We investigate an algorithm for a common point of fixed points of a finite family of Lipschitz pseudocontractive mappings and solutions of a finite family of γ-inverse strongly accretive mappings. Our theorems improve and unify most of the results that have been proved in this direction for this important class of nonlinear mappings.
1. Introduction
Let C be a subset of a real Hilbert space H. Let A:C→H be a nonlinear mapping. The variational inequality problem for A and C is to
(1)findx*∈Csuchthat〈Ax*,v-x*〉≥0,∀v∈C.
The set of solutions of variational inequality problem is denoted by VI(C,A); that is,
(2)VI(C,A)={x*∈C:〈Ax*,x-x*〉≥0,∀x∈C}.
It is well known that variational inequality theory has emerged as an important tool in studying a wide class of numerous problems in variational inequalities, minimax problems, optimization, physics, and the Nash equilibrium problems in noncooperative games. Several numerical methods have been developed for solving variational inequalities and related optimization problems; see, for instance, [1–5] and the references therein.
A mapping A:C⊆H→H is said to be γ-inverse strongly accretive (or γ-inverse strongly monotone) if there exists a positive real number γ such that
(3)〈x-y,Ax-Ay〉≥γ∥Ax-Ay∥2,∀x,y∈C.
If A is γ-inverse strongly accretive, then inequality (3) implies that A is Lipschitzian with constant L:=1/γ; that is, ∥Ax-Ay∥≤(1/γ)∥x-y∥, for all x,y∈C. If in (3) we have that γ=0, then A is called accretive (or monotone).
Let C be a closed and convex subset of a real Hilbert space H. A mapping T:C→H is called a contraction mapping if there exists L∈[0,1) such that ∥Tx-Ty∥≤L∥x-y∥ for all x,y∈C. If L=1, then T is called nonexpansive. A mapping T:C→E is called λ-strictly pseudocontractive of Browder-Petryshyn type [6] if and only if there exists λ∈(0,1) such that
(4)∥Tx-Ty∥2≤∥x-y∥2+λ∥(I-T)x-(I-T)y∥2hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh∀x,y∈C.Tis called pseudocontractive if
(5)∥Tx-Ty∥2≤∥x-y∥2+∥(I-T)x-(I-T)y∥2,hhhhhhhhhhhhhhhhhhhhhhhhhhhhhh∀x,y∈C.
We note that inequalities (4) and (5) can be equivalently written as
(6)〈Tx-Ty,x-y〉≤∥x-y∥2-k∥(x-Tx)-(y-Ty)∥2hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh∀x,y∈C,
for some k>0 and
(7)〈Tx-Ty,x-y〉≤∥x-y∥2∀x,y∈C,
respectively. We remark that T is pseudocontractive if and only if A:=(I-T) is accretive. A point x∈C is a fixed point of T if Tx=x and we denote by F(T) the set of fixed points of T; that is, F(T)={x∈C:Tx=x}.
We observe that in a real Hilbert space H a class of pseudocontractive mappings includes the class of λ-strictly pseudocontractive mappings and hence the classes of nonexpansive and contraction mappings.
Closely related to the variational inequality problems is the problem of finding fixed points of nonexpansive mappings, λ-strict pseudocontraction mappings or pseudocontractive mappings which is the current interest in functional analysis. Several researchers considered a unified approach that approximates a common point of fixed point of nonlinear problems and solutions of variational inequality problems and solutions of variational inequality problems; see, for example, [7–18] and the references therein.
In [19], Takahashi and Toyoda studied the problem of finding a common point of fixed points of a nonexpansive mapping and solutions of a variational inequality problem (1) by considering the following iterative algorithm:
(8)x0∈C,xn+1=αnxn+(1-αn)TPC(xn-λnAxn),n=0,1,…,
where {αn} is a sequence in (0,1), {λn} is a positive sequence, T:C→C is a nonexpansive mapping, and A:C→H is an γ-inverse strongly accretive mapping. They showed that the sequence {xn} generated by (8) converges weakly to some z∈VI(C,A)∩F(S) provided that the control sequences satisfy some restrictions.
Iiduka and Takahashi [20] reconsidered the common element problem via the following iterative algorithm:
(9)x1=x∈C,xn+1=αnx+(1-αn)TPC(xn-λnAxn),n=0,1,…,
where T:C→C is a nonexpansive mapping, A:C→H is a γ-inverse-strongly accretive mapping, {αn} is a sequence in (0,1), and {λn} is a sequence in (0,2α). They proved that the sequence {xn} strongly converges to some point z∈F(T)∩VI(C,A).
