Skip to main content

Iterative approximation for split common fixed point problem involving an asymptotically nonexpansive semigroup and a total asymptotically strict pseudocontraction

Abstract

In this paper, we prove the strong convergence theorem for split feasibility problem involving a uniformly asymptotically regular nonexpansive semigroup and a total asymptotically strict pseudocontractive mapping in Hilbert spaces. Our main results improve and extend some recent results in the literature.

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

1 Introduction

In this paper, we assume that H is a real Hilbert space with the inner product , and the norm . Let I denote the identity operator on H. Let C and Q be nonempty, closed and convex subsets of real Hilbert spaces H 1 and H 2 , respectively. The split feasibility problem (SFP) is to find a point

xC such that AxQ,
(1.1)

where A: H 1 H 2 is a bounded linear operator. The SFP in finite-dimensional Hilbert spaces was first introduced by Censor and Elfving [1] for modeling inverse problems which arise from phase retrievals and in medical image reconstruction [2]. The SFP attracts the attention of many authors due to its application in signal processing. Various algorithms have been invented to solve it (see, for example, [311] and references therein).

Note that the split feasibility problem (1.1) can be formulated as a fixed point equation by using the fact

P C ( I γ A ( I P Q ) A ) x = x ;
(1.2)

that is, x solves SFP (1.1) if and only if x solves fixed point equation (1.2) (see [8] for details). This implies that we can use fixed point algorithms (see [1214]) to solve SFP. A popular algorithm that solves SFP (1.1) is due to Byrne’s CQ algorithm [2] which is found to be a gradient-projection method (GPM) in convex minimization. Subsequently, Byrne [3] applied KM iteration to the CQ algorithm, and Zhao and Yang [15] applied KM iteration to the perturbed CQ algorithm to solve the SFP. It is well known that the CQ algorithm and the KM algorithm for a split feasibility problem do not necessarily converge strongly in the infinite-dimensional Hilbert spaces.

Now let us recall the definitions of some operators that will be used in this paper.

Let T:HH be a mapping. A point xH is said to be a fixed point of T provided that Tx=x, and denote by F(T) the fixed point set of T.

Definition 1.1 The mapping T:HH is said to be

  1. (a)

    nonexpansive if

    TxTyxy,x,yH;
  2. (b)

    strictly pseudocontractive if there exists a constant k[0,1) such that

    T x T y 2 x y 2 +k ( x y ) ( T x T y ) 2 ,x,yH;
  3. (c)

    (k,{ μ n },{ ξ n },ϕ)-total asymptotically strict pseudocontractive if there exists a constant k[0,1) and sequences { μ n }[0,), { ξ n }[0,) with μ n 0 and ξ n 0 as n, and a continuous and strictly increasing function ϕ:[0,)[0,) with ϕ(0)=0 such that for all n1, x,yH,

    T n x T n y 2 x y 2 +k ( x y ) ( T x T y ) 2 + μ n ϕ ( x y ) + ξ n ,x,yH.

One parameter family Γ:={S(t):0t<} is said to be a (continuous) Lipschitzian semigroup on a real Hilbert space H if the following conditions are satisfied:

  1. (1)

    S(0)x=x for all xH;

  2. (2)

    S(s+t)=S(s)S(t) for all s,t0;

  3. (3)

    for each t>0, there exists a bounded measurable function L t :(0,)[0,) such that

    S ( t ) x S ( t ) y L t xy,x,yH;
  4. (4)

    for each xH, the mapping S()x from [0,) into H is continuous.

A Lipschitzian semigroup Γ is called nonexpansive (or contractive) if L t =1 for all t>0 and asymptotically nonexpansive if lim sup t L t 1, respectively. Let F(Γ) denote the common fixed point set of the semigroup Γ, i.e., F(Γ):={xK:S(t)x=x,t>0}.

Let H be a real Hilbert space, Γ:={S(t):0t<} be a continuous operator semigroup on H. Then Γ is said to be uniformly asymptotically regular (in short, u.a.r.) on H if for all h0 and any bounded subset C of H,

lim t sup x C S ( h ) ( S ( t ) x ) S ( t ) x =0.

The nonexpansive semigroup { σ t :t>0} defined by the following lemma is an example of u.a.r. operator semigroup. Other examples of u.a.r. operator semigroup can be found in Examples 17, 18 of [16].

Lemma 1.2 (See Lemma 2.7 of [17])

Let D be a bounded closed convex subset of H, and Γ:={S(t):t>0} be a nonexpansive semigroup on H such that F(Γ) is nonempty. For each h>0, set σ t (x)= 1 t 0 t S(s)xds, then

lim t sup x D S ( h ) ( σ t x ) σ t x =0.

Example 1.3 (See [18])

The set { σ t :t>0} defined by Lemma 1.2 is an u.a.r. nonexpansive semigroup.

Several authors have proved several convergence theorems using several iterative schemes for fixed points of nonexpansive semigroups in the literature. See, for example, [1622] and the references contained therein.

