- Research
- Open access
- Published:
Strong convergence theorems for a class of split feasibility problems and fixed point problem in Hilbert spaces
Journal of Inequalities and Applications volume 2018, Article number: 289 (2018)
Abstract
In this paper we consider a class of split feasibility problem by focusing on the solution sets of two important problems in the setting of Hilbert spaces. One of them is the set of zero points of the sum of two monotone operators and the other is the set of fixed points of mappings. By using the modified forward–backward splitting method, we propose a viscosity iterative algorithm. Under suitable conditions, some strong convergence theorems of the sequence generated by the algorithm to a common solution of the problem are proved. At the end of the paper, some applications and the constructed algorithm are also discussed.
1 Introduction
Many applications of the split feasibility problem (SFP), which was first introduced by Censor and Elfving [1], have appeared in various fields of science and technology, such as in signal processing, medical image reconstruction and intensity-modulated radiation therapy (for more information, see [2, 3] and the references therein). In fact, Censor and Elfving [1] studied SFP in a finite-dimensional space, by considering the problem of finding a point
where C and Q are nonempty closed convex subsets of \(\Bbb {R}^{n}\), and A is an \(n\times n\) matrix. They introduced an iterative method for solving SFP.
On the other hand, variational inclusion problems are being used as mathematical programming models to study a large number of optimization problems arising in finance, economics, network, transportation and engineering science. The formal form of a variational inclusion problem is the problem of finding \(x^{*}\in H\) such that
where \(B:H \to 2^{H}\) is a set-valued operator. If B is a maximal monotone operator, the elements in the solution set of problem (1.2) are called the zeros of this maximal monotone operator. This problem was introduced by Martinet [4], and later it has been studied by many authors. It is well known that the popular iteration method that was used for solving problem (1.2) is the following proximal point algorithm: for a given \(x\in H\),
where \(\{\lambda_{n}\}\subset (0,\infty)\) and \(J^{B}_{\lambda_{n}}=(I+ \lambda_{n}B)^{-1}\) is the resolvent of the considered maximal monotone operator B corresponding to \(\lambda_{n}\) (see, also [5–9] for more details).
In view of SFP and the fixed point problem, very recently, Montira et al. [10] considered the problem of finding a point \(x^{*}\in H\) such
where \(A:H_{1}\to H_{1}\) is a monotone operator, and \(B:H_{1}\to 2^{H _{1}}\) is a maximal monotone operator, \(L:H_{1}\to H_{2}\) is a bounded linear operator and \(T:H_{2}\to H_{2}\) is a nonexpansive mapping.
They considered the following iterative algorithm: for any \(x_{0} \in H_{1}\),
where \(\{\lambda_{n}\}\) and \(\{\gamma_{n}\}\) satisfy some suitable control conditions, and \(J^{B}_{\lambda_{n}}\) is the resolvent of a maximal monotone operator B associated to \(\lambda_{n}\), and proved that sequence (1.4) weakly converges to a point \(x^{*}\in \Omega^{A+B} _{L,T}\), where \(\Omega^{A+B}_{L,T}\) is the solution set of problem (1.3).
Motivated by the work of Montira et al. [10] and the research in this direction, the purpose of this paper is to study the following split feasibility problem and fixed point problem: find \(x^{*} \in H\) such that
where A, B, L are the same as in (1.3) and \(S : H_{1} \to H_{1}\) is a nonexpansive mapping. By using a modified forward–backward splitting method, we propose a viscosity iterative algorithm (see (3.4) below). Under suitable conditions, some strong convergence theorems of the sequence generated by the algorithm to a zero of the sum of two monotone operators and fixed point of mappings are proved. At the end of the paper, some applications and the constructed algorithm are also discussed. The results presented in the paper extend and improve the main results of Montira et al. [10], Byrne et al. [11], Takahashi et al. [12] and Passty [13].
2 Preliminaries
Throughout this paper, we denote by \(\Bbb {N}\) the set of positive integers, and by \(\Bbb {R}\) the set of real numbers. Let H be a real Hilbert space with the inner product \(\langle \cdot,\cdot \rangle \) and norm \(\|\cdot \|\), respectively. When \(\{x_{n}\}\) is a sequence in H, we denote the weak convergence of \(\{x_{n}\}\) to x in H by \(x_{n}\rightharpoonup x\).
