Abstract
In this paper we introduce an iterative method for finding a common fixed point of an infinite family of nonexpansive mappings in q-uniformly real smooth Banach space which is also uniformly convex. We proved strong convergence of the proposed iterative algorithms to the unique solution of a variational inequality problem.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Let E be a real Banach space and \(E^*\) be the dual space of E. A mapping \(\varphi :[0,\infty )\rightarrow [0,\infty )\) is called a guage function if it is strictly increasing, continuous and \(\varphi (0)=0\). Let \(\varphi \) be a gauge function, a generalized duality mapping with respect to \(\varphi ,J_{\varphi }:E\rightarrow 2^{E^*}\) is defined by, \(x\in E,\)
where \(\langle .,.\rangle \) denotes the duality pairing between element of E and that of \(E^*\). If \(\varphi (t)=t\), then \(J_\varphi \) is simply called the normalized duality mapping and is denoted by J. For any \(x\in E\), an element of \(J_\varphi x\) is denoted by \(j_{\varphi }x\).
If however \(\varphi (t)=t^{q-1}\), for some \(q>1\), then \(J_\varphi \) is still called the generalized duality mapping and is denoted by \(J_{q}\) (see, for example [1, 5]).
Let \(S(E):= \{x \in E:\Vert x\Vert = 1\}\) be the unit sphere of E. Then space E is said to have Gâteaux differentiable norm if for any \(x \in S(E)\) the limit
exists \(\forall y \in S(E)\). The norm of E is said to be uniformly G\(\hat{a}\)teaux differentiable if for each \(y\in S(E),\) the limit (1.1) is attained uniformly for \(x \in S(E).\) If E has a uniformly G\(\hat{a}\)teaux differentiable norm, then \(j:E\rightarrow E^*\) is uniformly continuous on bounded subsets of E to the weak\(^*\) topology of \(E^*\).
A mapping \(G:D(G)\subset E\rightarrow E\) is said to be accretive if for all \(x,y\in D(G)\), there exists \(j_q(x-y)\in J_q(x-y)\) such that
where D(G) denote the domain of G. G is called \(\eta -strongly~accretive\) if for all \(x,y\in D(G)\), there exists \(j_q(x-y)\in J_q(x-y)\) and \(\eta \in (0,1)\) such that
Let K be a nonempty, closed and convex subset of E and \(G:K\rightarrow E\) be a nonlinear mapping. The variational inequality problem is to:
for some \(j_q(v-u)\in J_q(v-u).\) The set of solution of variational inequality problem is denoted by VI(K, G). If \(E:=H\), a real Hilbert space, the variational inequality problem reduces to:
which was introduced and studied by Stampacchia [16].
Variational inequality theory has emerged as an important tool in studying a wide class of related problems in Mathematical, Physical, regional, engineering and nonlinear optimization sciences (see, for instance, [8, 9, 11, 15, 24–26]).
A mapping \(T:E\rightarrow E\) is \(L-Lipschitian\) if for some \(L>0,||Tx-Ty||\le L||x-y||~\forall ~x,y\in E\). If \(L\in [0,1)\), then T is called contraction mapping, but if \(L\le 1\), then T is called nonexpansive mapping. A point \(x\in E\) is called a fixed point of T if \(Tx=x\). The set of fixed points of T is denoted by \(F(T):=\{x\in E:Tx=x\}\). In Hilbert spaces H, accretive operators are called monotone where inequality (1.2) and (1.3) hold with \(j_q\) replaced by the identity map on H.
In 2000, Moudafi [14] introduced the viscosity approximation method for nonexpansive mappings. Let f be a contraction on H, starting with an arbitrary \(x_0\in H,\) define a sequence \(\{x_n\}\) recursively by
where \(\{\alpha _n\}\) is a sequence in (0,1). Xu [21] proved that under certain appropriate conditions on \(\{\alpha _n\}\), the sequence \(\{x_n\}\) generated by (1.4) strongly converges to the unique solution \(x^*\) in F of the variational inequality
In [19], he proved, under some conditions on the real sequence \(\{\alpha _n\}\), that the sequence \(\{x_n\}\) defined by \(x_0\in H\) chosen arbitrary,
converges strongly to \(x^*\in F\) which is the unique solution of the minimization problem
where A is a strongly positive bounded linear operator. That is, there is a constant \(\bar{\gamma }>0\) with the property
Combining the iterative method (1.4) and (1.5), Marino and Xu [13] consider the following general iterative method:
they proved that if the sequence \(\{\alpha _n\}\) of parameters satisfies appropriate conditions, then the sequence \(\{x_n\}\) generated by (1.6) converges strongly to \(x^*\in F\) which solves the variational inequality
which is the optimality condition for the minimization problem
where h is a potential function for \(\gamma f\) (i.e. \(h'(x)=\gamma f(x)\) for \(x\in H\)).