In this paper, we shall focus our attention on the following split common fixed point problem (SCFP):

find xC such that AxQ,
(1.3)

where A: H 1 H 2 is a bounded linear operator, {S(t):t0} is a uniformly asymptotically regular nonexpansive semigroup on H 1 and T is a uniformly L-Lipschitzian continuous and (k,{ μ n },{ ξ n },ϕ)-total asymptotically strict pseudocontractive mapping with nonempty fixed point sets C:= t 0 F(S(t)) and Q:=F(T), and denote the solution set of the two-operator SCFP by

Ω:={yC:AyQ}=C A 1 (Q).
(1.4)

Recall that t 0 F(S(t)) and F(T) are closed and convex subsets of H 1 and H 2 , respectively. If Ω, we have that Ω is a closed and convex subset of H 1 . The split common fixed point problem (SCFP) is a generalization of the split feasibility problem (SFP) and the convex feasibility problem (CFP) (see [2, 23]).

In order to solve (1.3), Censor and Segal [23] proposed and proved, in finite-dimensional spaces, the convergence of the following algorithm:

x n + 1 =S ( x n + γ A t ( T I ) A x n ) ,n1,
(1.5)

where γ(0, 2 λ ), with λ being the largest eigenvalue of the matrix A t A ( A t stands for matrix transposition) and S and T are quasi-nonexpansive operators.

In 2011, Moudafi [5] introduced the following relaxed algorithm:

x n + 1 =(1 α n ) y n + α n S y n ,n1,
(1.6)

where y n = x n +γ A (TI)A x n , β(0,1), α n (0,1), and γ(0, 1 λ β ), with λ being the spectral radius of the operator A A. Moudafi proved weak convergence result of algorithm (1.6) in Hilbert spaces where S and T are quasi-nonexpansive operators. We observe that strong convergence result can be obtained in the results of Moudafi [5] if a compactness-type condition like demicompactness is imposed on the operator S. Furthermore, we can also obtain a strong convergence result by suitably modifying algorithm (1.6).

Recently, Zhao and He [24] introduced the following viscosity approximation algorithm

x n + 1 = α n f( x n )+(1 α n ) ( ( 1 w n ) x n + w n S ( x n + γ A ( T I ) A x n ) ) ,n1,
(1.7)

where f: H 1 H 1 is a contraction of modulus ρ>0, w n (0, 1 2 ), γ(0, 1 λ ), with λ being the spectral radius of the operator A A, and they proved strong convergence results concerning (1.3) for quasi-nonexpansive operators S and T in real Hilbert spaces. Inspired by the work of Zhao and He [24], Moudafi [6] quite recently revisited the viscosity-type approximation method (1.7) above introduced in [24]. First, he proposed a simple proof of the strong convergence of the iterative sequence { x n } defined by (1.7) based on attracting operator properties and then proposed a modification of this algorithm (1.7) and proved its strong convergence (see Theorem 2.1 of [6]).

Very recently Chang et al. [25] proved the following convergence theorem for multiple-set split feasibility problem (MSSFP) (1.3) for a family of multi-valued quasi-nonexpansive mappings and a total asymptotically pseudocontractive mapping in infinitely dimensional Hilbert spaces.

Theorem 1.4 Let H 1 and H 2 be two real Hilbert spaces, A: H 1 H 2 be a bounded linear operator and A : H 2 H 1 be the adjoint of A. Let { S i } i = 1 : H 1 CB( H 1 ) be a family of multi-valued quasi-nonexpansive mappings and for each i1, S i is demiclosed at  0. Let T: H 2 H 2 be a uniformly L-Lipschitzian continuous and (k,{ μ n },{ ξ n },ϕ)-total asymptotically strict pseudocontractive mapping satisfying n = 1 μ n < and n = 1 ξ n <. Suppose that there exist constants M 0 >0, M 1 >0 such that ϕ(λ) M 0 λ 2 , λ> M 1 . Let C:= i = 1 F( S i ) and Q:=F(T). Assume that for each pC, S i p={p} for each i1. Let { x n } be the sequence generated by

{ x 1 H 1 chosen arbitrarily , x n + 1 = α 0 , n y n + i = 1 α i , n w i , n , w i , n S i y n , y n = x n + γ A ( T n I ) A x n , n 1 ,
(1.8)

where { α i , n }(0,1) and γ>0 satisfy the following conditions:

  1. (a)

    i = 0 α i , n =1 for each n1;

  2. (b)

    for each i1, lim inf n α 0 , n α i , n >0;

  3. (c)

    γ(0, 1 k A 2 ).

If Ω (the set of solutions of multiple-set split feasibility problem (1.3)) is nonempty, then both { x n } n = 1 and { y n } n = 1 converge weakly to some point xΩ. In addition, if there exists a positive integer m such that S m is semi-compact, then both { x n } n = 1 and { y n } n = 1 converge strongly to xΩ.