Let \(T:H\to H\) be a mapping. We say that T is a Lipschitz mapping if there exists an \(L > 0\) such that
The number L, associated with T, is called a Lipschitz constant. If \(L =1\),we say that T is a nonexpansive mapping, that is,
We say that T is firmly nonexpansive if
A mapping \(T:H\to H\) is said to be an averaged mapping if it can be written as the average of the identity I and a nonexpansive mapping, that is,
where \(\alpha \in (0,1)\) and \(S:H\to H\) is a nonexpansive mapping [14]. More precisely, when (2.1) holds, we say that T is α-averaged. It should be observed that a mapping is firmly nonexpansive if and only if it is a \(\frac{1}{2}\)-averaged mapping.
Let \(A:H\to H\) be a single-valued mapping. For a positive real number β, we say that A is β-inverse strongly monotone (β-ism) if
We now collect some important conclusions and properties, which will be needed in proving our main results.
Lemma 2.1
The following conclusions hold:
-
(i)
The composition of finitely many averaged mappings is averaged. In particular, if \(T_{i}\) is \(\alpha_{i}\)-averaged, where \(\alpha_{i} \in (0,1)\) for \(i =1,2\), then the composition \(T_{1}T_{2}\) is α-averaged, where \(\alpha =\alpha_{1}+\alpha_{2}-\alpha_{1} \alpha_{2}\).
-
(ii)
If A is β-ism and \(\gamma \in (0,\beta ]\), then \(T:= I-\gamma A\) is firmly nonexpansive.
-
(iii)
A mapping \(T:H\to H\) is nonexpansive if and only if \(I-T\) is \(\frac{1}{2}\)-ism.
-
(iv)
If A is β-ism, then, for \(\gamma >0\), γA is \(\frac{\beta }{\gamma }\)-ism.
-
(v)
T is averaged if and only if the complement \(I-T\) is β-ism for some \(\beta >\frac{1}{2}\). Indeed, for \(\alpha \in (0,1)\), T is α-averaged if and only if \(I-T\) is \(\frac{1}{2 \alpha }\)-ism.
Lemma 2.2
([17])
Let \(T=(1-\alpha)A+\alpha N\) for some \(\alpha \in (0,1)\). If A is β-averaged and N is nonexpansive then T is \(\alpha +(1-\alpha)\beta \)-averaged.
Let \(B:H\to 2^{H}\) be a set-valued mapping. The effective domain of B is denoted by \(D(B)\), that is, \(D(B)=\{x\in H:Bx\neq \emptyset \}\). Recall that B is said to be monotone if
A monotone mapping B is said to be maximal if its graph is not properly contained in the graph of any other monotone operator. For a maximal monotone operator \(B:H\to 2^{H}\) and \(r>0\), its resolvent \(J^{B}_{r}\) is defined by
It is well known that, if B is a maximal monotone operator and r is a positive number, then the resolvent \(J^{B}_{r}\) is single-valued and firmly nonexpansive, and \(F(J^{B}_{r})=B^{-1}0\equiv \{x\in H:0\in Bx \}\), \(\forall r>0\) (see [12, 18, 19]).
Lemma 2.3
([20])
Let H be a Hilbert space and let B be a maximal monotone operator on H. Then for all \(s,t >0 \) and \(x\in H\),
Lemma 2.4
([12])
Let \(H_{1}\) and \(H_{2}\) be Hilbert spaces. Let \(L:H_{1}\to H_{2}\) be a nonzero bounded linear operator and \(T:H_{2}\to H_{2}\) be a nonexpansive mapping. If \(B:H_{1}\to 2^{H_{1}}\) is a maximal monotone operator, then
-
(i)
\(L^{*}(I-T)L\) is \(\frac{1}{2\|L\|^{2}}\)-ism,
-
(ii)
For \(0< r<\frac{1}{\|L\|}\),
-
(iia)
\(I-rL^{*}(I-T)L\) is \(r\|L\|^{2}\)-averaged,
-
(iib)
\(J^{B}_{\lambda }(I-rL^{*}(I-T)L)\) is \(\frac{1+r\|L\|^{2}}{2}\)-averaged, for \(\lambda >0\),
-
(iii)
If \(r=\|L\|^{-2}\), then \(I-rL^{*}(I-T)L\) is nonexpansive.
Lemma 2.5
([21])
Let \(B:H\to 2^{H}\) be a maximal monotone operator with the resolvent \(J^{B}_{\lambda }=(I+\lambda B)^{-1}\) for \(\lambda >0\). Then we have the following resolvent identity:
for all \(\mu >0\) and \(x\in H\).