On the other hand, Yamada [24] in 2001 introduced the following hybrid iterative method:
where G is a \(\kappa \)-Lipschitzian and \(\eta \)-strongly monotone operator with \(\kappa>0,\eta >0\) and \(0<\mu <2\eta /\kappa ^2\). Under some appropriate conditions, he proved that the sequence \(\{x_n\}\) generated by (1.7) converges strongly to the unique solution of the variational inequality
Recently, combining (1.6) and (1.7), Tian [18] considered the following general iterative method:
and proved that the sequence \(\{x_n\}\) generated by (1.8) converges strongly to the unique solution \(x^*\in F\) of the variational inequality
Most recently, Ali et al [4], extended the result of Tian [18] to q-uniformly smooth Banach space whose duality mapping is weakly sequentially continuous. Under some assumptions on \(\{\alpha _n\},\gamma ,\mu \) and G being \(\eta \)-accretive mapping in (1.8), they proved that the sequence \(\{x_n\}\) generated by (1.8) converges strongly to the unique solution \(x^*\in F\) of the variational inequality
Let \(\{T_i\}\) be countable family of nonexpansive mapping. We denote by a set \(\textit{N}_{I}:=\{i\in \mathbb {N}:T_i\ne I\}\) (I being the identity mapping on E). Maingé [12] studied the Halpern-type scheme for approximation of a common fixed point of countable infinite family of nonexpansive mappings in a real Hilbert space. He proved the following theorems.
Theorem 1.1
(Maingé [12]) Let K be a nonempty closed convex subset of a real Hilbert space H. Let \(\{T_i\}\) be countable family of nonexpansive self-mappings of K, \(\{t_n\}\) and \(\{\sigma _{i,t_n}\}\) be sequences in (0,1) satisfying the following conditions: (i) \(\lim t_n=0,\) (ii) \(\sum _{i\ge 1}\sigma _{i,t_n}=1-t_n\), (iii) \(\forall i\in \textit{N}_{I},\) \(\underset{n\rightarrow \infty }{\lim }\frac{t_n}{\sigma _{i,t_n}}=0.\) Define a fixed point sequence \(\{x_{t_n}\}\) by
where \(C:K\rightarrow K\) is a strict contraction. Assume \(F:=\cap ^{\infty }_{i=1}F(T_i)\ne \emptyset \), the \(\{x_{t_n}\}\) converges strongly to a unique fixed point of the contraction \(P_{F}\circ C\), where \(P_{F}\) is a metric projection from H onto F.
Theorem 1.2
(Maingé [12]) Let K be a nonempty closed convex subset of a real Hilbert space H. Let \(\{T_i\}\) be countable family of nonexpansive self-mappings of K, \(\{\alpha _n\}\) and \(\{\sigma _{i,n}\}\) be sequences in (0,1) satisfying the following conditions:
-
(i)
\(\sum \alpha _{n}=\infty ,\) \(\sum _{i\ge 1}\sigma _{i,n}=1-\alpha _n,\)
-
(ii)
$$\begin{aligned} \left\{ \begin{array}{ll} \frac{1}{\sigma _{i,n}}\Big \vert 1-\frac{\alpha _{n-1}}{\alpha _n}\Big \vert \rightarrow 0,~~\text {or}~~\sum _{n}\frac{1}{\sigma _{i,n}}|\alpha _{n-1}-\alpha _{n}|<\infty &{} \text { }\\ \frac{1}{\alpha _n}\Big \vert \frac{1}{\sigma _{i,n}}-\frac{1}{\sigma _{i,n-1}}\Big \vert \rightarrow 0,~~\text {or}~~\sum _n\Big \vert \frac{1}{\sigma _{i,n}}-\frac{1}{\sigma _{i,n-1}}\Big \vert<\infty &{} \text { }\\ \frac{1}{\sigma _{i,n}\alpha _n}\sum _{k\ge 0}|\sigma _{k,n}-\sigma _{k,n-1}|\rightarrow 0,~~\text {or}~~ \frac{1}{\sigma _{i,n}}\sum _{k\ge 0}|\sigma _{k,n}-\sigma _{k,n-1}|<\infty . \end{array} \right. \end{aligned}$$
-
(iii)
\(\forall i\in \textit{N}_{I},\) \(\underset{n\rightarrow \infty }{\lim }\frac{\alpha _n}{\sigma _{i,n}}=0.\)
Then, the sequence \(\{x_n\}\) define iteratively by \(x_1\in K\),
where \(C:K\rightarrow K\) is a strict contraction. Assume \(F:=\cap ^{\infty }_{i=1}F(T_i)\ne \emptyset \), the \(\{x_{n}\}\) converges strongly to a unique fixed point of the contraction \(P_{F}\circ C\), where \(P_{F}\) is a metric projection from H onto F.