Recently, Zegeye and Shahzad [21] investigated the problem of finding a common point of fixed points of a Lipschitz pseudocontractive mapping T and solutions of a variational inequality problem for γ-inverse strongly accretive mapping A by considering the following iterative algorithm:
(10)yn=(1-βn)xn+βnTxn,xn+1=PC[(1-αn)(δnTyn+θnxn+γnPC[I-γA]xn)],
where PC is a metric projection from H onto C and {δn},{θn},{γn},{αn},{βn} are in (0,1) satisfying certain conditions. Then, they proved that the sequence {xn} converges strongly to the minimum-norm point of F(T)∩VI(C,A).
A natural question arises whether we can obtain an iterative scheme which converges strongly to a common point of fixed points of a finite family of pseudocontractive mappings and solutions of a finite family of variational inequality problems for γ-inverse strongly accretive mappings or not.
It is our purpose in this paper to introduce an algorithm and prove that the algorithm converges strongly to a common point of fixed points of a finite family of Lipschitz pseudocontractive mappings and solutions of a finite family of variational inequality problems for γ-inverse strongly accretive mappings. The results obtained in this paper improve and extend the results of Takahashi and Toyoda [19], Iiduka and Takahashi [20], and Zegeye and Shahzad [21], Theorem 3.2 of Yao et al. [22], and some other results in this direction.
2. Preliminaries
In what follows we will make use of the following lemmas.
Lemma 1.
Letting H be a real Hilbert space, the following identity holds:
(11)∥x+y∥2≤∥x∥2+2〈y,x+y〉,∀x,y∈H.
Lemma 2 (see [<xref ref-type="bibr" rid="B4">23</xref>]).
Let C be a nonempty closed and convex subset of a real Hilbert space H. Let A:C→E be a γ-inverse strongly accretive mapping. Then, for 0<μ<2γ, the mapping Aμx:=(x-μAx) is nonexpansive.
Lemma 3 (see [<xref ref-type="bibr" rid="B2">24</xref>]).
Let C be a nonempty, closed, and convex subset of a smooth Banach space E. Let QC be a sunny nonexpansive retraction from E onto C and let A be an accretive operator of C into E. Then for all λ>0,
(12)VI(C,A)=F(QC(I-λA)).
Lemma 4 (see [<xref ref-type="bibr" rid="B8">25</xref>]).
Let C be a nonempty, closed, and convex subset of a real Hilbert space H. Let Ti:C→E, i=1,…,N, be nonexpansive mappings such that ∩i=1NF(Ti)≠∅. Let T:=θ1T1+θ2T2+⋯+θNTN with θ1+θ2+⋯+θN=1. Then T is nonexpansive and F(T)=∩i=1NF(Ti).
Lemma 5 (see [<xref ref-type="bibr" rid="B1">26</xref>]).
Let C be a convex subset of a real Hilbert space H. Let x∈H. Then x0=PCx if and only if
(13)〈z-x0,x-x0〉≤0,∀z∈C.
Lemma 6 (see [<xref ref-type="bibr" rid="B30">27</xref>]).
Let C be a closed convex subset of a real Hilbert space H and A:C→C be a continuous pseudo-contractive mapping. Then, for 0<μ<2γ, the mapping Aμx:=(x-μAx) is nonexpansive
F(T) is a closed convex subset of C;
(I-T) is demiclosed at zero; that is, if {xn} is a sequence in C such that xn⇀x and Txn-xn→0, as n→∞, then x=T(x).
Lemma 7 (see [<xref ref-type="bibr" rid="B22">28</xref>]).
Let H be a real Hilbert space. Then for all xi∈H and αi∈[0,1] for i=1,2,3 such that α1+α2+α3=1 the following equality holds:
(14)∥α1x1+α2x2+α3x3∥2=∑i=13αi∥xi∥2-∑1≤i,j≤3αiαj∥xi-xj∥2.
Lemma 8 (see [<xref ref-type="bibr" rid="B13">29</xref>]).
Let {an} be sequences of real numbers such that there exists a subsequence {ni} of {n} such that ani<ani+1 for all i∈N. Then there exists an increasing sequence {mk}⊂N such that mk→∞ and the following properties are satisfied by all (sufficiently large) numbers k∈N:
(15)amk≤amk+1,ak≤amk+1.
In fact, mk is the largest number n in the set {1,2,…,k} such that the condition an≤an+1 holds.
Lemma 9 (see [<xref ref-type="bibr" rid="B15">30</xref>]).
Let {an} be a sequence of nonnegative real numbers satisfying the following relation:
(16)an+1≤(1-αn)an+αnδn,n≥n0,
where {αn}⊂(0,1) and {δn}⊂R satisfying the following conditions: limn→∞αn=0, ∑n=1∞αn=∞ and limsupn→∞δn≤0. Then, limn→∞an=0.
3. Main Result
For the rest of this paper, let {an},{bn},{cn},⊂(c,1)⊂(0,1), for some c∈(0,1), and {αn}⊂(0,b)⊂(0,1), for some b∈(0,1), satisfy (i) an+bn+cn=1; (ii) limn→∞αn=0; and (iii) ∑αn=∞.
Theorem 10.
