- Research
- Open access
- Published:
Hybrid steepest iterative algorithm for a hierarchical fixed point problem
Fixed Point Theory and Applications volume 2017, Article number: 25 (2017)
Abstract
The purpose of this work is to introduce and study an iterative method to approximate solutions of a hierarchical fixed point problem and a variational inequality problem involving a finite family of nonexpansive mappings on a real Hilbert space. Further, we prove that the sequence generated by the proposed iterative method converges to a solution of the hierarchical fixed point problem for a finite family of nonexpansive mappings which is the unique solution of the variational inequality problem. The results presented in this paper are the extension and generalization of some previously known results in this area. An example which satisfies all the conditions of the iterative method and the convergence result is given.
1 Introduction
Throughout this paper, we always assume that \(\mathcal{V}\) is a real Hilbert space with the inner product \(\langle\cdot,\cdot\rangle\) and the norm \(\|\cdot\|\), respectively. Let a nonlinear mapping \(S : \mathcal {V}\rightarrow\mathcal{V}\) be a nonexpansive operator if
A point \(u\in\mathcal{V}\) is said to be a fixed point of S provided \(Su=u\). In this paper, we use \(\overline{F}(S)\) to denote the fixed point set which is closed and convex, see [1].
Let \(S:W\rightarrow\mathcal{V}\) be a nonexpansive mapping, where W is a nonempty closed convex subset of \(\mathcal{V}\). The hierarchical fixed point problem (in short, HFPP) is to find \(u\in\overline{F}(S)\) such that
Many authors solve (1.1) by various methods, see [2–9] and the references therein.
Yao et al. [2] proposed the following iterative algorithm to solve HFPP (1.1):
where \(\{a_{n}\}\) and \(\{b_{n}\}\) are sequences in \((0,1)\) and \(g:W\rightarrow\mathcal{V}\) is a contraction mapping, and the sequence \(\{u_{n}\}\) generated by (1.2) converges strongly to \(z\in\overline {F}(S)\), which is also a unique solution of the variational inequality problem (VIP), i.e., to find \(z\in\overline{F}(S)\) such that
After that, Ceng et al. [6] introduced the following algorithm:
where F is a Lipschitz continuous and strongly monotone mapping, g is a Lipschitz continuous mapping. Compute an iterative sequence \(\{ u_{n}\}\) generated by (1.4) converging strongly to \(z\in\overline {F}(S)\), which is also a unique solution of the following variational inequality problem (VIP), i.e., to find \(z\in\overline{F}(S)\) such that
By using a \(T_{n}\)-mapping [10], Yao [11] proposed the following iterative method:
where \(c>0\), A is a strongly positive bounded linear operator and \(g:W\rightarrow\mathcal{V}\) is a contraction mapping.
Further, Ceng et al. [12] proposed explicit and implicit iterative schemes for finding a common solution for the set of fixed points of a nonexpansive mapping. Buong and Duong [13] studied the explicit iterative algorithm for finding the approximate solution of a VIP defined over the set of common fixed points of a finite number of nonexpansive mappings:
where \(S^{k}_{i}=(1-b^{i}_{k})u_{k}+b^{i}_{k}S^{i}\) for \(1\leq i\leq p\), \(\{S_{i}\}^{p}_{i=1}\) are p-nonexpansive mappings on a real Hilbert space \(\mathcal{V}\), \(S^{k}_{0}=I-\lambda_{k}\mu F\), and F is an η-strongly monotone and L-Lipschitz continuous mapping.
Very recently, Zhang and Yang [14] studied the more general explicit iterative algorithm
where g is an α-Lipschitzian, F is an η-strongly monotone and L-Lipschitz continuous mapping and \(S^{k}_{i}=(1-b^{i}_{k})u_{k}+b^{i}_{k}S^{i}\) for \(1\leq i\leq p\). Under some assumptions, compute an iterative sequence \(\{u_{k}\}\) proposed by the iterative algorithm (1.8) that strongly converges to the solution of the VIP, i.e., to find \(z\in\bigcap^{p}_{i=1}\overline {F}(S_{i})\) such that
Inspired and motivated by the recent research, we develop an iterative algorithm for a hierarchical fixed point problem of a finite family of nonexpansive mappings on the real Hilbert space. We generate a strong convergence theorem for the sequence considered by the generalized method. Numerical examples are also given for the theoretical verification of the algorithm. The algorithm and results presented in this paper improve and extend some recent corresponding algorithms and results; see [15, 16] and the references therein.