Lemma 2.6
([22])
Let C be a closed convex subset of a Hilbert space H and let T be a nonexpansive mapping of C into itself. Then \(U:=I-T\) is demiclosed, i.e., \(x_{n}\rightharpoonup x_{0}\) and \(Ux_{n}\to y_{0}\) imply \(Ux_{0}=y_{0}\).
Lemma 2.7
([10])
Let \(H_{1}\) and \(H_{2}\) be Hilbert spaces. Let \(A: H_{1}\to H_{1}\) be a β-ism, \(B: H_{1}\to 2^{H_{1}}\) a maximal monotone operator, \(T:H_{2}\to H_{2}\) a nonexpansive mapping and \(L: H_{1}\to H_{2}\) a bounded linear operator. If \(\Omega^{A+B}_{L,T} \neq \emptyset \), then the following are equivalent:
-
(i)
\(z\in \Omega^{A+B}_{L,T}\),
-
(ii)
\(z=J^{B}_{\lambda }((I_{\lambda }-A)-\gamma L^{*}(I-T)L)z\),
-
(iii)
\(0\in L^{*}(I-T)Lz+(A+B)z\),
where \(\lambda,\gamma >0\) and \(z\in H_{1}\).
Lemma 2.8
([23])
Let \(\{a_{n}\}\) be a sequence of nonnegative real numbers such that
where \(\{\beta_{n}\}\) is a sequence in \((0,1)\) and \(\{\delta_{n}\}\) is a sequence in \(\Bbb {R}\) such that
-
(i)
\(\sum^{\infty }_{n=1}\beta_{n}=\infty\);
-
(ii)
\(\limsup_{n\to \infty }\frac{\delta_{n}}{\beta_{n}}\leq 0\) or \(\sum^{\infty }_{n=1}|\delta_{n}|<\infty \).
Then \(\lim_{n\to \infty }a_{n}=0\).
3 Main results
We are now in a position to give the main result of this paper.
Lemma 3.1
Let \(H_{1}\) and \(H_{2}\) be two real Hilbert spaces. Let \(A: H_{1}\to H_{1}\) be a β-ism, \(B:H_{1} \to 2^{H_{1}}\) be a maximal monotone operator, \(T: H_{2}\to H_{2}\) be a nonexpansive mapping, and \(L:H_{1}\to H_{2}\) be a bounded linear operator. Let \(S: H_{1}\to H_{1}\) be a nonexpansive mapping such that \(F(S)\cap \Omega^{A+B}_{L,T}\neq \emptyset \), where
is the set of solutions of problem (1.3). Let \(f: H_{1}\to H_{1}\) be a contraction mapping with a contractive constant \(\alpha \in (0,1)\). For any \(t\in (0,1]\), let \(W_{t}: H_{1} \to H_{1}\) be the mapping defined by
where \(L^{*}\) is the adjoint of L and the sequences \(\lambda_{n}\) and \(\gamma_{n}\) satisfy the following control conditions:
-
(i)
\(0< a\leq \lambda_{n}\leq b_{1}<\frac{\beta }{2}\),
-
(ii)
\(0< a\leq \gamma_{n}\leq b_{2}<\frac{1}{2\|L\|^{2}}\), for some \(a,b_{1},b_{2}\in \Bbb {R}\).
Then \(W_{t}\) is a contraction mapping with a contractive constant \([1-t(1-\alpha)]\). Therefore \(W_{t}\) has a unique fixed point for each \(t \in (0, 1)\).
Proof
Note that, for each \(n\in \Bbb {N}\), we have
Also, by condition (i) and Lemma 2.1(ii), we know that \(I-2\lambda _{n}A\) is a firmly nonexpansive mapping, and this implies that \(I-2\lambda_{n}A\) must be a nonexpansive mapping. On the other hand, by Lemma 2.4(iia), we know that \(I-2\gamma_{n}L^{*}(I-T)L\) is \(2\gamma _{n}\|L\|^{2}\)-averaged. Thus, by condition (ii) and Lemma 2.2, we see that \((I-\lambda_{n}A)-\gamma_{n}L^{*}(I-T)L\) is \(\frac{1+2\gamma_{n} \|L\|^{2}}{2}\)-averaged.