Let C be a nonempty, closed, and convex subset of a real Hilbert space H. Let Tj:C→C, j=1,2,…,M, be Lipschitz pseudocontractive mappings with Lipschitz constants Li, respectively. Let Aj:C→H, for j=1,2,…,N, be γj-inverse strongly accretive mappings. Let f:C→C be a contraction with constant α. Assume that F=[∩j=1MF(Tj)]⋂[∩j=1NVI(C,Aj)] is nonempty. Let a sequence {xn} be generated from an arbitrary x0∈C by
(17)yn=(1-λn)xn+λnTnxn;xn+1=αnf(xn)+(1-αn)(anxn+bnTnyn+cnGxn),
where Tn=Tn(modM) and G:=e0I+e1PC[I-γA1]+e2PC[I-γA2]+⋯+eNPC[I-γAr], for γ∈(0,2γ0), for γ0:=min1≤j≤N{γj} with e0+e1+⋯+er=1 and bn+cn≤λn≤λ<1/(1+L2+1),∀n≥0, for L=max{Lj:1≤j≤M}. Then, {xn} converges strongly to a point x*∈F which is the unique solution of the variational inequality 〈(I-f)(x*),x-x*〉≥0 for all x∈F.
Proof.
From Lemmas 2, 4, and 3 we get that G is nonexpansive mapping with F(G)=∩j=1NVI(C,Aj). Let p∈F. Then from (17), (5), and Lemma 7 we have that
(18)∥yn-p∥2=∥(1-λn)(xn-p)+λn(Tnxn-p)∥2=(1-λn)∥xn-p∥2+λn∥Tnxn-p∥2-λn(1-λn)∥xn-Tnxn∥2≤(1-λn)∥xn-p∥2+λn[∥xn-p∥2+∥xn-Tnxn∥2]-λn(1-λn)∥xn-Tnxn∥2=∥xn-p∥2+λn2∥xn-Tnxn∥2,(19)∥xn+1-p∥2=∥αnf(xn)+(1-αn)(anxn+bnTnyn+cnGxn)-p∥2=∥αn(f(xn)-p)+(1-αn)×(an(xn-p)+bn(Tnyn-p)+cn(Gxn-p))∥2≤αn∥f(xn)-p∥2+(1-αn)×∥an(xn-p)+bn(Tnyn-p)+cn(Gxn-p)∥2≤αn∥f(xn)-p∥2+(1-αn)×[an∥xn-p∥2+bn∥Tnyn-p∥2+cn∥Gxn-p∥2]-(1-αn)bnan∥Tnyn-xn∥2≤αn∥f(xn)-p∥2+(1-αn)[(an+cn)∥xn-p∥2hhhhhhhhhhhhhhhhhhhhhhhhhh+bn∥Tnyn-p∥2]-(1-αn)bnan∥Tnyn-xn∥2≤αn∥f(xn)-p∥2+(1-αn)[(an+cn)∥xn-p∥2hhhhhhhhhhhhh+bn(∥yn-p∥2+∥yn-Tnyn∥2)]-(1-αn)bnan∥Tnyn-xn∥2.
Now, substituting (18) in (19) we get that
(20)∥xn+1-p∥2≤αn∥f(xn)-p∥2+(1-αn)[(an+cn)∥xn-p∥2hhhhhhhhhhhh+bn(∥xn-p∥2+λn2∥xn-Tnxn∥2)hhhhhhhhhhhh+bn∥yn-Tnyn∥2]-(1-αn)bnan∥Tnyn-xn∥2=αn∥f(xn)-p∥2+(1-αn)∥xn-p∥2+(1-αn)λn2bn∥xn-Tnxn∥2+(1-αn)bn∥yn-Tnyn∥2-(1-αn)bnan∥Tnyn-xn∥2.
Moreover, from (17), Lemma 7, and Lipschitz property of Tn we get that
(21)∥yn-Tnyn∥2=∥(1-λn)(xn-Tnyn)+λn(Tnxn-Tnyn)∥2=(1-λn)∥xn-Tnyn∥2+λn∥Tnxn-Tnyn∥2-λn(1-λn)∥xn-Tnxn∥2≤(1-λn)∥xn-Tnyn∥2+λnL2∥xn-yn∥2-λn(1-λn)∥xn-Tnxn∥2=(1-λn)∥xn-Tnyn∥2+λn3L2∥xn-Tnxn∥2-λn(1-λn)∥xn-Tnxn∥2=(1-λn)∥xn-Tnyn∥2-λn(1-L2λn2-λn)∥xn-Tnxn∥2.