We observe on Theorem 1.4 that:

  1. (1)

    Theorem 1.4 gives a weak convergence result for multiple-set split feasibility problem (1.3) for a family of multi-valued quasi-nonexpansive mappings and a total asymptotically pseudocontractive mapping in infinitely dimensional Hilbert spaces. In order to get strong convergence, Chang et al. [25] imposed a compactness-type condition (semi-compactness) on the mappings { S i } i = 1 . This compactness condition appears strong as only few mappings are semi-compact.

  2. (2)

    It is an interesting problem to extend the results of Theorem 1.4 to the nonexpansive semigroup case so that strong convergence is obtained. In order to obtain a strong convergence result in Theorem 1.4 without compactness-type condition for the nonexpansive semigroup case, a modification of (1.8) is necessary. This modification could be an implicit iterative scheme or an explicit iterative scheme. In the implicit iterative scheme, the computation of the next iteration x n + 1 involves solving a nonlinear equation at every step of the iteration, a task which may pose the same difficulty level as the initial problem. Therefore, in order to get a strong convergence result for the split common fixed point problem for a nonexpansive semigroup case and a total asymptotically pseudocontractive mapping in infinitely dimensional Hilbert spaces without compactness-type condition, a modification of (1.8), which is an explicit iterative scheme, is necessary. This leads to the following natural question.

Question Can we modify the iterative scheme (1.8) so that strong convergence is guaranteed for a split common fixed point problem involving a uniformly asymptotically regular nonexpansive semigroup and a total asymptotically pseudocontractive mapping in infinitely dimensional Hilbert spaces without any compactness-type condition assumed?

Our interest in this paper is to answer the above question. We thus modify the iterative scheme (1.8) and prove a strong convergence result for the split common fixed point problem for a uniformly asymptotically regular nonexpansive semigroup and a total asymptotically pseudocontractive mapping in infinitely dimensional Hilbert spaces without any further compactness-type condition assumed. Our results improve the corresponding results of Chang et al. [25] and many recent and important results that the results of Chang et al. [25] improved and extended like Censor et al. [26, 27], Yang [10], Moudafi [28], Xu [9], Censor and Segal [23], Masad and Reich [29] and others.

2 Preliminaries

We first recall some definitions, notations and conclusions which will be needed in proving our main results.

  • x n x means that x n x strongly;

  • x n x means that x n x weakly.

Next, we state the following well-known lemmas which will be used in the sequel.

Lemma 2.1 Let H be a real Hilbert space. Then the following well-known results hold:

  1. (i)

    x + y 2 = x 2 +2x,y+ y 2 , x,yH;

  2. (ii)

    x + y 2 x 2 +2y,x+y, x,yH;

  3. (iii)

    λ x + ( 1 λ ) y 2 =λ x 2 +(1λ) y 2 λ(1λ)xy, x,yH, λ[0,1].

Lemma 2.2 (Chang et al. [25])

Let T:HH be a uniformly L-Lipschitzian continuous and (k,{ μ n },{ ξ n },ϕ)-total asymptotically strict pseudocontractive mapping, then IT is demiclosed at origin.

Lemma 2.3 (Alber et al. [30])

Let { λ n } and { γ n } be nonnegative, { α n } be positive real numbers such that

λ n + 1 λ n α n λ n + γ n ,n1.

Let for all n>1,

γ n α n c 1 and α n α.

Then λ n max{ λ 1 , K }, where K =(1+α) c 1 .

Lemma 2.4 (Xu [31])

Let { a n } be a sequence of nonnegative real numbers satisfying the following relation:

a n + 1 (1 α n ) a n + α n σ n + γ n ,n0,

where

  1. (i)

    { a n }[0,1], n = 1 α n =;

  2. (ii)

    lim sup n σ n 0;

  3. (iii)

    γ n 0, n = 1 γ n <.

Then a n 0 as n.

3 Main results

For solving the split common fixed point problem (1.3), we assume that the following conditions are satisfied:

  1. (1)

    H 1 and H 2 are two real Hilbert spaces, A: H 1 H 2 is a bounded linear operator and A : H 2 H 1 is the adjoint of A;

  2. (2)

    {S(t):t0} is a uniformly asymptotically regular nonexpansive semigroup on H 1 ;

  3. (3)

    T: H 2 H 2 is a uniformly L-Lipschitzian continuous and (k,{ μ n },{ ξ n },ϕ)-total asymptotically strict pseudocontractive mapping satisfying the following conditions:

  4. (i)

    n = 1 μ n <; n = 1 ξ n <;

  5. (ii)

    { α n } is a real sequence in (0,1) such that μ n =o( α n ), ξ n =o( α n ), lim n α n =0, n = 1 α n =;

  6. (iii)

    there exist constants M 0 >0, M 1 >0 such that ϕ(λ) M 0 λ 2 , λ> M 1 ;

  7. (iv)

    C:= t 0 F(S(t)), Q:=F(T) and Ω.