Motivated by the results above, we introduce an iterative method for finding a common fixed point of an infinite family of nonexpansive mappings in q-uniformly real smooth Banach space. We prove the strong convergence of the proposed iterative algorithm to the unique solution of a variational inequality problem.
2 Preliminaries
Let K be a nonempty, closed, convex and bounded subset of a Banach space E and let the diameter of K be defined by \(d(K):= \sup \{\Vert x-y\Vert :x , y \in K \}\). For each \(x \in K\), let \(r(x,K):=\sup \{\Vert x-y\Vert : y \in K \}\) and let \(r(K):=\inf \{r(x,K): x \in K \}\) denote the Chebyshev radius of K relative to itself. The normal structure coefficient N(E) of E (introduced in 1980 by Bynum [3], see also Lim [10] and the references contained therein) is defined by N(E):=\(\inf \){\(\frac{d(K)}{r(K)} \): K is a closed convex and bounded subset of E with \(d(K)>0\)}. A Banach space E such that \(N(E)>1\) is said to have uniform normal structure . It is known that every Banach space with a uniform normal structure is reflexive, and that all uniformly convex and uniformly smooth Banach spaces have uniform normal structure (see e.g., [5, 23]).
Let E be a normed space with dimE \(\ge 2\). The modulus of smoothness of E is the function \(\rho _{E}:[0,\infty )\rightarrow [0,\infty )\) defined by
The space E is called uniformly smooth if and only if \(\lim _{t\rightarrow 0^{+}}\frac{\rho _{E}(t)}{t}=0.\) For some positive constant \(q\in E\) is called \(q-uniformly~smooth\) if there exists a constant \(c>0\) such that \(\rho _{E}(t)\le ct^{q},t>0.\) It is well known that if E is smooth then the duality mapping is singled-valued, and if E is uniformly smooth then the duality mapping is norm-to-norm uniformly continuous on bounded subset of E.
Lemma 2.1
Let E be a real normed space. Then
for all \(x,y\in E\) and for all \(j(x+y)\in J(x+y).\)
Lemma 2.2
(Xu, [22]) Let E be a real q-uniformly smooth Banach space for some \(q>1\), then there exists some positive constant \(d_{q}\) such that
Lemma 2.3
(Xu, [21]) Let \(\{a_{n}\}\) be a sequence of nonegative real numbers satisfying the following relation:
where, (i) \(\{\alpha _{n}\}\subset [0,1], ~\sum \alpha _{n}=\infty ;\) (ii) \(\limsup \sigma _{n}\le 0;\) (iii) \(\gamma _{n}\ge 0;~(n\ge 0),~\sum \gamma _{n}<\infty .\) Then, \(a_{n}\rightarrow 0\) as \(n\rightarrow \infty .\)
Lemma 2.4
(Suzuki [17]) Let \(\{x_{n}\}\) and \(\{y_{n}\}\) be bounded sequence in a Banach space E and let \(\{\beta _{n}\}\) be a sequence in [0, 1] with \(0<\liminf \beta _{n}\le \limsup \beta _{n}<1.\) Suppose that \(x_{n+1}=\beta _{n}y_{n}+(1-\beta _{n})x_{n}\) for all integer \(n\ge 1\) and \({\limsup }_{n\rightarrow \infty }(||y_{n+1}-y_{n}||-||x_{n+1}-x_{n}||)\le 0.\) Then, \(\underset{n\rightarrow \infty }{\lim }||y_{n}-x_{n}||=0.\)
Lemma 2.5
(See Lemma 2.1 of Ali [2]) Let E be a real smooth and uniformly convex Banach space and let \(r>0\). Then there exists a strictly increasing, continuous and convex function \(g:[0,2r]\rightarrow R\) such that \(g(0)=0\) and \(g(||x-y)\le ||x||^2-2\langle x,jy\rangle +||y||^2\) for all \(B_{r}=\{x\in E:||x||\le r\}.\)
Lemma 2.6
Let E be a real Banach space, \(f:E\rightarrow E\) be contraction mapping with a coefficient \(0<\beta<\) and let \(G:E\rightarrow E\) be a \(\kappa -\)Lipschitzian and \(\eta -\)strongly accretive operator with \(\kappa >0,\eta \in (0,1)\). Then for \(\gamma \in (0,\frac{\mu \eta }{\beta })\),
That is, \((\mu G-\gamma f)\) is strongly accretive with coefficient \((\mu \eta -\gamma \beta )\).