Set
Since \(J^{B}_{\lambda_{n}}\) is \(\frac{1}{2}\)-averaged, by Lemma 2.1(i) we see that \(T_{n}\) is \(\frac{3+2\gamma_{n}\|L\|^{2}}{4}\)-averaged and hence it is nonexpansive. Further, for any \(x,y\in H_{1}\), we obtain
Since \(0<1-t(1-\alpha)<1\), it follows that \(W_{t}\) is a contraction mapping. Therefore, by Banach contraction principle, \(W_{t}\) has a unique fixed point \(x_{t}\) in \(H_{1}\). □
Theorem 3.2
Let \(H_{1}\), \(H_{2}\), A, B, T, L, S, f be the same as in Lemma 3.1. For any given \(x_{0}\in H_{1}\), let \(\{u_{n}\}\) and \(\{x_{n}\}\) be the sequences generated by
where \(\{\alpha_{n}\}\) is a sequence in \((0,1)\) such that \(\lim_{n\to \infty }\alpha_{n}=0,\sum^{\infty }_{n=0}\alpha_{n}=\infty \) and \(\sum^{\infty }_{n=1}|\alpha_{n}-\alpha_{n-1}|<\infty \) and \(L^{*}\) is the adjoint of L.
If \(F(S)\cap \Omega^{A+B}_{L,T}\neq \emptyset \) and the sequences \(\{\lambda_{n}\}\) and \(\{\gamma_{n}\}\) satisfy the following conditions:
-
(i)
\(0< a\leq \lambda_{n}\leq b_{1}<\frac{\beta }{2}\), and \(\Sigma^{\infty }_{n=1}|\lambda_{n}-\lambda_{n-1}|<\infty\),
-
(ii)
\(0< a\leq \gamma_{n}\leq b_{2}<\frac{1}{2\|L\|^{2}}\), and \(\Sigma^{\infty }_{n=1}|\gamma_{n}-\gamma_{n-1}|<\infty\), for some \(a, b_{1}, b_{2} \in \Bbb {R}\),
then the sequences \(\{u_{n}\}\) and \(\{x_{n}\}\) both converge strongly to \(z\in F(S)\cap \Omega^{A+B}_{L,T}\), where \(z=P_{F(S)\cap \Omega^{A+B}_{L,T}}f(z)\), i.e., z is a solution of problem (1.5).
Proof
Take
for each \(n\in \Bbb {N}\). By Lemma 2.7, we have \(\Omega^{A+B}_{L,T}=F(T _{n})\), for all \(n\in \Bbb {N}\). Thus, for each \(n\in \Bbb {N}\), we can write \(x_{n+1}=\alpha_{n}f(x_{n})+(1-\alpha_{n})ST_{n}x_{n}\). By the proof of Lemma 3.1, we see that \(T_{n}\) is \(\frac{3+2\gamma_{n}\|L\|^{2}}{4}\)-averaged. Thus, for each \(n\in \Bbb {N}\), we can write
where \(\xi_{n}=\frac{3+2\gamma_{n}\|L\|^{2}}{4}\) and \(V_{n}\) is a nonexpansive mapping. Consequently, we also have \(\Omega^{A+B}_{L,T}=F(T _{n})=F(V_{n})\), for all \(n\in \Bbb {N}\). Using this fact, for each \(p\in F(S)\cap \Omega^{A+B}_{L,T}\), we see that
for each \(n\in \Bbb {N}\). Since \(I-T_{n}=\xi_{n}(I-V_{n})\), in view of (3.5) we get
for each \(n\in \Bbb {N}\). Since \(\xi_{n}=\frac{3+2\gamma_{n}\|L\|^{2}}{4} \in (\frac{3}{4},1)\), we obtain
Next, we estimate
By induction, we can prove that
Hence \(\{x_{n}\}\) is bounded and so are \(\{u_{n}\}\), \(\{f(x_{n})\}\) and \(\{Su_{n}\}\).
Next, we show that
In fact, it follows from (3.4) that
where \(K:=\sup \{\|f(x_{n})\|+\|Su_{n}\|:n\in \Bbb {N}\}\).
Put
Since \(J^{B}_{\lambda_{n}}((I-\lambda_{n}A)-\gamma_{n}L^{*}(I-T)L)\) is nonexpansive, it follows from Lemma 2.3 that
where \(M_{1}\) and \(M_{2}\) are constants defined by
Therefore it follows from (3.10) and (3.11) that
Take
It follows from Lemma 2.8 that
Now, we write
Since \(\|x_{n+1}-x_{n}\|\to 0\) and \(\alpha_{n}\to 0\) as \(n\to \infty \), we obtain
Next, we prove that
In fact, it follows from (3.4) and (3.6) That
Hence, we obtain
Since \(\alpha_{n}\to 0\) as \(n\to \infty \), and \(\xi_{n}=\frac{3+2 \gamma_{n}\|L\|^{2}}{4}\in (\frac{3}{4},1)\), from (3.12) we obtain
Therefore we have
On the other hand, since \(\{x_{n}\}\) is bounded, let \(\{x_{n_{j}}\}\) be any subsequence of \(\{x_{n}\}\) with \(x_{n_{j}}\rightharpoonup \hat{x}\). Also, we assume that \(\lambda_{n_{j}}\rightarrow \hat{\lambda }\in (0,\frac{ \beta }{2})\) and \(\gamma_{n_{j}}\to \hat{\gamma }\in (0, \frac{1}{2\|L\|^{2}})\).