2 Preliminaries
We recall some concepts and results which are needed in the sequel.
Definition 2.1
Let \(S:W\rightarrow\mathcal{V}\) be a mapping which is said to be
-
(i)
monotone if
$$ \langle Su-Sv,u-v\rangle\geq0, \quad\forall u,v\in W; $$ -
(ii)
strongly monotone if there exists a constant \(\alpha>0\) such that
$$ \langle Su-Sv,u-v\rangle\geq\alpha\|u-v\|^{2}, \quad\forall u,v\in W; $$ -
(iii)
Lipschitz continuous if there exists a constant \(k>0\) such that
$$ \|Su-Sv\|\leq k\|u-v\|, \quad\forall u,v\in W. $$If \(k=1\), then S is called nonexpansive.
Definition 2.2
A mapping \(g:W\rightarrow\mathcal{V}\) is said to be σ-contraction if there exists a constant \(\sigma\in(0,1)\) such that
Lemma 2.1
([6])
Let \(F:W\rightarrow\mathcal{V}\) be an η-strongly monotone and k-Lipschitz continuous mapping and \(g:W\rightarrow \mathcal{V}\) be a τ-Lipschitz continuous mapping. Then the mapping \(\mu F-\rho g\) is \((\mu\eta-\rho\tau)\)-strongly monotone with condition \(\mu\eta>\rho\tau\geq0\), i.e.,
Definition 2.3
A mapping \(T:\mathcal{V}\rightarrow\mathcal{V}\) is said to be an averaged mapping if it can be written as the average of the identity I and a nonexpansive mapping, i.e.,
where \(\alpha\in(0,1)\) and \(S:\mathcal{V}\rightarrow\mathcal{V}\) is nonexpansive.
Lemma 2.2
If the mappings \(\{S_{i}\}^{p}_{i=1}\) are averaged and have a common fixed point, then
In particular, if \(p=2\), we have \(\overline{F}(S_{1})\cap\overline {F}(S_{2})=\overline{F}(S_{1}S_{2})=\overline{F}(S_{2}S_{1})\).
Lemma 2.3
([19])
Let \(\{\alpha_{n}\}\) be a sequence of nonnegative real numbers such that
where \(\{w_{n}\}\in(0,1)\) and \(\{t_{n}\}\) is a sequence such that
-
(i)
\(\sum^{\infty}_{n=1}w_{n}=\infty\);
-
(ii)
\(\lim\sup_{n\rightarrow\infty}\frac{t_{n}}{w_{n}}\leq0\) or \(\sum^{\infty}_{n=1}|t_{n}|<\infty\).
Then \(\lim_{n\rightarrow\infty}\alpha_{n}=0\).
Lemma 2.4
([1])
Let \(S:W\rightarrow W\) be a nonexpansive mapping with \(\overline{F}(S)\neq\varnothing\). Then the mapping \(I-S\) is demiclosed at 0, that is, if \(\{u_{n}\}\) is a sequence converging weakly to u and \(\{(I-S)u_{n}\}\) converges strongly to 0, then \((I-S)u=0\).
Lemma 2.5
([20])
Let \(F:W\rightarrow\mathcal{V}\) be an η-strongly monotone and k-Lipschitzian mapping. Let \(\frac{2\eta}{k^{2}}>\mu>0\), \(Q=I-\lambda\mu F\). Then Q is a \((1-\lambda\tau)\)-contraction mapping with \(\min\{1,\frac {1}{\tau}\}>\lambda>0\), that is,
where \(\tau=1-\sqrt{1-\mu(2\eta-\mu k^{2})}\in(0,1]\).
Lemma 2.6
Let \(\mathcal{V}\) be a real Hilbert space. The following inequality holds:
3 Main results
In this section, we establish an iterative method for finding the solution of hierarchical fixed point problem (1.1).
Let W be a nonempty closed convex subset of a real Hilbert space \(\mathcal{V}\), and let \(\{S_{i}\}^{p}_{i=1}\) be p nonexpansive mappings on W such that \(\Xi=\bigcap^{p}_{i=1}\overline{F}(S_{i})\neq \emptyset\). Let \(F:W\rightarrow W\) be an η-strongly monotone and k-Lipschitzian mapping and \(g:W\rightarrow W\) be a τ-contraction mapping.