In this section, we introduce the following algorithm and prove its strong convergence for solving split common fixed point problem (1.3).

Theorem 3.1 Let H 1 , H 2 , A, A , {S(t):t0}, T, C, Q, k, { μ n }, { ξ n }, ϕ and L satisfy the above conditions (i)-(iv). Let { x n } be the sequence generated by x 1 H 1 ,

{ u n = ( 1 α n ) x n , y n = u n + γ A ( T n I ) A u n , x n + 1 = β n y n + ( 1 β n ) S ( t n ) y n , n 1 ,
(3.1)

where t n and { β n }(0,1) and γ>0 satisfy the following conditions:

  1. (a)

    0<ϵ β n b<1;

  2. (b)

    γ(0, 1 k A 2 ).

If Ω is nonempty, then the sequence { x n } n = 1 converges strongly to an element of Ω.

Proof Since ϕ is continuous, it follows that ϕ attains maximum (say M) in [0, M 1 ] and by our assumption, ϕ(λ) M 0 λ 2 , λ> M 1 . In either case, we have that

ϕ(λ)M+ M 0 λ 2 ,λ[0,).

Let x Ω. Then, by the convexity of 2 , we obtain

u n x 2 = ( 1 α n ) x n x 2 = ( 1 α n ) ( x n x ) + α n ( x ) 2 ( 1 α n ) x n x 2 + α n x 2 .
(3.2)

From (3.1) and Lemma 2.1(i), we obtain that

y n x 2 = u n x + γ A ( T n I ) A u n 2 = u n x 2 + 2 γ u n x , A ( T n I ) A u n + γ 2 A ( T n I ) A u n 2 .
(3.3)

Since

γ 2 A ( T n I ) A u n 2 = γ 2 A ( T n I ) A u n , A ( T n I ) A u n = γ 2 A A ( T n I ) A u n , ( T n I ) A u n γ 2 A 2 ( T n I ) A u n 2 ,
(3.4)

A x Q=F(T) and T is a total asymptotically strict pseudocontractive mapping, then we obtain

u n x , A ( T n I ) A u n = A ( u n x ) , ( T n I ) A u n = A ( u n x ) + ( T n I ) A u n ( T n I ) A u n , ( T n I ) A u n = T n A u n A x , ( T n I ) A u n ( T n I ) A u n 2 = 1 2 [ T n A u n A x 2 + ( T n I ) A u n 2 A u n A x 2 ] ( T n I ) A u n 2 1 2 [ A u n A x 2 + k ( T I ) A x n 2 + μ n ϕ ( A u n A x ) + ξ n ] + 1 2 [ ( T n I ) A u n 2 A u n A x 2 ] ( T n I ) A u n 2 k 1 2 ( T n I ) A u n 2 + μ n 2 ( M + M 0 A u n A x 2 ) + ξ n 2 .
(3.5)

Substituting (3.5) and (3.4) into (3.3), we have

y n x 2 u n x 2 γ ( 1 k γ A 2 ) ( T n I ) A u n 2 + μ n γ ( M + M 0 A u n A x 2 ) + γ ξ n .
(3.6)

Putting (3.6) and (3.2) into (3.1), we obtain

x n + 1 x 2 = β n ( y n x ) + ( 1 β n ) ( S ( t n ) y n x ) 2 β n y n x 2 + ( 1 β n ) S ( t n ) y n x 2 β n y n x 2 + ( 1 β n ) y n x 2 = y n x 2 u n x 2 γ ( 1 k γ A 2 ) ( T n I ) A u n 2 + μ n γ ( M + M 0 A u n A x 2 ) + γ ξ n ( 1 α n ) x n x 2 + α n x 2 γ ( 1 k γ A 2 ) ( T n I ) A u n 2 + μ n γ ( M + M 0 A 2 ( ( 1 α n ) x n x 2 + α n x 2 ) ) + γ ξ n = x n x 2 ( α n μ n γ M 0 A 2 ( 1 α n ) ) x n x 2 + α n x 2 + μ n γ M + γ α n M 0 A 2 x 2 + γ ξ n γ ( 1 k γ A 2 ) ( T n I ) A u n 2 x n x 2 ( α n μ n γ M 0 A 2 ( 1 α n ) ) x n x 2 + α n x 2 + μ n γ M + γ α n M 0 A 2 x 2 + γ ξ n = x n x 2 ( α n μ n γ M 0 A 2 ( 1 α n ) ) x n x 2 + σ n ,
(3.7)

where σ n = α n x 2 μ n γM+γ α n M 0 A 2 x 2 +γ ξ n . Since μ n =o( α n ) and ξ n =o( α n ), we may assume without loss of generality that there exist constants k 0 (0,1) and M 2 >0 such that for all n1,

μ n α n 1 k 0 M 0 γ A 2 ( 1 α n ) and σ n α n M 2 .

Thus, we obtain

x n + 1 x 2 x n x 2 α n k 0 x n x 2 + σ n .