Letting
we know that T̂ is \(\frac{3+2\hat{\gamma }\|L\|^{2}}{4}\)-averaged and \(F(\hat{T})=\Omega^{A+B}_{L,T}\).
Hence, for each \(j\in \Bbb {N}\) we have
where \(z_{j}=((I-\lambda_{n_{j}}A)-\gamma_{n_{j}}L^{*}(I-T)L)x_{n_{j}}\). Now, we estimate the last term in (3.16). We have
for each \(j\in \Bbb {N}\). This implies that
Next, we estimate the second term in (3.16). By Lemma 2.5, we have
Also for each \(j\in \Bbb {N}\) we have
This shows that \(\{\|(J^{B}_{\lambda_{n_{j}}}z_{j}-z_{j})\|\}\) is a bounded sequence. This, together with (3.18), implies
Substituting (3.14), (3.17) and (3.19) into (3.16), we get
Thus, by Lemma 2.6, it follows that \(\hat{x}\in F(\hat{T})=\Omega^{A+B} _{L,T}\).
Furthermore, it follows from (3.13) and (3.14) that \(\{u_{n}\}\), \(\{x_{n}\}\) and \(\{S(u_{n})\}\) have the same asymptotical behavior, so \(\{u_{n}\}\) also converges weakly to x̂. Since S is nonexpansive, by (3.13) and Lemma 2.6, we obtain that \(\hat{x}\in F(S)\). Thus \(\hat{x}\in \Omega^{A+B}_{L,T}\cap F(S)\).
Next, we claim that
where \(z=P_{F(S)\cap \Omega^{A+B}_{L,T}}f(z)\).
Indeed, we have
since \(z=P_{F(S)\cap \Omega^{A+B}_{L,T}}f(z)\).
Finally, we show that \(x_{n}\to z\). Indeed, we have
which implies that
Now, by using (3.22) and Lemma 2.8, we deduce that \(x_{n}\to z\). Further it follows from \(\|u_{n}-x_{n}\|\to 0,u_{n}\rightharpoonup \hat{x} \in F(S)\cap \Omega^{A+B}_{L,T}\) and \(x_{n}\to z\) as \(n\to \infty \), that \(z=\hat{x}\). This completes the proof. □
If \(A:=0\), the zero operator, then the following result can be obtained from Theorem 3.2 immediately.
Corollary 3.3
Let \(H_{1}\) and \(H_{2}\) be Hilbert spaces. Let \(B : H_{1} \to 2^{H_{1}}\) be a maximal monotone operator, \(T : H_{2} \to H_{2}\) a nonexpansive mapping and \(L : H_{1}\to H_{2}\) a bounded linear operator. Let \(S: H_{1}\to H_{1}\) be a nonexpansive mapping such that \(\Gamma =F(S)\cap B^{-1}(0)\cap L^{-1}(F(T))\neq \emptyset \). Let \(f: H_{1}\to H_{1}\) be a contraction mapping with a contractive constant \(\alpha \in (0,1)\). For any given \(x_{0}\in H_{1}\), let \(\{u_{n}\}\) and \(\{x_{n}\}\) be the sequences generated by
If the sequences \(\{\alpha_{n}\}\), \(\{\lambda_{n}\}\) and \(\{\gamma _{n}\}\) satisfy all the conditions in Theorem 3.2, then the sequences \(\{u_{n}\}\) and \(\{x_{n}\}\) both converge strongly to \(z=P_{\Gamma }f(z)\) which is a solution of problem (1.5) with \(A =0\).
If \(H_{1}=H_{2}\), \(L=I\), then by applying Theorem 3.2, we can obtain the following result.