We consider the following hierarchical fixed point problem (in short, HFPP): find \(u\in\Xi\) such that
Now we define the following algorithm for finding a solution of HFPP (3.1).
Algorithm 3.1
Given arbitrarily \(u_{0}\in W\), compute sequences \(\{u_{n}\}\) and \(\{ v_{n}\}\) by the iterative schemes
where \(S^{n}_{i}=(1-d^{i}_{n})I+d^{i}_{n}S_{i}\) and \(d^{i}_{n}\in(0,1)\) for \(i=1,2,\ldots ,p\), let the parameters satisfy \(\frac{2\eta }{k^{2}}>\mu>0\) and \(\frac{\nu}{\tau}>\rho> 0\), where \(\nu=\mu(\eta -\frac{\mu k^{2}}{2})\) and \(\{a_{n}\}\), \(\{b_{n}\}\) and \(\{c_{n}\}\) are sequences in \((0,1)\) satisfying the following conditions:
-
(i)
\(\lim_{n\rightarrow\infty}a_{n}=0\) and \(\sum_{n=1}^{\infty }a_{n}=\infty\) and \(\sum_{n=1}^{\infty}|a_{n-1}-a_{n}|<\infty\).
-
(ii)
\(\{b_{n}\}\subset[\sigma,1)\) and \(\lim_{n\rightarrow\infty }b_{n}=b<1\).
-
(iii)
\(a_{n}+c_{n}<1\) and \(\lim_{n\rightarrow\infty}c_{n}=0\).
-
(iv)
\(\sum_{n=1}^{\infty}|c_{n-1}-c_{n}|<\infty\) and \(\sum_{n=1}^{\infty}|d_{n-1}^{i}-d_{n}^{i}|<\infty\) for \(i=1,2,\ldots,p\).
Lemma 3.1
Let \(u^{*}\in\Xi\). Then the sequences \(\{u_{n}\}\) and \(\{v_{n}\}\) defined in Algorithm 3.1 are bounded.
Proof
Let \(u^{*}\in\Xi\). So, we have
where the third and fifth inequalities follow from (3.3) and the second inequality follows from Lemma 2.5.
By induction on n and (3.4), we have
Hence, \(\{u_{n}\}\) is bounded; and consequently, we get \(\{v_{n}\}\), \(\{ Sv_{n}\}\), \(\{S_{1}u_{n+1}\}\), \(\|S_{1}^{n}u_{n+1}\|\), \(\|S_{2}S_{1}^{n}u_{n+1}\|,\ldots, \|S^{n}_{p-1}\cdots S^{n}_{1}u_{n+1}\|, \|S_{p}S^{n}_{p-1}\cdots S^{n}_{1}u_{n+1}\|, \|S^{n-1}_{p-1}\cdots S^{n-1}_{1}v_{n}\|+ \|S_{p}S^{n-1}_{p-1}\cdots S^{n-1}_{1}v_{n}\|+\|S^{n-1}_{p-1}\cdots S^{n-1}_{1}u_{n}\| +\|S_{p}S^{n-1}_{p-1}\cdots S^{n-1}_{1}u_{n}\|\) and \(\{g(v_{n})\}\) are bounded. □
Lemma 3.2
Let \(\{u_{n}\}\) be a sequence generated by Algorithm 3.1. Then
-
(i)
\(\lim_{n\rightarrow\infty}\|u_{n+1}-u_{n}\|=0\).
-
(ii)
\(\lim_{n\rightarrow\infty}\|u_{n}-S_{p}^{n}S^{n}_{p-1}\cdots S^{n}_{1}u_{n}\|=0\).
Proof
From the sequence \(\{v_{n}\}\) defined in Algorithm 3.1, we have
From the definition of \(S^{n}_{i}\) it follows that
and from (3.6), we have
By induction on p, it follows that
Similarly,
From (3.5), (3.8) and (3.9), it follows that
where
From conditions (i) and (iv) of Algorithm 3.1 and Lemma 2.3, we have
From (3.2), we have
From (3.11), we have
Since from (i), (ii), (iii) and (3.10), we have
 □
Lemma 3.3
Let
Then \(u_{n}\) converges strongly to \(\tilde{u}\in\Xi\) as \(n\rightarrow0\).