By Lemma 2.3, we have that

x n x 2 max { x 1 x 2 , ( 1 + k 0 ) M 2 } .

Therefore, { x n } is bounded. Furthermore, the sequences { y n } and { u n } are bounded.

The rest of the proof will be divided into two parts.

Case 1. Suppose that there exists n 0 N such that { x n x } n = n 0 is nonincreasing. Then { x n x } n = 1 converges and

x n x 2 x n + 1 x 2 0,n.

From (3.6), we have that

x n + 1 x 2 y n x 2 x n x 2 γ ( 1 k γ A 2 ) ( T n I ) A x n 2 + μ n γ ( M + M 0 A x n A x 2 ) + γ ξ n x n x 2 γ ( 1 k γ A 2 ) ( T n I ) A x n 2 + μ n γ ( M + M 0 A x n A x 2 ) + γ ξ n + α n x 2 .

This implies that

γ ( 1 k γ A 2 ) ( T n I ) A x n 2 x n x 2 x n + 1 x 2 + μ n γ ( M + M 0 A x n A x 2 ) + γ ξ n + α n x 2

and

γ ( 1 k γ A 2 ) ( T n I ) A x n 2 0,n.

Hence, we obtain

( T n I ) A x n 0,n.
(3.8)

Also, we observe that

y n x n =γ A ( T n I ) A x n 0,n.

Using (3.6) and Lemma 2.1(iii) in (3.1), we have

x n + 1 x 2 = ( 1 β n ) ( y n x ) + β n ( S ( t n ) y n x ) 2 ( 1 β n ) y n x 2 + β n S ( t n ) y n x 2 ( 1 β n ) β n y n S ( t n ) y n = y n x 2 ( 1 β n ) β n y n S ( t n ) y n x n x 2 γ ( 1 k γ A 2 ) ( T n I ) A x n 2 + μ n γ ( M + M 0 A x n A x 2 ) + γ ξ n ( 1 β n ) β n y n S ( t n ) y n .

This implies from (3.2) and condition (b) that

( 1 β n ) β n y n S ( t n ) y n x n x 2 x n + 1 x 2 + μ n γ ( M + M 0 A x n A x 2 ) + γ ξ n x n x 2 x n + 1 x 2 + α n x 2 + μ n γ ( M + M 0 A x n A x 2 ) + γ ξ n .

From condition (a) we have

lim n y n S ( t n ) y n =0.
(3.9)

Hence, for any t0,

S ( t ) y n y n S ( t ) y n S ( t ) S ( t n ) y n + S ( t ) S ( t n ) y n S ( t n ) y n + S ( t n ) y n y n S ( t ) S ( t n ) y n S ( t n ) y n + 2 S ( t n ) y n y n 0 , n .
(3.10)

We obtain from (3.1) that

x n + 1 x n 2 = β n ( y n x n ) + ( 1 β n ) ( S ( t n ) y n x n ) 2 β n y n x n 2 + ( 1 β n ) S ( t n ) y n x n 2 β n y n x n 2 + ( 1 β n ) ( S ( t n ) y n y n + y n x n ) 2 .

Since lim n y n x n =0 and lim n y n S( t n ) y n =0, we have

lim n x n + 1 x n =0.

Consequently,

u n + 1 u n = ( 1 α n + 1 ) x n + 1 ( 1 α n ) x n | α n + 1 α n | x n + 1 + ( 1 α n ) x n + 1 x n 0 , n .
(3.11)

Using the fact that T is uniformly L-Lipschitzian, we have

T A u n A u n T A u n T n + 1 A u n + T n + 1 A u n T n + 1 A u n + 1 + T n + 1 A u n + 1 A u n + 1 + A u n + 1 A u n L A u n T n A u n + ( L + 1 ) A u n + 1 A u n + T n + 1 A u n + 1 A u n + 1 L A u n T n A u n + ( L + 1 ) A u n + 1 u n + T n + 1 A u n + 1 A u n + 1 .

From (3.8) and (3.11), we obtain

( T I ) A u n 0,n.
(3.12)

Since { x n } is bounded, there exists { x n j } of { x n } such that x n j z H 1 . Using the fact that x n j z H 1 and y n x n 0, n, we have that y n j z H 1 . Similarly, u n j z H 1 since u n x n 0, n.

We next show that z t 0 F(S(t))=C. Assume the contrary that zS(t)z for some t0. Then, by Opial’s condition, we obtain from (3.10) that

lim inf j y n j z < lim inf j y n j S ( t ) z lim inf j ( y n j S ( t ) y n j + S ( t ) y n j S ( t ) z ) y n j z .

This is a contradiction. Hence, z t 0 F(S(t))=C. On the other hand, since A is a linear bounded operator, it follows from u n j z H 1 that A u n j Az H 2 . Hence, from (3.12), we have that

TA u n j A u n j =TA u n j A u n j 0,j.

Since T is demiclosed at zero, we have that AzF(T)=Q. Hence zΩ.

