators such that:(A5) 𝑇 1 𝑇 0 1 0 𝐸 𝑃 .

Recall that if 𝐶 is a nonempty closed convex subset of a Hilbert space 𝐻 , then the metric projection 𝑃 𝐶 𝐻 𝐶 of 𝐻 onto 𝐶 is nonexpansive. This fact actually characterizes Hilbert spaces and hence, it is not available in more general Banach spaces. In this connection, Alber [26] recently introduced a generalized projection operator Π 𝐶 in a Banach space 𝑋 which is an analogue of the metric projection in Hilbert spaces.

Consider the functional defined as in [26] by

𝜙 ( 𝑥 , 𝑦 ) = 𝑥 2 2 𝑥 , 𝐽 𝑦 + 𝑦 2 , 𝑥 , 𝑦 𝑋 . ( 2 . 1 ) It is clear that in a Hilbert space 𝐻 , (2.1) reduces to 𝜙 ( 𝑥 , 𝑦 ) = 𝑥 𝑦 2 , 𝑥 , 𝑦 𝐻 .

The generalized projection Π 𝐶 𝑋 𝐶 is a mapping that assigns to an arbitrary point 𝑥 𝑋 the minimum point of the functional 𝜙 ( 𝑦 , 𝑥 ) , that is, Π 𝐶 𝑥 = 𝑥 , where 𝑥 is the solution to the minimization problem

𝜙 𝑥 , 𝑥 = m i n 𝑦 𝐶 𝜙 ( 𝑦 , 𝑥 ) . ( 2 . 2 )

The existence and uniqueness of the operator Π 𝐶 follow from the properties of the functional 𝜙 ( 𝑥 , 𝑦 ) and strict monotonicity of the mapping 𝐽 ; See, for example, [27]. In a Hilbert space, Π 𝐶 = 𝑃 𝐶 . From [26], in a smooth, strictly convex and reflexive Banach space 𝑋 , we have

( ) 𝑦 𝑥 2 ) 𝜙 ( 𝑦 , 𝑥 ) ( 𝑦 + 𝑥 2 , 𝑥 , 𝑦 𝑋 . ( 2 . 3 )

Moreover, by the property of subdifferential of convex functions, we easily get the following inequality:

𝜙 ( 𝑥 , 𝑦 ) 𝜙 𝑥 , 𝐽 1 ( 𝐽 𝑦 + 𝐽 𝑧 ) 2 𝑦 𝑥 , 𝐽 𝑧 , 𝑥 , 𝑦 , 𝑧 𝑋 . ( 2 . 4 )

Let 𝑆 be a mapping from 𝐶 into itself. A point 𝑝 in 𝐶 is called an asymptotic fixed point of 𝑆 [28] if 𝐶 contains a sequence { 𝑥 𝑛 } which converges weakly to 𝑝 such that 𝑆 𝑥 𝑛 𝑥 𝑛 0 . The set of asymptotic fixed points of 𝑆 is denoted by 𝐹 ( 𝑆 ) . A mapping 𝑆 from 𝑆 into itself is called relatively nonexpansive [18, 29, 30] if 𝐹 ( 𝑆 ) = 𝐹 ( 𝑆 ) and 𝜙 ( 𝑝 , 𝑆 𝑥 ) 𝜙 ( 𝑝 , 𝑥 ) , for all 𝑥 𝐶 and 𝑝 𝐹 ( 𝑆 ) .

Observe that, if 𝑋 is a reflexive, strictly convex and smooth Banach space, then for any 𝑥 , 𝑦 𝑋 , 𝜙 ( 𝑥 , 𝑦 ) = 0 if and only if 𝑥 = 𝑦 . To this end, it is sufficient to show that if 𝜙 ( 𝑥 , 𝑦 ) = 0 , then 𝑥 = 𝑦 . Actually, from (2.3), we have 𝑥 = 𝑦 , which implies that 𝑥 , 𝐽 𝑦 = 𝑥 2 = 𝑦 2 . From the definition of 𝐽 , we have 𝐽 𝑥 = 𝐽 𝑦 and therefore, 𝑥 = 𝑦 . For further details, we refer to [31].