Corollary 3.4
Let \(H_{1}\) be Hilbert spaces. Let \(A : H_{1} \to H_{1}\) be a β-ism and \(B : H_{1} \to 2^{H_{1}}\) be a maximal monotone operator. Let \(S: H_{1}\to H_{1}\) be a nonexpansive mapping such that \(\Gamma_{1}=F(S)\cap (A+B)^{-1}0\cap F(T)\neq \emptyset \). Let \(f: H_{1}\to H_{1}\) be a contraction mapping with constant \(\alpha \in (0,1)\). For any \(x_{0}\in H_{1}\) arbitrarily, let the iterative sequences \(\{u_{n}\}\) and \(\{x_{n}\}\) be generated by
If the sequences \(\{\alpha_{n}\}\), \(\{\lambda_{n}\}\) and \(\{\gamma _{n}\}\) satisfy all the conditions in Theorem 3.2, then the sequences \(\{u_{n}\}\) and \(\{x_{n}\}\) both converge strongly to \(z\in \Gamma _{1}\), where \(z=P_{\Gamma_{1}} f(z)\).
4 Applications
In this section, we will utilize the results presented in the paper to study variational inequality problems, convex minimization problem and split common fixed point problem in Hilbert spaces.
4.1 Application to variational inequality problem
Let C be a nonempty closed and convex subset of a Hilbert space H. Recall that the normal cone to C at \(u\in C\) is defined by
It is well known that \(N_{C}\) is a maximal monotone operator. In the case \(B:=N_{C}:H\to 2^{H}\) we can verify that the problem of finding \(x^{*}\in H\) such that \(0\in Ax^{*}+Bx^{*}\) is reduced to the problem of finding \(x^{*}\in C\) such that
In the sequel, we denote by \(\operatorname{VIP}(C,A)\) the solution set of problem (4.1). In this case, we also have \(J^{B}_{\lambda }= P_{C}\) (the metric projection of H onto C). By the above consideration, problem (1.5) is reduced to finding
Therefore, the following convergence theorem can be immediately obtained from Theorem 3.2.
Theorem 4.1
Let \(H_{1}\) and \(H_{2}\) be Hilbert spaces. Let \(A : H_{1} \to H_{1}\) be a β-ism operator, \(T : H_{2}\to H_{2}\) a nonexpansive mapping and \(L : H_{1}\to H_{2}\) a bounded linear operator. Let \(S: H_{1}\to H_{1}\) be a nonexpansive mapping such that \(F(S)\cap \Omega^{A,C}_{L,T}\neq \emptyset \), where
Let \(f: H_{1}\to H_{1}\) be a contraction mapping with a contractive constant \(\alpha \in (0,1)\). For any given \(x_{0}\in H_{1}\), let the sequences \(\{u_{n}\}\) and \(\{x_{n}\}\) be generated by
where \(\{\alpha_{n}\}\) is a sequence in \((0,1)\) such that \(\lim_{n\to \infty }\alpha_{n}=0\), \(\sum^{\infty }_{n=0}\alpha_{n}=\infty\), \(\sum^{\infty }_{n=1}|\alpha_{n}-\alpha_{n-1}|<\infty\), \(L^{*}\) is the adjoint of L, and the sequences \(\{\lambda_{n}\}\) and \(\{\gamma_{n}\}\) satisfy conditions (i)–(ii) in Theorem 3.2. Then the sequences \(\{u_{n}\}\) and \(\{x_{n}\}\) both converge strongly to \(z=P_{F(S)\cap \Omega^{A,C}_{L,T}}f(z)\), which is a solution of problem (4.2).
4.2 Application to convex minimization problem
Let \(g:H\to R\) be a convex function, which is also Fréchet differentiable. Let C be a given closed convex subset of H. In this case, by setting \(A:=\nabla g\), the gradient of g, and \(B = N_{C}\), the problem of finding \(x^{*}\in (A+B)^{-1}0\) is equivalent to finding a point \(x^{*}\in C\) such that
Note that (4.4) is equivalent to the following minimization problem: find \(x^{*}\in C\) such that
Thus, in this situation, problem (1.5) is reduced to the problem of finding
Denote by
Then, by using Theorem 3.2, we can obtain the following result.