Next, we prove that { x n } converges strongly to z. From (3.6) and Lemma 2.1(ii), we have

x n + 1 z 2 y n z 2 u n z 2 γ ( 1 k γ A 2 ) ( T n I ) A u n 2 + μ n γ ( M + M 0 A u n A z 2 ) + γ ξ n u n z 2 + μ n γ ( M + M 0 A u n A z 2 ) + γ ξ n u n z 2 + M μ n + γ ξ n = ( 1 α n ) ( x n z ) α n z 2 + M μ n + γ ξ n ( 1 α n ) 2 x n z 2 2 α n u n z , z + M μ n + γ ξ n ( 1 α n ) x n z 2 2 α n u n z , z + M μ n + γ ξ n ,
(3.13)

where M >γ sup n 1 (M+ M 0 A u n A z 2 )>0. It is clear that 2 u n z,z0, n and n = 1 M μ n <; n = 1 γ ξ n <. Now, using Lemma 2.4 in (3.13), we have x n z0. So x n z as n.

Case 2. Assume that { x n x } is not a monotonically decreasing sequence. Set Γ n = x n x 2 and let τ:NN be a mapping for all n n 0 (for some n 0 large enough) by

τ(n):=max{kN:kn, Γ k Γ k + 1 }.

Clearly, τ is a nondecreasing sequence such that τ(n) as n and

Γ τ ( n ) Γ τ ( n ) + 1 ,n n 0 .

From (3.9), it is easy to see that

lim n y τ ( n ) S ( t τ ( n ) ) y τ ( n ) =0.

Furthermore, we can show that

( T I ) A u τ ( n ) 0,n.

By a similar argument as above in Case 1, we conclude immediately that x τ ( n ) , y τ ( n ) and u τ ( n ) weakly converge to z as τ(n). At the same time, from (3.13), we note that for all n n 0 ,

0 x τ ( n ) + 1 z 2 x τ ( n ) z 2 α τ ( n ) [ 2 u τ ( n ) z , z x τ ( n ) z 2 ] + M μ τ ( n ) + γ ξ τ ( n ) ,

which gives

x τ ( n ) z 2 2 u τ ( n ) z,z+ M μ τ ( n ) +γ ξ τ ( n ) .

Hence, we deduce that

lim n x τ ( n ) z=0.

Therefore,

lim n Γ τ ( n ) = lim n Γ τ ( n ) + 1 =0.

Furthermore, for n n 0 , it is easy to see that Γ τ ( n ) Γ τ ( n ) + 1 if nτ(n) (that is τ(n)<n) because Γ j Γ j + 1 for τ(n)+1jn. As a consequence, we obtain for all n n 0 ,

0 Γ n max{ Γ τ ( n ) , Γ τ ( n ) + 1 }= Γ τ ( n ) + 1 .

This shows that lim Γ n =0 and hence { x n } converges strongly to z. This completes the proof. □

Based on Lemma 1.2 and Example 1.3, we can deduce the following corollary from Theorem 3.1.

Corollary 3.2 Let H 1 and H 2 be two real Hilbert spaces, A: H 1 H 2 be a bounded linear operator and A : H 2 H 1 be the adjoint of A. Let :={S(t):0t<} be a one-parameter nonexpansive semigroup on H 1 . Let T: H 2 H 2 be a uniformly L-Lipschitzian continuous and (k,{ μ n },{ ξ n },ϕ)-total asymptotically strict pseudocontractive mapping satisfying the following conditions:

  1. (i)

    n = 1 μ n <; n = 1 ξ n <;

  2. (ii)

    { α n } is a real sequence in (0,1) such that μ n =o( α n ), ξ n =o( α n ), lim n α n =0; n = 1 α n =;

  3. (iii)

    there exist constants M 0 >0, M 1 >0 such that ϕ(λ) M 0 λ 2 , λ> M 1 .

Let C:= t 0 F(S(t)), Q:=F(T) and Ω. Let { x n } be the sequence generated by x 1 H 1 ,

{ u n = ( 1 α n ) x n , y n = u n + γ A ( T n I ) A u n , x n + 1 = β n y n + ( 1 β n ) ( 1 t n 0 t n T ( u ) d u ) y n , n 1 ,
(3.14)

where { β n }(0,1) and γ>0 satisfy the following conditions:

  1. (a)

    0<ϵ β n b<1;

  2. (b)

    γ(0, 1 k A 2 ).

If Ω is nonempty, then the sequence { x n } n = 1 converges strongly to an element of Ω.

A strong mean convergence theorem for nonexpansive mappings was first established for odd mappings by Baillon [32] and it was later generalized to that of nonlinear semigroups by Reich [33]. It follows from the above proof that Theorem 3.1 is valid for nonexpansive mappings. Thus, we also have the following mean ergodic theorem for nonexpansive mappings in Hilbert spaces.