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

Lemma 2.2 (see [32]). Let 𝑋 be a smooth and uniformly convex Banach space and let { 𝑥 𝑛 } and { 𝑦 𝑛 } be two sequences of 𝑋 . If 𝜙 ( 𝑥 𝑛 , 𝑦 𝑛 ) 0 and either { 𝑥 𝑛 } or { 𝑦 𝑛 } is bounded, then 𝑥 𝑛 𝑦 𝑛 0 .

Lemma 2.3 (see [26, 32]). Let 𝐶 be a nonempty closed convex subset of a smooth, strictly convex and reflexive Banach space 𝑋 , 𝑥 𝑋 and 𝑧 𝐶 . Then 𝑧 = Π 𝐶 𝑥 𝑦 𝑧 , 𝐽 𝑥 𝐽 𝑧 0 , 𝑦 𝐶 . ( 2 . 5 )

Lemma 2.4 (see [26, 32]). Let 𝐶 be a nonempty closed convex subset of a smooth, strictly convex and reflexive Banach space 𝑋 . Then 𝜙 𝑥 , Π 𝐶 𝑦 Π + 𝜙 𝐶 𝑦 , 𝑦 𝜙 ( 𝑥 , 𝑦 ) , 𝑥 𝐶 , 𝑦 𝑋 . ( 2 . 6 )

Lemma 2.5 (see [33]). Let 𝑋 be a reflexive, strictly convex and smooth Banach space and let 𝑇 𝑋 2 𝑋 be a multivalued operator. Then (i) 𝑇 1 0 is closed and convex if 𝑇 is maximal monotone such that 𝑇 1 0 ;(ii) 𝑇 is maximal monotone if and only if 𝑇 is monotone with 𝑅 ( 𝐽 + 𝑟 𝑇 ) = 𝑋 for all 𝑟 > 0 .

Lemma 2.6 (see [34]). Let 𝑋 be a uniformly convex Banach space and let 𝑟 > 0 . Then there exists a strictly increasing, continuous and convex function 𝑔 [ 0 , 2 𝑟 ] such that 𝑔 ( 0 ) = 0 and 𝑡 𝑥 + ( 1 𝑡 ) 𝑦 2 𝑡 𝑥 2 + ( 1 𝑡 ) 𝑦 2 𝑡 ( 1 𝑡 ) 𝑔 ( 𝑥 𝑦 ) , ( 2 . 7 ) for all 𝑥 , 𝑦 𝐵 𝑟 and 𝑡 [ 0 , 1 ] , where 𝐵 𝑟 = { 𝑧 𝑋 𝑧 𝑟 } .

Lemma 2.7 (see [32]). Let 𝑋 be a smooth and uniformly convex Banach space and let 𝑟 > 0 . Then there exists a strictly increasing, continuous, and convex function 𝑔 [ 0 , 2 𝑟 ] such that 𝑔 ( 0 ) = 0 and 𝑔 ( 𝑥 𝑦 ) 𝜙 ( 𝑥 , 𝑦 ) , 𝑥 , 𝑦 𝐵 𝑟 . ( 2 . 8 )

The following result is due to Blum and Oettli [14].

Lemma 2.8 (see [14]). Let 𝐶 be a nonempty closed convex subset of a smooth, strictly convex and reflexive Banach space 𝑋 , 𝑓 𝐶 × 𝐶 a bifunction satisfying conditions (A1)–(A4), and 𝑟 > 0 and 𝑥 𝑋 . Then, there exists 𝑧 𝐶 such that 1 𝑓 ( 𝑧 , 𝑦 ) + 𝑟 𝑦 𝑧 , 𝐽 𝑧 𝐽 𝑥 0 , 𝑦 𝐶 . ( 2 . 9 )

Motivated by a result in [35] in a Hilbert space setting, Takahashi and Zembayashi [18] established the following lemma.