Theorem 4.2
Let \(H_{1}\) and \(H_{2}\) be Hilbert spaces and let C be a nonempty closed convex subset of \(H_{1}\). Let \(g: H_{1} \to \Bbb {R}\) be a convex and Fréchet differentiable function, ∇g be β-Lipschitz, \(T:H_{2}\to H_{2}\) be a nonexpansive mapping, and let \(L : H_{1}\to H_{2}\) be a bounded linear operator. Let \(S: H_{1}\to H_{1}\) be a nonexpansive mapping such that \(F(S)\cap \Omega^{g,C}_{L,T}\neq \emptyset \). Let \(f: H_{1}\to H_{1}\) be a contraction mapping with a contractive constant \(\alpha \in (0,1)\). For any given \(x_{0}\in H_{1}\), let \(\{u_{n}\}\) and \(\{x_{n}\}\) be the sequences generated by
If the sequences \(\{\alpha_{n}\}\), \(\{\lambda_{n}\}\) and \(\{\gamma _{n}\}\) satisfy all the conditions in Theorem 3.2, then the sequences \(\{u_{n}\}\) and \(\{x_{n}\}\) both converge strongly to \(z\in F(S) \cap \Omega^{g,C}_{L,T}\), where \(z=P_{F(S)\cap \Omega^{g,C}_{L,T}}f(z)\), which is a solution of problem (4.5).
Proof
Note that if \(g: H\to \Bbb {R}\) is convex and \(\nabla g:H \to H\) is β-Lipschitz continuous for \(\beta >0\) then ∇g is \(\frac{1}{\beta }\)-ism (see [24]). Thus, the required result can be obtained immediately from Theorem 3.2. □
4.3 Application to split common fixed point problem
Let \(V:H_{1}\to H_{1}\) be a nonexpansive mapping. Then, by Lemma 2.1(iii), we know that \(A:=I-V\) is \(\frac{1}{2}\)-ism. Furthermore, \(Ax^{*}=0\) if and only if \(x^{*}\in F(V)\). Hence problem (1.5) can be reduced to the problem of finding
where \(T:H_{2}\to H_{2}\), \(L:H_{1}\to H_{2}\) and \(S: H_{1} \to H_{1}\) are mappings as in Theorem 3.2.
This problem is called the split common fixed point problem (SCFP), and was studied by many authors (see [25–28], for example). By using Theorem 3.2, we can obtain the following result.
Theorem 4.3
Let \(H_{1}\) and \(H_{2}\) be Hilbert spaces. Let \(V:H_{1} \to H_{1}\) and \(T : H_{2}\to H_{2}\) be nonexpansive mappings and \(L : H_{1}\to H_{2}\) a bounded linear operator. Let \(S: H_{1} \to H_{1}\) be a nonexpansive mapping such that \(F(S)\cap \Omega ^{V}_{L,T}\neq \emptyset\), where
Let \(f: H_{1}\to H_{1}\) be a contraction mapping with a contractive constant \(\alpha \in (0,1)\). For any given \(x_{0}\in H_{1}\), let be \(\{u_{n}\}\) and \(\{x_{n}\}\) be the iterative sequences generated by
where the sequences \(\{\alpha_{n}\}\), \(\{\lambda_{n}\}\) and \(\{\gamma_{n}\}\) satisfy all the conditions in Theorem 3.2. Then the sequences \(\{u_{n}\}\) and \(\{x_{n}\}\) both converge strongly to a point \(z=P_{F(S)\cap \Omega^{V}_{L,T}}f(z)\), which is a solution of problem (4.7).
Proof
We consider \(B:=0\), the zero operator. The required result follows from the fact that the zero operator is monotone and continuous, hence it is maximal monotone. Moreover, in this case, we see that \(J^{B}_{\lambda }\) is the identity operator on \(H_{1}\), for each \(\lambda >0\). Thus algorithm (3.4) reduces to (4.8), by setting \(A:=I-V\) and \(B:=0\). □
References
Censor, Y., Elfving, T.: A multiprojection algorithm using Bregman projections in product space. Numer. Algorithms 8, 221–239 (1994)
Byrne, C.: Iterative oblique projection onto convex sets and the split feasibility problem. Inverse Probl. 18, 441–453 (2002)
Censor, Y., Bortfeld, T., Martin, B., Trofimov, A.: A unified approach for inversion problems in intensity-modulated radiation therapy. Phys. Med. Biol. 51, 2353–2365 (2006)
Martinet, B.: Régularisation dinéquations variationnelles par approximations successives. Rev. Fr. Inform. Rech. Opér. 3, 154–158 (1970)
Bruck, R.E., Reich, S.: Nonexpansive projections and resolvents of accretive operators in Banach spaces. Houst. J. Math. 3, 459–470 (1977)
Eckstein, J., Bertsckas, D.P.: On the Douglas–Rachford splitting method and the proximal point algorithm for maximal monotone operators. Math. Program. 55, 293–318 (1992)
Marino, G., Xu, H.K.: Convergence of generalized proximal point algorithm. Commun. Pure Appl. Anal. 3, 791–808 (2004)