Corollary 3.3 Let H 1 and H 2 be two real Hilbert spaces, A: H 1 H 2 be a bounded linear operator and A : H 2 H 1 be the adjoint of A. Let S be a nonexpansive mapping on  H 1 . Let T: H 2 H 2 be a uniformly L-Lipschitzian continuous and (k,{ μ n },{ ξ n },ϕ)-total asymptotically strict pseudocontractive mapping satisfying the following conditions:

  1. (i)

    n = 1 μ n <; n = 1 ξ n <;

  2. (ii)

    { α n } is a real sequence in (0,1) such that μ n =o( α n ), ξ n =o( α n ), lim n α n =0, n = 1 α n =;

  3. (iii)

    there exist constants M 0 >0, M 1 >0 such that ϕ(λ) M 0 λ 2 , λ> M 1 .

Let C:= t 0 F(S(t)), Q:=F(T) and Ω. Let { x n } be the sequence generated by x 1 H 1 ,

{ u n = ( 1 α n ) x n , y n = u n + γ A ( T n I ) A u n , x n + 1 = β n y n + ( 1 β n ) ( 1 n + 1 j = 0 n S j y n ) , n 1 ,
(3.15)

where { β n }(0,1) and γ>0 satisfy the following conditions:

  1. (a)

    0<ϵ β n b<1;

  2. (b)

    γ(0, 1 k A 2 ).

If Ω is nonempty, then the sequence { x n } n = 1 converges strongly to an element of Ω.

References

  1. Censor Y, Elfving T: A multiprojection algorithm using Bregman projections in a product space. Numer. Algorithms 1994, 8(2–4):221–239.

    Article  MathSciNet  Google Scholar 

  2. Byrne C: Iterative oblique projection onto convex sets and the split feasibility problem. Inverse Probl. 2002, 18(2):441–453. 10.1088/0266-5611/18/2/310

    Article  MathSciNet  Google Scholar 

  3. Byrne C: A unified treatment of some iterative algorithms in signal processing and image reconstruction. Inverse Probl. 2004, 20(1):103–120. 10.1088/0266-5611/20/1/006

    Article  MathSciNet  Google Scholar 

  4. Maingé PE: The viscosity approximation process for quasi-nonexpansive mappings in Hilbert spaces. Comput. Math. Appl. 2010, 59(1):74–79. 10.1016/j.camwa.2009.09.003

    Article  MathSciNet  Google Scholar 

  5. Moudafi A: A note on the split common fixed-point problem for quasi-nonexpansive operators. Nonlinear Anal. 2011, 74(12):4083–4087. 10.1016/j.na.2011.03.041

    Article  MathSciNet  Google Scholar 

  6. Moudafi A: Viscosity-type algorithms for the split common fixed-point problem. Adv. Nonlinear Var. Inequal. 2013, 16: 61–68.

    MathSciNet  Google Scholar 

  7. Qu B, Xiu N: A note on the CQ algorithm for the split feasibility problem. Inverse Probl. 2005, 21(5):1655–1665. 10.1088/0266-5611/21/5/009

    Article  MathSciNet  Google Scholar 

  8. Xu H-K: Iterative methods for the split feasibility problem in infinite-dimensional Hilbert spaces. Inverse Probl. 2010., 26(10): Article ID 105018

    Google Scholar 

  9. Xu H-K: A variable Krasnosel’skii-Mann algorithm and the multiple-set split feasibility problem. Inverse Probl. 2006, 22(6):2021–2034. 10.1088/0266-5611/22/6/007

    Article  Google Scholar 

  10. Yang Q: The relaxed CQ algorithm solving the split feasibility problem. Inverse Probl. 2004, 20(4):1261–1266. 10.1088/0266-5611/20/4/014

    Article  Google Scholar 

  11. Yang Q, Zhao J: Generalized KM theorems and their applications. Inverse Probl. 2006, 22(3):833–844. 10.1088/0266-5611/22/3/006

    Article  Google Scholar 

  12. Yao Y, Chen R, Liou Y-C: A unified implicit algorithm for solving the triple-hierarchical constrained optimization problem. Math. Comput. Model. 2012, 55(3–4):1506–1515. 10.1016/j.mcm.2011.10.041

    Article  MathSciNet  Google Scholar 

  13. Yao Y, Cho Y-J, Liou Y-C: Hierarchical convergence of an implicit double-net algorithm for nonexpansive semigroups and variational inequalities. Fixed Point Theory Appl. 2011., 2011: Article ID 101

    Google Scholar 

  14. Yao Y, Liou Y-C, Kang SM: Two-step projection methods for a system of variational inequality problems in Banach spaces. J. Glob. Optim. 2013, 55(4):801–811. 10.1007/s10898-011-9804-0

    Article  MathSciNet  Google Scholar 

  15. Zhao J, Yang Q: Several solution methods for the split feasibility problem. Inverse Probl. 2005, 21(5):1791–1799. 10.1088/0266-5611/21/5/017

    Article  Google Scholar 

  16. Aleyner A, Censor Y: Best approximation to common fixed points of a semigroup of nonexpansive operators. J. Nonlinear Convex Anal. 2005, 6(1):137–151.