Substituting (21) into (20) we obtain that
(22)∥xn+1-p∥2≤αn∥f(xn)-p∥2+(1-αn)∥xn-p∥2+(1-αn)bnλn2∥xn-Tnxn∥2+(1-αn)bn[(1-λn)∥xn-Tnyn∥2-λn(1-L2λn2-λn)∥xn-Tnxn∥2]-(1-αn)bnan∥Tnyn-xn∥2,=αn∥f(xn)-p∥2+(1-αn)∥xn-p∥2-(1-αn)λnbn[1-L2λn2-2λn]∥xn-Tnxn∥2+(1-αn)bn[(1-an)-λn]∥Tnyn-xn∥2=αn∥f(xn)-p∥2+(1-αn)∥xn-p∥2-(1-αn)λnbn[1-L2λn2-2λn]∥xn-Tnxn∥2+(1-αn)bn[bn+cn-λn]∥Tnyn-xn∥2.
But, from the hypothesis we have that
(23)1-2λn-L2λn2≥1-2λ-L2λ2>0,bn+cn≤λn,∀n≥0,
and hence inequality (22) gives that
(24)∥xn+1-p∥2≤αn∥f(xn)-p∥2+(1-αn)∥xn-p∥2.
But we have that
(25)∥f(xn)-p∥2=[∥f(xn)-f(p)∥+∥f(p)-p∥]2≤[α∥xn-p∥+∥f(p)-p∥]2≤α2∥xn-p∥2+∥f(p)-p∥2+2α∥xn-p∥∥f(p)-p∥≤α(1+α)∥xn-p∥2+(1+α)∥f(p)-p∥2.
Substituting (25) into (24) we get that
(26)∥xn+1-p∥2≤(1-αn(1-α(1+α)))∥xn-p∥2+αn(1+α)∥f(p)-p∥2.
Therefore, by induction we get that
(27)∥xn+1-p∥2≤max{∥x0-p∥2,1+α1-α(1+α)∥f(p)-p∥2},hhhhhhhhhhhhhhhhhhhhhhhhhhhhhh∀n≥0,
which implies that {xn} and hence {yn} are bounded.
Let x*=PFf(x*). Then, from (17), Lemmas 1 and 7, and the methods used to get (22) we obtain that
(28)∥xn+1-x*∥2=∥αn(f(xn)-x*)+(1-αn)[anxn+bnTnyn+cnGxn-x*]∥2≤(1-αn)∥an(xn-x*)+bn(Tnyn-x*)hhhhhhhhhhh+cn(Gxn-x*)∥2+2αn〈f(xn)-x*,xn+1-x*〉≤(1-αn)bn∥Tnyn-x*∥2+(1-αn)an∥xn-x*∥2×(1-αn)cn∥Gxn-p∥2-(1-αn)anbn∥Tnyn-xn∥2-(1-αn)ancn∥Gxn-xn∥2+2αn〈f(xn)-x*,xn+1-x*〉,∥xn+1-x*∥2≤(1-αn)bn[∥yn-x*∥2+∥yn-Tnyn∥2]+(1-αn)(an+cn)∥xn-x*∥2-(1-αn)anbn∥Tnyn-xn∥2-(1-αn)ancn∥Gxn-xn∥2+2αn〈f(xn)-x*,xn+1-x*〉,≤(1-αn)bn[∥xn-x*∥2+λn2∥xn-Tnxn∥2]+(1-αn)bn[(1-λn)∥xn-Tnyn∥2hhhhhhhhhhhhhhh-λn(1-L2λn2-λn)hhhhhhhhhhhhhhh×∥xn-Tnxn∥2]+(1-αn)(an+cn)∥xn-x*∥2-(1-αn)anbn∥Tnyn-xn∥2-(1-αn)ancn∥Gxn-xn∥2+2αn〈f(xn)-x*,xn+1-x*〉
which implies that
(29)∥xn+1-x*∥2≤(1-αn)∥xn-x*∥2-(1-αn)bnλn[1-L2λn2-2λn]×∥xn-Tnxn∥2+(1-αn)bn(bn+cn-λn)∥xn-Tnyn∥2-(1-αn)andn∥Gxn-xn∥2+2αn〈f(xn)-x*,xn+1-x*〉(30)≤(1-αn)∥xn-x*∥2+2αn〈f(xn)-x*,xn+1-x*〉.
But
(31)〈f(xn)-x*,xn+1-x*〉=〈f(xn)-x*,xn-x*〉+〈f(xn)-x*,xn+1-xn〉≤〈f(xn)-f(x*),xn-x*〉+〈f(x*)-x*,xn-x*〉+∥xn+1-xn∥∥f(xn)-x*∥≤α∥xn-x*∥2+〈f(x*)-x*,xn-x*〉+∥xn+1-xn∥∥f(xn)-x*∥.
Thus, substituting (31) in (30) we obtain that
(32)∥xn+1-x*∥2≤(1-αn(1-2α))∥xn-x*∥2+2αn〈f(x*)-x*,xn-x*〉+2αn∥xn+1-xn∥·∥f(xn)-x*∥.
Next, we consider two cases.