Let \(\mu \) be a linear continuous functional on \(l^{\infty }\) and let \(a=(a_1,a_2,\cdots )\in l^{\infty }.\) We will sometimes write \(\mu _{n}(a_n)\) in place of the value \(\mu (a)\). A linear continuous functional \(\mu \) such that \(||\mu ||=1=\mu (1)\) and \(\mu _{n}(a_n)=\mu _{n}(a_{n+1})\) for every \(a=(a_1,a_2,\cdots )\in l^{\infty }\) is called a Banach limit. It is known that if \(\mu \) is a Banach limit, then
for every \(a=(a_1,a_2,\ldots )\in l^{\infty }\) (see, for example, [5, 6])
3 Main results
In the sequel we assume for each \(\alpha \in (0,1)\), the sequence \(\{\sigma _{i,\alpha }\}\) satisfies \(\sum _{i\ge 1}\sigma _{i,\alpha }=1-\alpha \) and for the sequence \(\{\alpha _{n}\}\subset (0,1),\{\sigma _{i,n}\}=1-\alpha _{n}\).
Lemma 3.1
Let E be a \(q-\)uniformly smooth real Banach space with constant \(d_q, q>1\). Let \(f:E\rightarrow E\) be a \(\beta -\)contraction mapping with a coefficient \(\beta \in (0,1)\). Let \(T_{i}:E\rightarrow E~i\in N,\) be a family of nonexpansive maps such that \(F:=\cap ^{\infty }_{i=1}F(T_i)\ne \emptyset \) and \(G:E\rightarrow E\) be an \(\eta -\)strongly accretive mapping which is also \(\kappa -\)Lipschizian. Let \(\mu \in \Big (0,\min \Big \{1,(\frac{q\eta }{d_{q}\kappa ^{q}})^{\frac{1}{q-1}}\Big \}\Big )\) and \(\tau :=\mu \Big (\eta -\frac{\mu ^{q-1}d_{q}\kappa ^{q}}{q}\Big )\). For each \(t,\alpha \in (0,1)\) with \(\alpha < t\) and \(\beta \in (0,\frac{\tau }{\gamma })\). Assume that \(S:=\alpha u+(1-\delta )(1-\alpha )I+\delta \sum _{i\ge 1}\sigma _{i,\alpha }T_i,\) where \(\delta \) is some fixed number in (0, 1) and \(u\in E\), \(\alpha ,\sigma _{i,t}\) are in (0, 1). Define the following mapping \(W_{t}\) on E by
where t is in (0,1). Then \(W_t\) is a strict contraction mapping. Furthermore, for any \(x,y\in E,\)
Proof
Observe that, for any \(x,y\in E\)
Without loss of generality, assume \(\eta <\frac{1}{q}\). Then, as \(\mu <(\frac{q\eta }{d_{q}\kappa ^{q}})^{\frac{1}{q-1}}\), we have \(0<q\eta -\mu ^{q-1}d_{q}\kappa ^{q}\). Furthermore, from \(\eta <\frac{1}{q}\) we have \(q\eta -\mu ^{q-1}d_{q}\kappa ^{q}<1\) so that \(0<q\eta -\mu ^{q-1}d_{q}\kappa ^{q}<1\). Also as \(\mu <1\) and \(t\in (0,1)\), we obtain that \(0<t\mu (q\eta -\mu ^{q-1}d_{q}\kappa ^{q})<1\).