Lemma 2.9 (see [18]). Let 𝐶 be a nonempty closed convex subset of a uniformly smooth, strictly convex and reflexive Banach space 𝑋 , and 𝑓 𝐶 × 𝐶 a bifunction satisfying conditions (A1)–(A4). For 𝑟 > 0 and 𝑥 𝑋 , define a mapping 𝑇 𝑟 𝑋 𝐶 as follows: 𝑇 𝑟 1 ( 𝑥 ) = 𝑧 𝐶 𝑓 ( 𝑧 , 𝑦 ) + 𝑟 𝑦 𝑧 , 𝐽 𝑧 𝐽 𝑥 0 , 𝑦 𝐶 ( 2 . 1 0 ) for all 𝑥 𝑋 . Then (i) 𝑇 𝑟 is single-valued;(ii) 𝑇 𝑟 is a firmly nonexpansive-type mapping, that is, for all 𝑥 , 𝑦 𝑋 , 𝑇 𝑟 𝑥 𝑇 𝑟 𝑦 , 𝐽 𝑇 𝑟 𝑥 𝐽 𝑇 𝑟 𝑦 𝑇 𝑟 𝑥 𝑇 𝑟 𝑦 , 𝐽 𝑥 𝐽 𝑦 ; ( 2 . 1 1 ) (iii) 𝐹 ( 𝑇 𝑟 ) = 𝐹 ( 𝑇 𝑟 ) = 𝐸 𝑃 ( 𝑓 ) ;(iv) 𝐸 𝑃 ( 𝑓 ) is closed and convex.

Using Lemma 2.9, we have the following result.

Lemma 2.10 (see [18]). Let 𝐶 be a nonempty closed convex subset of a smooth, strictly convex and reflexive Banach space 𝑋 , 𝑓 𝐶 × 𝐶 a bifunction satisfying conditions (A1)–(A4), and 𝑟 > 0 . Then, for 𝑥 𝑋 and 𝑞 𝐹 ( 𝑇 𝑟 ) , 𝜙 𝑞 , 𝑇 𝑟 𝑥 𝑇 + 𝜙 𝑟 𝑥 , 𝑥 𝜙 ( 𝑞 , 𝑥 ) . ( 2 . 1 2 )

Utilizing Lemmas 2.8, 2.9, and 2.10, Zhang [1] derived the following result.

Proposition 2.11 (see [1]). Let 𝑋 be a smooth, strictly convex and reflexive Banach space and let 𝐶 be a nonempty closed convex subset of 𝑋 . Let 𝐴 𝐶 𝑋 be an 𝛼 -inverse-strongly monotone mapping, 𝑓 𝐶 × 𝐶 a bifunction satisfying conditions (A1)–(A4), and 𝑟 > 0 . Then (I) for 𝑥 𝑋 , there exists 𝑢 𝐶 such that 1 𝑓 ( 𝑢 , 𝑦 ) + 𝐴 𝑢 , 𝑦 𝑢 + 𝑟 𝑦 𝑢 , 𝐽 𝑢 𝐽 𝑥 0 , 𝑦 𝐶 ; ( 2 . 1 3 ) (II) if 𝑋 is additionally uniformly smooth and 𝐾 𝑟 𝐶 𝐶 is defined as 𝐾 𝑟 1 ( 𝑥 ) = 𝑢 𝐶 𝑓 ( 𝑢 , 𝑦 ) + 𝐴 𝑢 , 𝑦 𝑢 + 𝑟 𝑦 𝑢 , 𝐽 𝑢 𝐽 𝑥 0 , 𝑦 𝐶 , 𝑥 𝐶 , ( 2 . 1 4 ) then the mapping 𝐾 𝑟 has the following properties:(i) 𝐾 𝑟 is single-valued,(ii) 𝐾 𝑟 is a firmly nonexpansive-type mapping, that is, 𝐾 𝑟 𝑥 𝐾 𝑟 𝑦 , 𝐽 𝐾 𝑟 𝑥 𝐽 𝐾 𝑟 𝑦 𝐾 𝑟 𝑥 𝐾 𝑟 𝑦 , 𝐽 𝑥 𝐽 𝑦 , 𝑥 , 𝑦 𝑋 , ( 2 . 1 5 ) (iii) 𝐹 ( 𝐾 𝑟 ) = 𝐹 ( 𝐾 𝑟 ) = 𝐸 𝑃 ,(iv) 𝐸 𝑃 is a closed convex subset of 𝐶 ,(v) 𝜙 ( 𝑝 , 𝐾 𝑟 𝑥 ) + 𝜙 ( 𝐾 𝑟 𝑥 , 𝑥 ) 𝜙 ( 𝑝 , 𝑥 ) , for all 𝑝 𝐹 ( 𝐾 𝑟 ) .