Case 1. Suppose that there exists n0∈N such that {∥xn-x*∥} is decreasing for all n≥n0. Then, we get that {∥xn-x*∥} is convergent. Thus, from (29) and (23) we have that
(33)xn-Tnxn⟶0,Gxn-xn⟶0asn⟶∞.
Furthermore, from (17) and (33) we obtain that
(34)∥yn-xn∥=λn∥xn-Tnxn∥⟶0asn⟶∞,
and hence Lipschitz continuity of Tn, (34), and (33) implies that
(35)∥Tnyn-xn∥≤∥Tnyn-Tnxn∥+∥Tnxn-xn∥≤L∥yn-xn∥+∥Tnxn-xn∥⟶0hhhhhhhhhhhhhhhhhhhasn⟶∞.
Thus, from (33) and (35) we have that
(36)∥xn+1-xn∥=∥αn(f(xn)-xn)+(1-αn)×(anxn+bnTnyn+cnGxn)-xn∥≤αn∥f(xn)-xn∥+(1-αn)bn∥Tnyn-xn∥+(1-αn)cn∥Gxn-xn∥⟶0asn⟶∞.
Therefore, ∥xn+j-xn∥→0, as n→∞, for all j=1,2,…,M, and hence
(37)∥xn-Tn+jxn∥≤∥xn-xn+j∥+∥xn+j-Tn+jxn+j∥+L∥xn+j-xn∥⟶0,
as n→∞, for all j∈{1,2,…,M}.
Now, since {xn} is bounded subset of H, we can choose a subsequence {xnm} of {xn} such that xnm⇀x and limsupn→∞〈f(x*)-x*,xn-x*〉=limm→∞〈f(x*)-x*,xnm-x*〉. Then, from (37) and Lemma 6 we have that x∈F(Tj), for each j=1,2,…,M. Hence, x∈∩j=1MF(Tj).
In addition, since G is nonexpansive, from Lemma 6 we get that x∈F(G) and hence by Lemmas 4 and 3 we obtain that x∈VI(C,Aj), for each j∈{1,2,…,N}.
Therefore, by Lemma 5, we immediately obtain that
(38)limsupn→∞〈f(x*)-x*,xn-x*〉=limm→∞〈f(x*)-x*,xnm-x*〉=〈f(x*)-x*,x-x*〉≤0.
Then, it follows from (32), (38), and Lemma 9 that ∥xn-x*∥→0 as n→∞. Consequently, xn→x*=PF(f(x*)).
Case 2. Suppose that there exists a subsequence {ni} of {n} such that
(39)∥xni-x*∥<∥xni+1-x*∥,
for all i∈N. Then, by Lemma 8, there exists a nondecreasing sequence {mk}⊂N such that mk→∞, and
(40)∥xmk-x*∥≤∥xmk+1-x*∥,∥xk-x*∥≤∥xmk+1-x*∥,
for all k∈N. Now, from (29) and (23) we get that xmk-Tmkxmk→0 and Gxnk-xnk→0 as k→∞. Thus, following the method in Case 1, we obtain that xmk+1-xmk→0, xmk-Tjxmk→0, and
(41)limsupk→∞〈f(x*)-x*,xmk-x*〉≤0.
Furthermore, from (32) and (40) we obtain that
(42)αmk(1-2α)∥xmk-x*∥2≤∥xmk-x*∥2-∥xmk+1-x*∥2+2αmk〈f(x*)-x*,xmk-x*〉+2αmk∥xmk+1-xmk∥∥f(xmk)-x*∥≤2αmk〈f(x*)-x*,xmk-x*〉+2αmk∥xmk+1-xmk∥∥f(xmk)-x*∥.
Now, using the fact that αmk>0 and (41) we get that
(43)∥xmk-x*∥2⟶0ask⟶∞,
and this together with (32) implies that ∥xmk+1-x*∥→0 as k→∞. Since ∥xk-x*∥≤∥xmk+1-x*∥ for all k∈N, we obtain that xk→x*. Hence, from the above two cases, we can conclude that {xn} converges strongly to a point x*=PFf(x*), which satisfies the variational inequality 〈(I-f)(x*),x-x*〉≥0, for all x∈F. The proof is complete.
If, in Theorem 10, we assume that f(x)=u∈C, a constant mapping, then we get the following corollary.
Corollary 11.
Let C be a nonempty, closed, and convex subset of a real Hilbert space H. Let Tj:C→C, j=1,2,…,M, be Lipschitz pseudocontractive mappings with Lipschitz constants Lj, respectively. Let Aj:C→H, for j=1,2,…,N, be γj-inverse strongly accretive mappings. Assume that F=[∩j=1MF(Tj)]⋂[∩j=1NVI(C,Aj)] is nonempty. Let a sequence {xn} be generated from an arbitrary x0,u∈C by
(44)yn=(1-λn)xn+λnTnxn;xn+1=αnu+(1-αn)(anxn+bnTnyn+cnGxn),
where Tn=Tn(modM), G:=e0I+e1PC[I-γA1]+e2PC[I-γA2]+⋯+eNPC[I-γAr], for γ∈(0,2γ0), for γ0:=min1≤j≤N{γj} with e0+e1+⋯+er=1, and bn+cn≤λn≤λ<1/(1+L2+1),∀n≥0, for L=max{Lj:1≤j≤M}. Then, {xn} converges strongly to a unique point x*∈C satisfying x*=PF(u), which is the unique solution of the variational inequality 〈x*-u,x-x*〉≥0 for all x∈F.