For each \(t,\alpha \in (0,1)\), then for any \(x,y\in E\), define \(K_{t}x=(1-t\mu G)Sx\), then from (3.1), we obtain
therefore
Hence
which implies that \(W_t\) is a strict contraction, by Banach contraction mapping principle, there exists a unique fixed point \(x_t\) of \(W_t\) in E. That is,
\(\square \)
Theorem 3.2
Let E be a q-uniformly real smooth Banach space which is also uniformly convex. Let \(T_{i},f,G,\mu ,\tau ,\beta ,\gamma ,S\) and F be as in Lemma 3.1. Let \(\{t_n\},\{\alpha _n\}\) be sequences in (0,1), such that \(\underset{n\rightarrow \infty }{\lim }t_n=0\) and \(\underset{n\rightarrow \infty }{\lim }\frac{\alpha _n}{t_n}=0\). Let \(\{x_{t_n}\}\) be a sequence satisfying (3.3), then
-
(i)
\(\{x_{t_n}\}\) is bounded for \(t_n\in (0,\frac{1}{\tau })\).
-
(ii)
\(\underset{n\rightarrow \infty }{\lim }||x_{t_n}-T_{i}x_{t_n}||=0, \quad \forall i\in \mathbb {N}.\)
-
(iii)
then \(\{x_{t_0}\}\) converges strongly to a common fixed point p in F which is a unique solution of the variational inequality
$$\begin{aligned} \langle (\mu G-\gamma f)p,j(p-x)\rangle \le 0,~~\forall x\in F. \end{aligned}$$(3.4)
Proof
Let \(p\in F\) and \(\alpha _n\le t_n\), then
Since \((1-\tau t_n)(\alpha _n/t_n)\rightarrow 0\) as \(n\rightarrow \infty ,\) then there exists \(n_0\in \mathbb {N}\) such that \((1-\tau t_{n})(\alpha _n/t_n)<(\tau -\gamma \beta )/2\) for \(n\ge n_0.\) Furthermore
for all \(n\ge n_0\). That is, \(||x_{t_n}-p||\le (\frac{||\gamma f(p)-\mu G(p)||}{\tau -\gamma \beta }+\frac{||u-p||}{2})\) for all \(n\ge n_0\). Thus \(\{x_{t_n}\}\) is bounded, so are \(\{f(x_{t_n})\},\{G(x_{t_n})\},\{T_i(x_{t_n})\}\) and \(\{G(T_ix_{t_n})\}\).
(ii) From (3.3), we have
Using Lemma 2.5, we have the following estimate
Therefore
Then we immediately obtain \(\underset{n\rightarrow \infty }{\lim }\sum _{i\ge 1}\sigma _{i,n}g(||T_{i}x_{t_n}-x_{t_n}||)=0,\) it follows that \(\underset{n\rightarrow \infty }{\lim }g(||T_{i}x_{t_n}-x_{t_n}||)=0~~\forall i\in \mathbb {N}.\) By the property of g we have that
(iii) By Lemma 2.6, \((\mu G-\gamma f)\) is strongly accretive, so the variational inequality (3.4) has a unique solution in F. Below we use \(q\in F\) to denote the unique solution of (3.4). Next, we prove that \(x_t\rightarrow q~(t\rightarrow 0).\)
Let \(\{t_{n}\}\) be a sequence in (0, 1) such that \(\{x_{t_{n}}\}\) satisfies (3.3). By writing \(\{x_{n}\}\) instead of \(\{x_{t_{n}}\},\) define a map \(\phi :E\rightarrow \mathbb {R}\) by
Then, \(\phi (y)\rightarrow \infty \) as \(||y||\rightarrow \infty \), \(\phi \) is continuous and convex, so as E is reflexive, there exists \(q \in E\) such that \(\phi (q)=\underset{u \in E}{\min }\phi (u)\). Hence, the set
Since \(\underset{n\rightarrow \infty }{\lim }||x_{n}-T_{i}x_{n}||=0,\) \(\underset{n\rightarrow \infty }{\lim }||x_{n}-T_{i}^{m}x_{n}||=0,~ \mathrm{for~ any}~m\ge 1\) and \(i\in \mathbb {N}\) by induction. Now let \(v\in K^*\), we have for any \(i\in \mathbb {N}\)
and hence \(T_{i}v\in K^{*}\).