Proof. Define a bifunction 𝐹 𝐶 × 𝐶 by 𝐹 ( 𝑥 , 𝑦 ) = 𝑓 ( 𝑥 , 𝑦 ) + 𝐴 𝑥 , 𝑦 𝑥 , 𝑥 , 𝑦 𝐶 . ( 2 . 1 6 ) It is easy to verify that 𝐹 satisfies the conditions (A1)–(A4). Therefore, the conclusions (I) and (II) follow immediately from Lemmas 2.8, 2.9, and 2.10.

Let 𝑇 , 𝑇 𝑋 2 𝑋 be two maximal monotone operators in a smooth Banach space 𝑋 . We denote the resolvent operators of 𝑇 and 𝑇 by 𝐽 𝑟 = ( 𝐽 + 𝑟 𝑇 ) 1 𝐽 and 𝐽 𝑟 = ( 𝐽 + 𝑟 𝑇 ) 1 𝐽 for each 𝑟 > 0 , respectively. Then 𝐽 𝑟 𝑋 𝐷 ( 𝑇 ) and 𝐽 𝑟 𝑋 𝐷 ( 𝑇 ) are two single-valued mappings. Also, 𝑇 1 0 = 𝐹 ( 𝐽 𝑟 ) and 𝑇 1 𝐽 0 = 𝐹 ( 𝑟 ) for each 𝑟 > 0 , where 𝐹 ( 𝐽 𝑟 ) and 𝐽 𝐹 ( 𝑟 ) are the sets of fixed points of 𝐽 𝑟 and 𝐽 𝑟 , respectively. For each 𝑟 > 0 , the Yosida approximations of 𝑇 and 𝑇 are defined by 𝐴 𝑟 = ( 𝐽 𝐽 𝐽 𝑟 ) / 𝑟 and 𝐴 𝑟 𝐽 = ( 𝐽 𝐽 𝑟 ) / 𝑟 , respectively. It is known that

𝐴 𝑟 𝐽 𝑥 𝑇 𝑟 𝑥 , 𝐴 𝑟 𝑇 𝐽 𝑥 𝑟 𝑥 , f o r e a c h 𝑟 > 0 , 𝑥 𝑋 . ( 2 . 1 7 )

Lemma 2.12 (see [23]). Let 𝑋 be a reflexive, strictly convex and smooth Banach space, and let 𝑇 𝑋 2 𝑋 be a maximal monotone operator with 𝑇 1 0 . Then, 𝜙 𝑧 , 𝐽 𝑟 𝑥 𝐽 + 𝜙 𝑟 𝑥 , 𝑥 𝜙 ( 𝑧 , 𝑥 ) , 𝑟 > 0 , 𝑧 𝑇 1 0 , 𝑥 𝑋 . ( 2 . 1 8 )

Lemma 2.13 (see [36]). Let { 𝑎 𝑛 } and { 𝑏 𝑛 } be two sequences of nonnegative real numbers such that 𝑎 𝑛 + 1 𝑎 𝑛 + 𝑏 𝑛 for all 𝑛 0 . If 𝑛 = 0 𝑏 𝑛 < , then l i m 𝑛 𝑎 𝑛 exists.

3. Strong Convergence Theorem

In this section, we prove a strong convergence theorem for finding a common element of the set of solutions for a generalized equilibrium problem and the set 𝑇 1 𝑇 0 1 0 for two maximal monotone operators 𝑇 and 𝑇 .