If, in Theorem 10, we assume that N=1 and M=1, then we get the following corollary which is Theorem 3.1 of [21].
Corollary 12.
Let C be a nonempty, closed, and convex subset of a real Hilbert space H. Let T:C→C be Lipschitz pseudocontractive mappings with Lipschitz constant L and A:C→H an γ-inverse strongly accretive mapping. Let f:C→C be a contraction with constant α. Assume that F=F(T)⋂VI(C,A) is nonempty. Let a sequence {xn} be generated from an arbitrary x0∈C by
(45)yn=(1-λn)xn+λnTxn;xn+1=αnf(xn)+(1-αn)(anxn+bnTyn+cnPC[I-rA]xn),
where r∈(0,2γ) and bn+cn≤λn≤λ<1/(1+L2+1),∀n≥0. Then, {xn} converges strongly to a point x*∈F, which is the unique solution of the variational inequality 〈(I-f)(x*),x-x*〉≥0 for all x∈F.
If, in Theorem 10, we assume that Ti′s are strictly pseudocontractive mappins, then we get the following corollary.
Corollary 13.
Let C be a nonempty, closed, and convex subset of a real Hilbert space H. Let Ti:C→C, i=1,2,…M, be λi-strictly pseudocontractive mappings and let Ai:C→H, for i=1,2,…,N, be an γi-inverse strongly accretive mappings. Let f:C→C be a contraction with constant α. Assume that F=[∩i=1MF(Ti)]⋂[∩i=1NVI(C,Ai)] is nonempty. Let a sequence {xn} be generated from an arbitrary x0∈C by
(46)yn=(1-λn)xn+λnTnxn;xn+1=αnf(xn)+(1-αn)(anxn+bnTnyn+cnGxn),
where Tn=Tn(modM), G:=c0I+e1PC[I-γA1]+e2PC[I-γA2]+⋯+eNPC[I-γAr], for γ∈(0,2γ0), for γ0:=min1≤i≤N{γi} with e0+e1+⋯+er=1, and bn+cn≤λn≤λ<1/(1+L2+1),∀n≥0, L=max{(1+λi)/λi}. Then, {xn} converges strongly to a point x*∈F, which is the unique solution of the variational inequality 〈(I-f)(x*),x-x*〉≥0 for all x∈F.
If, in Theorem 10, we assume that Ti′s are nonexpansive mapping, then we get the following corollary.
Corollary 14.
Let C be a nonempty, closed, and convex subset of a real Hilbert space H. Let Ti:C→C, i=1,2,…,M, be nonexpansive mappings and let Ai:C→H, for i=1,2,…,N, be an γi-inverse strongly accretive mappings. Let f:C→C be a contraction with constant α. Assume that F=[∩i=1MF(Ti)]⋂[∩i=1NVI(C,Ai)] is nonempty. Let a sequence {xn} be generated from an arbitrary x0∈C by
(47)yn=(1-λn)xn+λnTnxn;xn+1=αnf(xn)+(1-αn)(anxn+bnTnyn+cnGxn),
where Tn=Tn(modM), G=c0I+e1PC[I-γA1]+e2PC[I-γA2]+⋯+eNPC[I-γAr], for γ∈(0,2γ0), for γ0:=min1≤i≤N{γi} with e0+e1+⋯+er=1, an+bn+cn=1, and bn+cn≤λn≤λ<1/(2+1),∀n≥0. Then, {xn} converges strongly to point x*∈F, which is the unique solution of the variational inequality 〈(I-f)(x*),x-x*〉≥0 for all x∈F.
We note that the method of proof of Theorem 10 provides the following theorem which is a convergence theorem for a minimum norm point of common fixed points of a finite family of Lipschitz pseudocontractive mappings and common solutions of a finite family of variational inequality problems for accretive mappings.
Theorem 15.