    MathSciNet  Google Scholar 

  17. Chen R, Song Y: Convergence to common fixed point of nonexpansive semigroup. J. Comput. Appl. Math. 2007, 200: 566–575. 10.1016/j.cam.2006.01.009

    Article  MathSciNet  Google Scholar 

  18. Song Y, Xu S: Strong convergence theorems for nonexpansive semigroup in Banach spaces. J. Math. Anal. Appl. 2008, 338: 152–161. 10.1016/j.jmaa.2007.05.021

    Article  MathSciNet  Google Scholar 

  19. Cianciaruso F, Marino G, Muglia L: Iterative methods for equilibrium and fixed point problems for nonexpansive semigroups in Hilbert spaces. J. Optim. Theory Appl. 2010, 146: 491–509. 10.1007/s10957-009-9628-y

    Article  MathSciNet  Google Scholar 

  20. Li S, Li L, Su Y: General iterative methods for a one-parameter nonexpansive semigroup in Hilbert space. Nonlinear Anal. 2009, 70: 3065–3071. 10.1016/j.na.2008.04.007

    Article  MathSciNet  Google Scholar 

  21. Suzuki T: On strong convergence to common fixed points of nonexpansive semigroup in Hilbert spaces. Proc. Am. Math. Soc. 2002, 131: 2133–2136.

    Article  Google Scholar 

  22. Wang X, Marino G, Muglia L: On the convergence of Mann and Ishikawa iterative processes for asymptotically ϕ -pseudocontractive mappings. Abstr. Appl. Anal. 2012. 10.1155/2012/850104

    Google Scholar 

  23. Censor Y, Segal A: The split common fixed point problem for directed operators. J. Convex Anal. 2009, 16(2):587–600.

    MathSciNet  Google Scholar 

  24. Zhao J, He S: Strong convergence of the viscosity approximation process for the split common fixed-point problem of quasi-nonexpansive mappings. J. Appl. Math. 2012., 2012: Article ID 438023

    Google Scholar 

  25. Chang SS, Lee HWJ, Chan CK, Wang L, Qin LJ: Split feasibility problem for quasi-nonexpansive multi-valued mappings and total asymptotically strict pseudo-contractive mapping. Appl. Math. Comput. 2013, 219: 10416–10424. 10.1016/j.amc.2013.04.020

    Article  MathSciNet  Google Scholar 

  26. Censor Y, Elfving T, Kopf N, Bortfeld T: The multiple-sets split feasibility problem and its applications. Inverse Probl. 2005, 21: 2071–2084. 10.1088/0266-5611/21/6/017

    Article  MathSciNet  Google Scholar 

  27. Censor Y, Motova A, Segal A: Perturbed projections and subgradient projections for the multiple-sets split feasibility problem. J. Math. Anal. Appl. 2007, 327: 1244–1256. 10.1016/j.jmaa.2006.05.010

    Article  MathSciNet  Google Scholar 

  28. Moudafi A: The split common fixed-point problem for demicontractive mappings. Inverse Probl. 2010, 26: 587–600.

    Article  MathSciNet  Google Scholar 

  29. Masad E, Reich S: A note on the multiple-set split feasibility problem in Hilbert spaces. J. Nonlinear Convex Anal. 2007, 8: 367–371.

    MathSciNet  Google Scholar 

  30. Alber Y, Espinola R, Lorenzo P: Strongly convergent approximations to fixed points of total asymptotically nonexpansive mappings. Acta Math. Sin. Engl. Ser. 2008, 24(6):1005–1022. 10.1007/s10114-007-6367-6

    Article  MathSciNet  Google Scholar 

  31. Xu H-K: Iterative algorithm for nonlinear operators. J. Lond. Math. Soc. 2002, 66(2):1–17.

    Google Scholar 

  32. Baillon JB: Un théorème de type ergodique pour les contractions non linéaires dans un espace de Hilbert. C. R. Acad. Sci. Paris 1975, 280: 1511–1514.

    MathSciNet  Google Scholar 

  33. Reich S: Almost convergence and nonlinear ergodic theorems. J. Approx. Theory 1978, 24: 269–272. 10.1016/0021-9045(78)90012-6

    Article  MathSciNet  Google Scholar 

Download references

Acknowledgements

This research was supported by National Research Council of Thailand (NRCT) and University of Phayao under Grant R020057216003.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Prasit Cholamjiak.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

All authors contributed equally to the writing of this paper. All authors read and approved the final manuscript.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 2.0 International License (https://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Cholamjiak, P., Shehu, Y. Iterative approximation for split common fixed point problem involving an asymptotically nonexpansive semigroup and a total asymptotically strict pseudocontraction. Fixed Point Theory Appl 2014, 131 (2014). https://doi.org/10.1186/1687-1812-2014-131

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/1687-1812-2014-131

Keywords