Theorem 3.1. Suppose that Assumption 2.1 is satisfied. Let 𝑥 0 𝑋 be chosen arbitrarily. Consider the sequence 𝑥 𝑛 + 1 = Π 𝐻 𝑛 𝑊 𝑛 𝑥 0 , 𝑛 = 0 , 1 , 2 , , ( 3 . 1 ) where 𝐻 𝑛 = 𝑧 𝐶 𝜙 𝑧 , 𝐾 𝑟 𝑛 𝑦 𝑛 𝛼 𝑛 + 𝛼 𝑛 𝛼 𝑛 𝛼 𝑛 𝜙 𝑧 , 𝑥 0 + 1 𝛼 𝑛 1 𝛼 𝑛 𝜙 𝑧 , 𝑥 𝑛 , 𝑊 𝑛 = 𝑧 𝐶 𝑥 𝑛 𝑧 , 𝐽 𝑥 0 𝐽 𝑥 𝑛 , 0 ̃ 𝑥 𝑛 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 0 + 1 𝛼 𝑛 𝛽 𝑛 𝐽 𝑥 𝑛 + 1 𝛽 𝑛 𝐽 𝐽 𝑟 𝑛 𝑥 𝑛 , 𝑦 𝑛 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 0 + 1 𝛼 𝑛 ̃ 𝛽 𝑛 𝐽 ̃ 𝑥 𝑛 + ̃ 𝛽 1 𝑛 𝐽 𝐽 𝑟 𝑛 ̃ 𝑥 𝑛 , ( 3 . 2 ) 𝐾 𝑟 is defined by (2.14), { 𝛼 𝑛 } , { 𝛽 𝑛 } , { 𝛼 𝑛 ̃ 𝛽 } , { 𝑛 } [ 0 , 1 ] satisfy l i m 𝑛 𝛼 𝑛 = 0 , l i m 𝑛 𝛼 𝑛 = 0 , l i m i n f 𝑛 𝛽 𝑛 1 𝛽 𝑛 > 0 , l i m i n f 𝑛 ̃ 𝛽 𝑛 ̃ 𝛽 1 𝑛 > 0 , ( 3 . 3 ) and { 𝑟 𝑛 } ( 0 , ) satisfies l i m i n f 𝑛 𝑟 𝑛 > 0 . Then, the sequence { 𝑥 𝑛 } converges strongly to Π 𝑇 1 0 𝑇 1 0 𝐸 𝑃 𝑥 0 , where Π 𝑇 1 0 𝑇 1 0 𝐸 𝑃 is the generalized projection of 𝑋 onto 𝑇 1 𝑇 0 1 0 𝐸 𝑃 .

Proof. For the sake of simplicity, we define 𝑢 𝑛 = 𝐾 𝑟 𝑛 𝑦 𝑛 , 𝑧 𝑛 = 𝐽 1 𝛽 𝑛 𝐽 𝑥 𝑛 + 1 𝛽 𝑛 𝐽 𝐽 𝑟 𝑛 𝑥 𝑛 , ̃ 𝑧 𝑛 = 𝐽 1 ̃ 𝛽 𝑛 𝐽 ̃ 𝑥 𝑛 + ̃ 𝛽 1 𝑛 𝐽 𝐽 𝑟 𝑛 ̃ 𝑥 𝑛 , ( 3 . 4 ) so that ̃ 𝑥 𝑛 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 0 + 1 𝛼 𝑛 𝐽 𝑧 𝑛 , 𝑦 𝑛 = 𝐽 1 𝛼 𝑛 𝐽 𝑥 0 + 1 𝛼 𝑛 𝐽 ̃ 𝑧 𝑛 . ( 3 . 5 ) We divide the proof into several steps.Step 1. We claim that 𝐻 𝑛 𝑊 𝑛 is closed and convex for each 𝑛 0 .
Indeed, it is obvious that 𝐻 𝑛 is closed and 𝑊 𝑛 is closed and convex for each 𝑛 0 . Let us show that 𝐻 𝑛 is convex. For 𝑧 1 , 𝑧 2 𝐻 𝑛 and 𝑡 ( 0 , 1 ) , put 𝑧 = 𝑡 𝑧 1 + ( 1 𝑡 ) 𝑧 2 . It is sufficient to show that 𝑧 𝐻 𝑛 . We first write 𝛾 𝑛 = 𝛼 𝑛 + 𝛼 𝑛 𝛼 𝑛 𝛼 𝑛 for each 𝑛 0 . Next, we prove that
𝜙 𝑧 , 𝑢 𝑛 𝛾 𝑛 𝜙 𝑧 , 𝑥 0 + 1 𝛾 𝑛 𝜙 𝑧 , 𝑥 𝑛 ( 3 . 6 ) is equivalent to 2 𝛾 𝑛 𝑧 , 𝐽 𝑥 0 + 2 1 𝛾 𝑛 𝑧 , 𝐽 𝑥