Preconditioning methods for solving a general split feasibility problem
Journal of Inequalities and Applications volume 2014, Article number: 435 (2014)
We introduce and study a new general split feasibility problem (GSFP) in a Hilbert space. This problem generalizes the split feasibility problem (SFP). The GSFP extends the SFP with a nonlinear continuous operator. We apply the preconditioning methods to increase the efficiency of the CQ algorithm, two general preconditioning CQ algorithms for solving the GSFP are presented. We also propose a new inexact method to approximate the preconditioner. The convergence theorems are established under the projections with respect to special norms. Some numerical results illustrate the efficiency of the proposed methods.
MSC:46E20, 47J20, 47J25, 47L50.
As preconditioning methods can improve the condition number of the ill-posed system matrix, the convergence rate of the iterative algorithm can also be improved . In [2, 3], a preconditioning method is applied to modify the projected Landweber algorithm for solving a linear feasibility problem (LFP). The modified algorithm is
where , is a linear and continuous operator, means 2-norm, X and Y are Hilbert spaces and is the datum of the problem, corrupted by noise or experimental errors.
where is the projection operator onto S with respect to the norm . Bertsekas and Gafni  and Marcotte and Wu  improved it with variable symmetric positive defined matrices , Fukushima  modified it by a relaxed projection method with half-space; then in , Yang established the convergence of Auslender’s algorithm under the weak co-coercivity of F.
where K is a nonempty closed convex set in , are nonlinear operators. In , Santos and Scheimberg extended and applied (1.1) to solve the GVI.
However, a general split feasibility problem (GSFP) equals to a GVI, and preconditioning methods for solving the GSFP have not been studied. By introducing a convex minimization problem, the split feasibility problem (SFP) is equivalent to a variational inequality problem (VIP), which involves a Lipschitz continuous and inversely strong monotone (ism) operator, see [14–22]. Similarly, by the same way, in this paper we introduce that a new GSFP equals to a GVI involving a Lipschitz continuous and co-coercive operator [9, 17–19].
Otherwise, Mohammad and Abdul  considered a general split feasibility in infinite-dimensional real Hilbert spaces. It is to find such that
where is a bounded linear operator, and are the families of nonempty closed convex subsets of and , respectively.
Let C and Q be nonempty closed convex subsets in real Hilbert spaces and , respectively. We consider a general split feasibility problem which is different from the one in . Our GSFP is to find
where is a bounded linear operator and is a continuous operator. We see that the SFP in  and the GSFP in  are particular cases of GSFP (1.2). It has applications in many special fields such as signal decryption, demodulating the digital signal and noise processing, etc. In order to solve GSFP (1.2), two preconditioning algorithms are developed in this paper following the iterative scheme
where the two general constraints C and Q deal with projections with respect to the norms corresponding to some symmetric positive definite matrices. Define the solution set of (1.2) , as Γ is nonempty, and by virtue of the related G-co-coercive operator, we can establish the convergence of the proposed algorithms.
The paper is organized as follows. Section 2 presents two useful propositions. In Section 3, we define the algorithms with fixed preconditioner, variable preconditioner, relaxed projection and preconditioner approximation and analyze the convergence. Numerical results are reported in Section 4. Finally, Section 5 gives some concluding remarks.
In what follows, we state some concepts and propositions.
The SFP is to find a point such that , where is a bounded linear operator .
Let G be a symmetric positive definite matrix, and set . Then the norm is defined by for . We denote by the projection operator onto C with respect to the norm , i.e.,
Let C be a nonempty closed convex subset in H, for and , the G-projection operator onto C has the following properties:
Let be a symmetric positive definite matrix, and . Then the norm is defined by for . We denote by the projection operator onto Q with respect to the norm . According to the SFP, when GSFP (1.2) has no solution (refer to [2, 14, 15]), we can define
is also convex and continuously differentiable in H. Its gradient operator is
As , we define
is also Lipschitz continuous.
Proposition 2.2 If we consider the constrained minimization problem
its stationary point satisfies
which is a general variational inequality involving a Lipschitz continuous and G-co-coercive operator.
Proof For , from (2.1) and Lemma 8.1 in , we have
where L is the largest eigenvalue of ; therefore, the operator is Lipschitz continuous,
Thus, the operator is co-coercive. □
3 Main results
In this section, we propose several modified CQ algorithms with preconditioning techniques and prove the convergence.