Proof
Since \(\{u_{n}\}\) is bounded, we assume that \(\{u_{n}\}\) converges weakly to a point \(\tilde{u}\in W\). From Lemma 2.4, we have \(\tilde {u}\in\Xi\). Now, for \(\tilde{u}\in\Xi\), we get
Hence,
Since \(u_{n}\rightharpoonup\tilde{u}\), from (3.13) we obtain \(u_{n}\rightarrow\tilde{u}\). □
Theorem 3.1
The sequence \(\{u_{n}\}\) generated by Algorithm 3.1 converges strongly to \(z\in\Xi=\bigcap^{p}_{i=1}\overline{F}(S_{i})\), which is also a unique solution of the HFPP
Proof
Let \(u_{t}\in W\) be a unique fixed point. Now, we claim that
where \(z=\lim_{t\rightarrow0}u_{t}\). It follows from Lemma 3.3 that \(z\in\Xi\).
By using Lemma 2.6, we get
From the above we have
where \(\mathcal{A}_{n}(t)=[1-[(1-a_{n}\nu)^{2}+2a_{n}\rho\tau]]\).
Further,
where \(\mathcal{M}>0\) is a constant such that \(\mathcal{M}\geq\| u_{n}-u_{t}\|^{2}\).
Taking the limsup as \(t\rightarrow0\) in (3.14), we get
Now, we have to show that \(u_{n}\rightarrow z\).
Further,
which implies that
Let \(w_{n}= [\frac{2a_{n}(\nu-\rho\tau)}{1+a_{n}(\nu-\rho\tau)} ]\) and
We have \(\sum_{n=1}^{\infty}a_{n}=\infty\) and \(\lim_{n\rightarrow\infty }\sup \{\frac{1}{\nu-\rho\tau}\langle\rho g(z)-\mu F(z),u_{n+1}-z\rangle \}\leq0\). It follows that \(\sum_{n=1}^{\infty}w_{n}=\infty\) and \(\lim_{n\rightarrow\infty}\sup\frac{t_{n}}{w_{n}}\leq0\). Thus, all the conditions of Lemma 2.3 are fulfilled. Hence, \(u_{n}\rightarrow z\). □
4 Examples
The following example ensures that all the conditions of Algorithm 3.1 and the convergence result are fulfilled.
Example 4.1
Let \(a_{n}=\frac{1}{3n}\), \(b_{n}=\frac{2n-1}{3n}\) and \(c_{n}=\frac {1}{3n}\). Then
and
The sequence \(\{a_{n}\}\) satisfies condition (i) of Algorithm 3.1.
Now we compute
So,
Similarly, we can show
The sequences \(\{a_{n}\}\), \(\{b_{n}\}\) and \(\{c_{n}\}\) satisfy conditions (i), (ii) and (iii).
Let \(d^{i}_{n}=\frac{n}{n+i}\) for \(i=1,2\). Then
Hence the sequence \(\{d^{i}_{n}\}\) also satisfies condition (iv) of Algorithm 3.1.
Let \(S_{1},S_{2}:\mathbb{R}\rightarrow\mathbb{R}\) be defined by
and
and let the mapping \(g:\mathbb{R}\rightarrow\mathbb{R}\) be defined by
It is easy to verify that \(S_{1}\) and \(S_{2}\) are \(\frac {1}{2}\)-nonexpansive and g is a \(\frac{1}{2}\)-contraction mapping.
Further,
Suppose that the mapping \(F:\mathbb{R}\rightarrow\mathbb{R}\) is defined by
Hence, F is 2-strongly monotone and 2-Lipschitzian.
Assume that \(\rho=\frac{1}{15}\) and \(\mu=\frac{1}{5}\) and they satisfy \(0<\mu<\frac{2\eta}{k^{2}}\) and \(0\leq\rho\tau<\nu\), where \(\nu=1-\sqrt {1-\mu(2\eta-\mu k^{2})}\).
All codes were written in Matlab, the values of \(\{v_{n}\}\) and \(\{ u_{n}\}\) with different n are given in Table 1.
Remark 4.1
Table 1 and Figures 1 and 2 show that the sequences \(\{v_{n}\}\) and \(\{ u_{n}\}\) converge to 0. Also, \(\{0\}\in\Xi\).