Now let \(z\in F,\) then \(z=T_{i}z,\) for any \(i\in \mathbb {N}\). Since \(K^{*}\) is a closed convex set, there exists a unique \(v^{*}\in K^{*}\) such that
But, for any \(i\in \mathbb {N}\)
which implies \(v^{*}=T_{i}v^{*}\) and so \(K^{*}\cap F\ne \emptyset .\)
Let \(p \in K^* \cap F\) and \(\epsilon \in (0,1)\). Then, it follows that \(\phi (p)\le \phi (p-\epsilon (G-\gamma f)p)\) and using Lemma 2.1, we obtain that
which implies
Moreover,
Since j is norm-to-norm uniformly continuous on bounded subsets of E, and \(\epsilon \rightarrow 0\) we have that
Now from (3.5) and since \(\underset{n\rightarrow \infty }{\lim }\frac{\alpha _n}{t_n}=0\), we have
and so
Thus there exist a subsequence say \(\{x_{n_{j}}\}\) of \(\{x_{n}\}\) such that \({lim}_{j\rightarrow \infty }x_{n_{j}}=p.\)
By definition of \(S_{\alpha }\) as \(S_{\alpha _n}x_n:=\alpha _n u+ (1-\delta )(1-\alpha _n)x_n+\delta \sum _{i\ge 1}\sigma _{i,n}T_ix_n,\) which implies \(S_{\alpha _{n_j}}x_{n_j}:=x_{n_j}+\alpha _{n_j} (u-x_{n_j})+ \delta \sum _{i\ge 1}\sigma _{i,n_j}(T_ix_{n_j}-x_{n_j}),\) then \({lim}_{j\rightarrow \infty }Sx_{n_{j}}={lim}_{j\rightarrow \infty }x_{n_{j}}=p\) and \(S_{\alpha }p=p.\) Thus for any \(z\in F,\) using (3.3) we have
since \(\langle (I-S)x_{n_{j}}-(I-S)p, j(x_{n_{j}}-z)\rangle \ge 0.\) As G is Lipschitzian and the fact that \(\Vert x_{n_{j}}-Sx_{n_{j}}\Vert \le \alpha _{n_j} \Vert u-x_{n_j}\Vert + \delta \sum _{i\ge 1}\sigma _{i,n_j}\Vert (T_ix_{n_j}-x_{n_j})\Vert \rightarrow 0\) as \(j\rightarrow \infty ,\) we have \(Gx_{n_{j}}-GSx_{n_{j}}\rightarrow 0~\mathrm{as}~j\rightarrow \infty .\) From this and (3.9), taking limit as \(j\rightarrow \infty \) we obtain
Hence p is the unique solution of the variational inequality (3.4). Now assume there exists another subsequence of \(\{z_{n}\}\) say \(\{x_{n_{k}}\}\) such that \({lim}_{k\rightarrow \infty }x_{n_{k}}=p^{*}.\) Then, using (3.8) we have \(p^{*}\in F.\) Repeating the above argument with p replaced by \(p^{*}\) we can easily obtain that \(p^{*}\) also solved the variational inequality (3.4). By uniqueness of the solution of the variational inequality, we obtained that \(p=p^{*}\) and this completes the proof. \(\square \)
Theorem 3.3
Let E be a q-uniformly real smooth Banach space which is also uniformly convex. Let \(T_{i}:E\rightarrow E~~i\in \{1,2,\ldots \}\) be a family of nonexpansive mappings with \(F:=\cap ^{\infty }_{i=1}F(T_{i})\ne \emptyset \). Let \(G:E\rightarrow E\) be an \(\eta \)-strongly accretive map which is also \(\kappa \)-Lipschitzian. Let \(f:E\rightarrow E\) be a contraction map with coefficient \(0<\beta <1\). Let \(\{\alpha _{n}\}\) and \(\{\beta _{n}\}\) be sequences in (0,1) satisfying:
-
(i)
\(\underset{n\rightarrow \infty }{\lim }\beta _{n}=0~\) and \(~\sum ^{\infty }_{n=0}\beta _{n}=\infty ;\)
-
(ii)
\(\sum \alpha _{n}<\infty \) and \(\underset{n\rightarrow \infty }{\lim }\frac{\alpha _n}{\beta _n}=0\);
-
(iii)
\(\underset{n\rightarrow \infty }{\lim }\sum _{i\ge 1}|\sigma _{i,n+1}-\sigma _{i,n}|=0\) and \(\sum _{i\ge 1}\sigma _{i,n}=1-\alpha _{n}.\)
Let \(\mu , \gamma ,\) and \(\tau \) be as in Lemma 3.1 and \(\delta \in (0,1)\) be fixed. define a sequence \(\{x_{n}\}_{n=1}^{\infty }\) iteratively in E by \(x_{0}\in E\)
Then, \(\{x_{n}\}^{\infty }_{n=1}\) converges strongly to \(x^{*}\in F\) which is also a solution to the following variational inequality
Proof
Since \((\mu G-\gamma f)\) is strongly accretive, then the variational inequality (3.11) has a unique solution in F. Now we show that \(\{x_{n}\}^{\infty }_{n=1}\) is bounded. Let \(p\in F\) then, for every \(i\in \mathbb {N}, T_{i}p=p\). From (3.10), we obtain
Also from (3.10) and (3.12), we obtain
Therefore, \(\{x_{n}\}\) is bounded. Hence \(\{y_n\},\{T_ix_n\},\{Gy_n\},\{GT_iy_n\}\) and \(\{f(y_n)\}\) are also bounded.