3.1 General preconditioning CQ algorithm
In this part, we have our first algorithm with fixed stepsize and preconditioner to solve GSFP (1.2). The algorithm is as follows.
Algorithm 3.1 Choose such that , and let such that , then we calculate such that
where , L and are the largest eigenvalues of and D, respectively.
Now we establish the weak convergence of Algorithm 3.1.
Theorem 3.1 Suppose that the operators and are continuous. If , then the sequence generated by Algorithm 3.1 converges to the solution of GSFP (1.2).
Proof Firstly, for , we have
From (3.1), (iii), (ii) and the definition of ism, we have
as , which implies that the sequence is monotonically decreasing, then we can obtain that the sequence is also convergent, especially, the sequence is bounded. Consequently, we get from (3.2)
Moreover, for each , from (iii) and (2.1), we have
Then by virtue of (3.3) we have
Hence, there exists a subsequence of such that
Thus, is also bounded.
Let be an accumulation point of , then the subsequence of , as . Because of the continuity of g, there exists an accumulation point of the sequence ; for the subsequence , we have as . After that, from (3.3) we obtain
that is, .
We use in place of in (3.2) and obtain that is convergent. Because its subsequence , then we get that converges to as . As well as is continuous, we finally have
Therefore, is a solution of GSFP (1.2). □
From Algorithm 3.1 and Theorem 3.1 we can deduce the following results easily.
Corollary 3.1 If , then GSFP (1.2) reduces to SFP, Algorithm 3.1 also reduces to a preconditioning CQ (PCQ) algorithm for :
where , L and are the largest eigenvalues of and D, respectively. and are still the projection operators onto C and Q with respect to the norm and , respectively.
Corollary 3.2 If , , , then GSFP reduces to SFP, then Algorithm 3.1 reduces to the CQ algorithm proposed in .
Corollary 3.3 If , , and are the projections onto C and Q with respect to the norm , set , (3.5) transforms into the algorithm in 
where , . Then the GSFP reduces to the extended split feasibility problem (ESFP) in .
3.2 An algorithm with variable projection metric
The algorithms above can speed the convergence of CQ algorithm, but the stepsize and preconditioner are fixed. In this subsection, we extend the results in  and construct an iterative scheme with variable stepsize and preconditioner from one iteration to the next. As a key role, will change arbitrarily or following some rules to achieve the convergence progress and better results.
Let and be two symmetric positive definite matrices for . Denote by and the projections onto C and Q with respect to the norm and , respectively. Let χ be a set of symmetric positive definite matrices, we have the following algorithm.
Algorithm 3.2 Choose such that , and let such that ; for , we compute such that
where , L is the largest eigenvalue of , is the minimum value of all the largest eigenvalue values to matrices .
Remark 3.1 Define , then for the next iteration, is either chosen arbitrarily from χ or equivalent to . It is conditional on whether has decreased or not. More particularly, we define a scalar with initial value . Having chosen a scalar at the n th iteration, then is calculated by
then we select
Theorem 3.2 If , then the sequence generated by Algorithm 3.2 converges to the solution of GSFP (1.2).
Proof To obtain variable at each iteration and keep the convergence of Algorithm 3.2, must be a descending behavior for . We first show that
Indeed if (3.7) is not true, we have , then must have changed a finite number of times, we set that this number is . Therefore, let be a solution of GSFP, refer to (3.2) and for , we have
as , . Then, following the proof of Theorem 3.1, we also get
Equation (3.9) contradicts the above hypothesis, so (3.7) is true.
By using (3.6) and (iii), for the n th iteration, we have
where . Then from (2.1) and (3.8) we know that
so . By virtue of (3.7), we also have
This means that at least a subsequence of converges to a solution of . Similar to the argumentation of accumulation in the proof of Theorem 3.1, we know that converges to a solution of GSFP. □
3.3 Some methods for execution
In Algorithms 3.1 and 3.2, there still exists difficulty to implement the projections and with respect to the defined norms, especially when C and Q are general closed convex sets. According to the relaxed method in [8, 27, 28], we consider the above algorithm in which the closed convex subsets C and Q are the following particular formula:
where and are convex functions. and are given as
where , .