5 Conclusion
We have analyzed an iterative method for finding an approximate solution of hierarchical fixed point problem (1.1) and variational inequality problem (1.5) involving a finite family of nonexpansive mappings in a real Hilbert space. This method can be viewed as a modification and improvement of some existing methods [15, 16] for solving the variational inequality problem and the hierarchical fixed point problem. Therefore, Algorithm 3.1 is expected to be widely applicable.
References
Geobel, K, Kirk, WA: Topics in Metric Fixed Point Theory. Cambridge University Press, Cambridge (1990)
Yao, Y, Cho, YJ, Liou, YC: Iterative algorithms for hierarchical fixed points problems and variational inequalities. Math. Comput. Model. 52(9-10), 1697-1705 (2010)
Crombez, G: A hierarchical presentation of operators with fixed points on Hilbert spaces. Numer. Funct. Anal. Optim. 27, 259-277 (2006)
Moudafi, A: Krasnoselski-Mann iteration for hierarchical fixed-point problems. Inverse Probl. 23(4), 1635-1640 (2007)
Gu, G, Wang, S, Cho, YJ: Strong convergence algorithms for hierarchical fixed points problems and variational inequalities. J. Appl. Math. 2011, Article ID 164978 (2011)
Ceng, LC, Ansari, QH, Yao, JC: Iterative methods for hierarchical variational inequalities in Hilbert spaces. J. Optim. Theory Appl. 151, 489-512 (2011)
Xu, HK: Viscosity method for hierarchical fixed point approach to variational inequalities. Taiwan. J. Math. 14(2), 463-478 (2010)
Wang, Y, Xu, W: Strong convergence of a modified iterative algorithm for hierarchical fixed point problems and variational inequalities. Fixed Point Theory Appl. 2013, Article ID 121 (2013)
Ansari, QH, Ceng, LC, Gupta, H: Triple hierarchical variational inequalities. In: Ansari, QH (ed.) Nonlinear Analysis: Approximation Theory, Optimization and Applications, pp. 231-280. Springer, New York (2014)
Takahashi, W, Atsushiba, S: Strong convergence theorems for a finite family of nonexpansive mappings and applications. Indian J. Math. 41(3), 435-453 (1999)
Yao, Y: A general iterative method for a finite family of nonexpansive mappings. Nonlinear Anal. 66, 2676-2678 (2007)
Ceng, LC, Khan, AR, Ansari, QH, Yao, JC: Viscosity approximation methods for strongly positive and monotone operators. Fixed Point Theory 10(1), 35-71 (2009)
Buong, N, Duong, LT: An explicit iterative algorithm for a class of variational inequalities in Hilbert spaces. J. Optim. Theory Appl. 151, 513-524 (2011)
Zhang, C, Yang, C: A new explicit iterative algorithm for solving a class of variational inequalities over the common fixed points set of a finite family of nonexpansive mappings. Fixed Point Theory Appl. 2014, Article ID 60 (2014)
Bnouhachem, A, Ansari, QH, Yao, JC: An iterative algorithm for hierarchical fixed point problems for a finite family of nonexpansive mappings. Fixed Point Theory Appl. 2015, Article ID 111 (2015)
Bnouhachem, A, Ansari, QH, Yao, JC: Strong convergence algorithm for hierarchical fixed point problems of a finite family of nonexpansive mappings. Fixed Point Theory 17, 47-62 (2016)
Byrne, C: A unified treatment of some iterative algorithms in signal processing and image reconstruction. Inverse Probl. 20, 103-120 (2004)
Combettes, PL: Solving monotone inclusions via compositions of nonexpansive averaged operators. Optimization 53, 475-504 (2004)
Xu, HK: Iterative algorithms for nonlinear operators. J. Lond. Math. Soc. 66, 240-256 (2002)
Deng, BC, Chen, T, Li, ZF: Cyclic iterative method for strictly pseudononspreading in Hilbert spaces. J. Appl. Math. 2012, Article ID 435676 (2012)
Author information
Authors and Affiliations
Contributions
Both the authors contributed equally and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare that they have no competing interests.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Cite this article
Husain, S., Singh, N. Hybrid steepest iterative algorithm for a hierarchical fixed point problem. Fixed Point Theory Appl 2017, 25 (2017). https://doi.org/10.1186/s13663-017-0618-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s13663-017-0618-8