Next, we show that \({\lim }_{n\rightarrow \infty }||x_{n+1}-x_{n}||=0.\) Define two sequences \(\{\lambda _{n}\}\) and \(\{z_{n}\}\) by \(\lambda _n:=(1-\delta )\alpha _n+\delta \) and
Observe that \(\{z_n\}\) is bounded and that
for some real number \(M:={\sup }_{n\ge 1}\{||\gamma f(x_n)-\mu G(y_n)||,||T_{i}x_n||,i=1,2,...\}\).
This implies
and by Lemma 2.4, we obtain
Hence
and from (3.10), we also obtain
from (3.13) and (3.14), we have
Next we show that \(\underset{n\rightarrow \infty }{\lim }||T_ix_n-x_n||=0\) for all \(i\in \mathbb {N}.\) Since \(p\in F\), using the same argument in (3.7), we obtain
From (3.15) and \(\lim _{n\rightarrow \infty }\alpha _{n}=0\), we obtain
it follows that for every \(i\in \mathbb {N},\)
Let \(z_t=t\gamma f(z_t)+(1-t\mu G)Sz_t\), where \(S:=\alpha u+(1-\delta )(1-\alpha )I+\delta \sum _{i\ge 1}\sigma _{i,\alpha }T_i,\) as in Theorem 3.1. Then,
Hence
Therefore
Now, taking limit superior as \(n\rightarrow \infty \) firstly, and then as \(t\rightarrow 0\), we have
Moreover, we note that
Taking limit superior as \(n\rightarrow \infty \) in (3.18), we have
since E has a uniformly Gâteaux differentiable norm, so j is norm-to-norm\(^{*}\) uniformly continuous on bounded subset of E. Then, from Theorem 3.1 (i.e., \(z_t\rightarrow p~~(t\rightarrow 0^{+}\))), we obtain
hence, using (3.17) in (3.19), we obtain
Finally, we show that \(x_n\rightarrow p.\) From the recursion formula (3.10), by using (2.1) and taking \(n\ge N\) where \(N\in \mathbb {N}\) is large enough, we obtain
On the other hand
Since \(\{x_n\}\) and \(\{f(x_n)\}\) are bounded, we pick a constant \(G_0>0\) such that
Therefore
Hence
where \(\theta _n:=\beta _n\Big (\alpha _n/\beta _n||u-p||^2+2G_{0}(\sqrt{\alpha _n}+\sqrt{\beta _n}). +2\langle \gamma f(p)-\mu G(p),j(x_{n+1}-p)\rangle \Big )\) By using Lemma 2.3 we obtain \(x_n\rightarrow p\) as \(n\rightarrow \infty \). This complete the proof. \(\square \)
Corollary 3.4
Let H be a real Hilbert space, \(\{z_{t}\}_{t\in (0,1)},\) be as in Theorem 3.2. Then \(\{z_{t}\}\) converges strongly to a common fixed points of the family \(\{T_{i}\}_{i=1}^{\infty }\) say p which is a unique solution of the variational inequality
Corollary 3.5
Let H be a real Hilbert space and let C a nonempty closed convex subset of H. Let \(G:H \rightarrow H,\) \(f:E \rightarrow E,\) \(\{T_{i}\}^{\infty }_{i=1}\) F, \(\{\alpha _{n}\}^{\infty }_{n = 1}\), \(\{\beta _{n}\}^{\infty }_{n =1}\) and \(\{x_{n}\}^{\infty }_{n = 1}\) be as in Theorem (3.1), then \(\{x_{n}\}^{\infty }_{n = 1}\) converges strongly to \(p \in F\), which is also the unique solution of the variational inequality
References
Agarwal, R.P., ORegan, D., Sahu, D.R.: Fixed point theory for lipschitzian-type mappings with applications. Springer, Dordrecht Heidelberg, London, New York (2000)
Ali, B.: Iterarive approximation of common fixed points for families of nonexpansive mappings and solutions of variational inequalitites. Adv. Nonlinear Var. Inequal. 12(2), 65–81 (2009)
Bynum, W.L.: Normal structure coefficients for Banach spaces. Pacific J. Math. 86, 427–436 (1980)
Ali, B., Ugwunnadi, G.C., Shehu, Y.: A general iterative algorithm for nonexpansive mappings in Banach spaces. Ann. Funct. Anal. 2(2), 11–22 (2011)
Chidume, C.E.: Geometric properties of Banach spaces and nonlinear iterations, Springer Verlag Series: Lecture Notes in Mathematics, vol. 1965, XVII, p. 326 (2009). ISBN: 978-1-84882-189-7