Xu, H.K.: Iterative algorithms for nonlinear operators. J. Lond. Math. Soc. 66, 240–256 (2002)
Yao, Y., Noor, M.A.: On convergence criteria of generalized proximal point algorithms. J. Comput. Appl. Math. 217, 46–55 (2008)
Montira, S., Narin, P., Suthep, S.: Weak convergence theorems for split feasibility problems on zeros of the sum of monotone operators and fixed point sets in Hilbert spaces. Fixed Point Theory Appl. 2017, Article ID 6 (2017)
Byrne, C., Censor, Y., Gibali, A., Reich, S.: Weak and strong convergence of algorithms for the split common null point problem. J. Nonlinear Convex Anal. 13, 759–775 (2012)
Takahashi, W., Xu, H.K., Yao, J.C.: Iterative methods for generalized split feasibility problems in Hilbert spaces. Set-Valued Var. Anal. 23, 205–221 (2015)
Passty, G.B.: Ergodic convergence to a zero of the sum of monotone operators in Hilbert space. J. Math. Anal. Appl. 72, 383–390 (1979)
Baillon, J.B., Bruck, R.E., Reich, S.: On the asymptotic behavior of nonexpansive mappings and semigroups in Banach spaces. Houst. J. Math. 4, 1–9 (1978)
Boikanyo, O.A.: The viscosity approximation forward–backward splitting method for zeros of the sum of monotone operators. Abstr. Appl. Anal. 2016, Article ID 2371857 (2016)
Xu, H.K.: Averaged mappings and the gradient-projection algorithm. J. Optim. Theory Appl. 150, 360–378 (2011)
Byrne, C.: A unified treatment of some iterative algorithms in signal processing and image reconstruction. Inverse Probl. 20, 103–120 (2004)
Goebel, K., Reich, S.: Uniform Convexity, Hyperbolic Geometry, and Nonexpansive Mappings. Dekker, New York (1984)
Takahashi, W.: Introduction to Nonlinear and Convex Analysis. Yokohama Publishers, Yokohama (2009)
Takahashi, S., Takahashi, W., Toyoda, M.: Strong convergence theorems for maximal monotone operators with nonlinear mappings in Hilbert spaces. J. Optim. Theory Appl. 147, 27–41 (2010)
Barbu, V.: Nonlinear Semigroups and Differential Equations in Banach Spaces. Noordhoff, Leiden (1976)
Takahashi, W.: Nonlinear Functional Analysis: Fixed Point Theory and Its Applications. Yokohama Publishers, Yokohama (2000)
Xu, H.K.: Viscosity approximation methods for nonexpansive mapping. J. Math. Anal. Appl. 298, 279–291 (2004)
Baillon, J.B., Haddad, G.: Quelques propriétés des opérateurs angle-bornés et n-cycliquement monotones. Isr. J. Math. 26(2), 137–150 (1977)
Cui, H., Wang, F.: Iterative methods for the split common fixed point problem in Hilbert spaces. Fixed Point Theory Appl. 2014, Article ID 78 (2014)
Moudafi, A.: A note on the split common fixed-point problem for quasi-nonexpansive operators. Nonlinear Anal., Theory Methods Appl. 74, 4083–4087 (2011)
Shimizu, T., Takahashi, W.: Strong convergence to common fixed points of families of nonexpansive mappings. J. Math. Anal. Appl. 211, 71–83 (1997)
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, Article ID 438023 (2012)
Acknowledgements
The authors would like to express their thanks to the Editor and the Referees for their helpful comments.
Availability of data and materials
Not applicable.
Funding
The first author was supported by Scientific Research Fund of Sichuan Provincial Department of Science and Technology (2015JY0165), the second author was supported by Scientific Research Fund of Sichuan Provincial Education Department (16ZA0331) and the third author was supported by The Natural Science Foundation of China Medical University, Taichung, Taiwan.
Author information
Authors and Affiliations
Contributions
All authors contributed equally and significantly in writing this article. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
None of the authors have any competing interests in the manuscript.
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
Zhu, J., Tang, J. & Chang, Ss. Strong convergence theorems for a class of split feasibility problems and fixed point problem in Hilbert spaces. J Inequal Appl 2018, 289 (2018). https://doi.org/10.1186/s13660-018-1881-x
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s13660-018-1881-x