Let C be a nonempty, closed, and convex subset of a real Hilbert space H. Let Tj:C→C, j=1,2,…,M, be Lipschitz pseudocontractive mappings with Lipschitz constants Lj, respectively. Let Aj:C→H, for j=1,2,…,N, be γj-inverse strongly accretive mappings. Assume that F=[∩j=1MF(Tj)]⋂[∩j=1NVI(C,Aj)] is nonempty. Let a sequence {xn} be generated from an arbitrary x0∈C by
(48)yn=(1-λn)xn+λnTnxn;xn+1=PC[(1-αn)(anxn+bnTnyn+cnGxn)],
where Tn=Tn(modM), G:=e0I+e1PC[I-γA1]+e2PC[I-γA2]+⋯+eNPC[I-γAr], for γ∈(0,2γ0), for γ0:=min1≤j≤N{γj} with e0+e1+⋯+er=1, and bn+cn≤λn≤λ<1/(1+L2+1),∀n≥0, for L=max{Lj:1≤j≤M}. Then, {xn} converges strongly to a unique minimum norm point x* of F(i.e., x*=PF(0)), which is the unique solution of the variational inequality 〈x*,x-x*〉≥0 for all x∈F.
4. Numerical Example
Now, we give an example of two Lipschitz pseudocontractive mappings and two γ-inverse strongly accretive mappings satisfying Theorem 10 and some numerical experiment result to explain the conclusion of the theorem as follows.
Example 1.
Let H=R with absolute value norm. Let C=[-2,2] and let T1,T2:C→C be defined by
(49)T1x:={x+x2,x∈[-2,0],x,x∈(0,2],T2x:={x,x∈[-2,12],x-(169)(x-12)2,x∈(12,2].
Clearly, for x,y∈C we have that
(50)〈(I-T1)x-(I-T1)y,x-y〉≥0,〈(I-T2)x-(I-T2)y,x-y〉≥0
which show that both mappings are pseudocontractive. Next, we show that T1 is Lipschitz with L=5. If x,y∈[-2,0], then
(51)|T1x-T1y|=|x+x2-y-y2|=|(x+y)+1||x-y|≤3|x-y|.
If x,y∈(0,2], then
(52)|T1x-T1y|=|x-y|.
If x∈[-2,0] and y∈(0,2], then
(53)|T1x-T1y|=|x+x2-y|=|x-y+x2|=|x-y+x2-y2+y2|=|x-y+x2-y2|+y2≤|x+y+1|·|x-y|+|y-x|2=(|x+y+1|+|x+y|)·|x-y|≤5|x-y|.
Thus, we get that T1 is Lipschitz pseudocontractive with L=5 and F(T1)=[0,2] which is not nonexpansive, since if we take x=-2 and y=-1.9, we have that |T1x-T2y|=0.29>0.1=|x-y|. Similarly, we can show that T2 is Lipschitz pseudocontractive with L=4 and F(T2)=[-2,1/2] which is not nonexpansive.
Furthermore, for C=[-2,2], let A1,A2:C→R be defined by
(54)A1x:={-(x-12)2,x∈[-2,12),0,x∈[12,2],A2x:={0,x∈[-2,23],3(x-23)2,x∈(23,2].
Then we first show that A1 is γ-inverse strongly accretive mapping with γ=1/5.
If x,y∈[-2,1/2), then
(55)〈A1x-A1y,x-y〉=〈-(x-12)2+(y-12)2,x-y〉=[(x-12)2-(y-12)2](y-x)=[(x-12)2-(y-12)2][(y-12)-(x-12)]=[(x-12)2-(y-12)2][(y-1/2)2-(x-1/2)2](y-1/2)+(x-1/2)=[(x-12)2-(y-12)2][(x-1/2)2-(y-1/2)2](1/2-x)+(1/2-y)≥15|(x-12)2-(y-12)2|2=15|A1x-A1y|2.
If x∈[-2,1/2) and y∈[1/2,2], we get that
(56)〈A1x-A1y,x-y〉=〈-(x-12)2,x-y〉=(x-12)2(y-x)=(x-12)2[(y-12)-(x-12)]≥(x-12)2(12-x)=(x-12)2(1/2-x)2(1/2-x)≥25|(x-12)2|2≥15|A1x-A1y|2.
If x,y∈[1/2,2], then we get that |A1x-A1y|=0 and hence
(57)〈A1x-A1y,x-y〉≥15|A1x-A1y|2.
Therefore, A1 is γ-inverse strongly accretive mapping with γ=1/5 and VI(C,A1)=[1/2,2]. Similarly, we can show that A2 is γ-inverse strongly accretive mapping with γ=1/2 and VI(C,A2)=[-2,2/3].
Note that we have F(T1)∩F(T2)∩VI(C,A1)∩VI(C,A2)={1/2}.
Thus, taking αn=1/(10n+100), λn=2/(n+100)+0.065, bn=cn=1/(n+100)+0.01, an=1-2/(n+100)-0.02, and f(x)=u∈C, we observe that conditions of Theorem 10 are satisfied and Scheme (17) provides the following Table 1 and Figures 1(a) and 1(b) for u=0.6 and u=0.8, respectively.