Chidume, C.E., Li, J., Udomene, A.: Convergence of paths and approximation of fixed points of asymptotically nonexpansive mappings. Proc. Am. Math. Soc. 133(2), 473480 (2005)
Jung, J.S.: Iterative approaches to common fixed points of nonexpansive mappings in Banach spaces. J. Math. Anal. Appl. 302(2), 509520 (2005)
Iiduka, H., Takahashi, W.: Strong convergence theorems for nonexpansive mappings and inverse-strongly monotone mappings. Nonlinear Anal. 61, 341–350 (2005)
Korpelevich, G.M.: An extragradient method for finding saddle points and for other problems. Ekonomika i Matematicheskie Metody 12, 747–756 (1976)
Lim, T.C.: Characterization of normal structure. Proc. Am. Math. Soc. 43, 313–319 (1974)
Maingé, P.E.: Strong convergence of projected subgradient methods for nonsmooth and non- strictly convex minimization. Set Valued Anal. 16, 899–912 (2008)
Maingé, P.E.: Approximation methods for common xed points of nonexpan- sive mappings in Hilbert space. J. Math. Anal. Appl. 325, 469–479 (2007)
Marino, G., Xu, H.K.: A general iterative method for nonexpansive mappings in Hilbert spaces. J. Math. Anal. Appl. 318(1), 4352 (2006)
Moudafi, A.: Viscosity approximation methods for fixed-point problems. J. Math. Anal. Appl. 241(1), 4655 (2000)
Noor, M.A.: A class of new iterative methods for solving mixed variational inequalities. Math. Comput. Modell. 31, 11–19 (2000)
Stampacchi, G.: Formes bilineaires coercivites sur les ensembles convexes, vol. 258, pp. 4413–4416. C. R. Acad. Sciences, Paris (1964)
Suzuki, T.: Strong convergence of Krasnoselskii and Manns type sequences for one-parameter nonexpansive semigroups without Bochner integrals. J. Math. Anal. Appl. 305(1), 227–239 (2005)
Tian, M.A.: A general iterative method for nonexpansive mappings in Hilbert space. Nonlinear Anal. 73, 689–694 (2010)
Xu, H.K.: An iterative approach to quadratic optimization. J. Optim. Theory Appl. 116, 659678 (2003)
Xu, H.K.: Inequalities in Banach spaces with applications. Nonlinear Anal. 16, 1127–1138 (1991)
Xu, H.K.: Iterative algorithms for nonlinear operators. J. London Math. Soc. (2) 66(1), 240–256 (2002)
Xu, H.-K.: Inequality in Banach spaces with applications. Nonlinear Anal. 16(12), 1127–1138 (1991)
Xu, Z.B., Roach, G.F.: Characteristic inequalities of uniformly smooth Banach spaces. J. Math. Anal. Appl. 157(1), 189210 (1991)
Yamada, I.: The hybrid steepest-descent method for variational inequality problems over the intersection of the fixed point sets of nonexpansive mappings, Inherently parallel algorithms in feasibility and optimization and their applications. In: Butnariu, D., Censor, Y., Reich, S. (eds.), pp. 473–504. North-Holland, Amsterdam, Holland (2001)
Yao, Y., Liou, Y.-C., Li, C-L., Lin H.-T.,: Extended extragradient methods for generalized varitional inequalities, fixed point Theory and applications. Volume 2012, Article ID 237083, p. 14
Yao, Y., Xu, H.-K.: Iterative methods for finding minimum-norm fixed points of nonexpan- sive mappings with applications. Optimization 60(6), 645–658 (2011)
Acknowledgments
The author is grateful to the editor and the reviewers suggestions which improved the contents of the article.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Chidi, U.G. Modified general iterative algorithm for an infinite family of nonexpansive mappings in Banach spaces. Afr. Mat. 28, 221–235 (2017). https://doi.org/10.1007/s13370-016-0441-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13370-016-0441-0