We observe that the data provides strong convergence of the sequence to the common point of fixed points of both pseudocontractive mappings and solutions of both variational inequality problems for γ-inverse strongly accretive mappings.
u=0.6
x0=1
u=0.8
x0=-1
n
xn
n
xn
0
1.0000
0
−1.0000
500
0.6112
5000
0.0627
10,000
0.5137
10,000
0.4282
12,000
0.5121
15,000
0.4540
14,000
0.5110
20,000
0.4686
18,000
0.5093
25,000
0.4782
20,000
0.5087
35,000
0.4905
Remark 2.
Theorem 10 provides an iteration scheme which converges strongly to a common point of fixed points of a finite family of Lipschitzian pseudocontractive mappings and solutions of a finite family of variational inequality problems in Hilbert spaces.
Remark 3.
Theorem 10 improves Theorem 3.1 of Takahashi and Toyoda [19], Iiduka and Takahashi [20], and Zegeye and Shahzad [21] and Theorem 3.2 of Yao et al. [22] in the sense that our convergence is to a common point of fixed points of a finite family of Lipschitzian pseudocontractive mappings and solutions of a finite family of variational inequality problems.
Conflict of Interests
The authors declare that they have no conflict of interests regarding the publication of this paper.
Acknowledgments
This project was funded by the Deanship of Scientific Research (DSR), King Abdulaziz University, Jeddah, under grant no. 167/130/1434. The authors, therefore, acknowledge with thanks DSR technical and financial support. The authors also thank the referees for their valuable comments and suggestions, which improved the presentation of this paper.
KinderlehrerD.StampacciaG.ZegeyeH.ShahzadN.A hybrid scheme for finite families of equilibrium, variational inequality and fixed point problemsZegeyeH.ShahzadN.Approximating common solution of variational inequality problems for two monotone mappings in Banach spacesZegeyeH.ShahzadN.Strong convergence theorems for variational inequality problems and quasi-ϕ-asymptotically nonexpansive mappingsZegeyeH.ShahzadN.An iteration to a common point of solution of variational inequality and fixed point-problems in Banach spacesBrowderF. E.PetryshynW. V.Construction of fixed points of nonlinear mappings in Hilbert spaceCengL. C.PetruşelA.YaoJ. C.Composite viscosity approximation methods for equilibrium problem, variational inequality and common fixed pointsCengL. C.PetruselA.WongM. M.YaoJ. C.Hybrid algorithms for solving variational inequalities, variational inclusions, mixed equilibria, and fixed point problemsCengL.PetruselA.WongM.Strong convergence theorem for a generalized equilibrium problem and a pseudocontractive mapping in a Hilbert spaceSahuD. R.ColaoV.MarinoG.Strong convergence theorems for approximating common fixed points of families of nonexpansive mappings and applicationsSahuD. R.PetruşelA.Strong convergence of iterative methods by strictly pseudocontractive mappings in Banach spacesYaoY.LiouY.ShahzadN.Construction of iterative methods for variational inequality and fixed point problemsYaoY.MarinoG.MugliaL.A modified Korpelevich's method convergent to the minimum-norm solution of a variational inequalityYaoY.PostolacheM.Iterative methods for pseudomonotone variational inequalities and fixed-point problemsZegeyeH.OfoeduE. U.ShahzadN.Convergence theorems for equilibrium problem, variational inequality problem and countably infinite relatively quasi-nonexpansive mappingsZegeyeH.ShahzadN.A hybrid approximation method for equilibrium, variational inequality and fixed point problemsZegeyeH.ShahzadN.Strong convergence theorems for a solution of finite families of equilibrium and variational inequality problemsZegeyeH.ShahzadN.Extragradient method for solutions of variational inequality problems in Banach spacesTakahashiW.ToyodaM.Weak convergence theorems for nonexpansive mappings and monotone mappingsIidukaH.TakahashiW.Strong convergence theorems for nonexpansive mappings and inverse-strongly monotone mappingsZegeyeH.ShahzadN.Solutions of variational inequality problems in the set of fixed points of pseudocontractive mappingsYaoY.LiouY. C.KangS. M.Algorithms construction for variational inequalitiesCaiY.TangY.LiuL.Iterative algorithms for minimum-norm fixed point of non-expansive mapping in Hilbert spaceAoyamaK.IidukaH.TakahashiW.Weak convergence of an iterative sequence for accretive operators in Banach spacesChidumeC. E.ZegeyeH.PrempehE.Strong convergence theorems for a finite family of nonexpansive mappings in Banach spacesAlberY. I.Metric and generalized projection operators in Banach spaces: properties and applicationsZhangQ. B.ChengC. Z.Strong convergence theorem for a family of Lipschitz pseudocontractive mappings in a Hilbert spaceZegeyeH.ShahzadN.Convergence of Mann's type iteration method for generalized asymptotically nonexpansive mappingsMaingéP. E.Strong convergence of projected subgradient methods for nonsmooth and nonstrictly convex minimizationXuH. K.Another control condition in an iterative method for nonexpansive mappings