Here, we also replace and by and . However, in this paper, take Algorithm 3.2 for example, the projections are with respect to the norms corresponding to and , we should use the following methods to calculate them. For and ,
Set , , let be an accumulation of . From the proof above, it is easy to deduce that
Therefore, , , with the projections and , is a solution of GSFP.
Next, we present a new approximation method to estimate and in Algorithm 3.2.
If , for and , such that
under the ideal condition, if , the solution is done, but unfortunately, cannot be calculated directly when A is a large matrix in practice. As
where λ is an eigenvalue of . Let , for the n th iteration, we have the next approximation of
where . So, at the n th iteration, let be the minimum eigenvalue of , take , , the variable stepsize is approximated by
4 Numerical results
We consider the following problem from  in a finite dimensional Hilbert space:
Let , where , and , where . is a random matrix where every element of A is in satisfying . Let be a random vector in where every element of is in .
We set as the stop rule, and let , , , , for . Using the methods in Section 3.3, we compare Algorithm 3.2 with the relaxed CQ algorithm (RCQ) in , with different ε and initial values. The results can be seen in Table 1. We see that the proposed methods in this paper behave better.
5 Concluding remarks
In this paper, we have discussed a new general split feasibility problem, which is related to the general variational inequalities involving a co-coercive operator. By using the G-norm method, variable modulus method and relaxed method, two modified projection algorithms for solving the GSFP and some approximate methods for algorithm executing have been presented. The numerical results show that by preconditioning method, the convergence speed of CQ algorithm can be improved, but the way to obtain variable stepsize in the paper is inexact. To continue to improve it or combine it with the methods in  and  is another interesting subject.
Chen K: Matrix Preconditioning Techniques and Applications. Cambridge University Press, New York; 2005.
Piana M, Bertero M: Projected Landweber method and preconditioning. Inverse Probl. 1997, 13: 441–463. 10.1088/0266-5611/13/2/016
Strand ON: Theory and methods related to the singular-function expansion and Landweber’s iteration for integral equations of the first kind. SIAM J. Numer. Anal. 1974, 11: 798–824. 10.1137/0711066
Auslender A: Optimisation: Méthodes Numérique. Masson, Paris; 1976.
Dafermos S: Traffic equilibrium and variational inequalities. Transp. Sci. 1980, 14: 42–54. 10.1287/trsc.14.1.42
Bertsekas PD, Gafni EM: Projection methods for variational inequities with application to the traffic assignment problem. Math. Program. Stud. 1982, 17: 139–159. 10.1007/BFb0120965
Marcotte P, Wu JH: On the convergence of projection methods: application to the decomposition of affine variational inequalities. J. Optim. Theory Appl. 1995, 85: 347–362. 10.1007/BF02192231
Fukushima M: A relaxed projection method for variational inequalities. Math. Program. 1986, 35: 58–70. 10.1007/BF01589441
Yang Q: The revisit of a projection algorithm with variable steps for variational inequalities. J. Ind. Manag. Optim. 2005, 1: 211–217.
He BS: Inexact implicit methods for monotone general variational inequalities. Math. Program. 1999, 86: 199–217. 10.1007/s101070050086
Noor MA, Wang YJ, Xiu N: Projection iterative schemes for general variational inequalities. J. Inequal. Pure Appl. Math. 2002., 3: Article ID 34
Santos PSM, Scheimberg S: A projection algorithm for general variational inequalities with perturbed constraint sets. Appl. Math. Comput. 2006, 181: 649–661. 10.1016/j.amc.2006.01.050
Muhammad AN, Abdellah B, Saleem U: Self-adaptive methods for general variational inequalities. Nonlinear Anal. 2009, 71: 3728–3738. 10.1016/j.na.2009.02.033
Qu B, Xiu N: A note on the CQ algorithm for the split feasibility problem. Inverse Probl. 2005, 21: 1655–1665. 10.1088/0266-5611/21/5/009
Byrne C: A unified treatment of some iterative algorithms in signal processing and image reconstruction. Inverse Probl. 2004, 20: 103–120. 10.1088/0266-5611/20/1/006
Dolidze Z: Solution of variational inequalities associated with a class of monotone maps. Èkon. Mat. Metody 1982, 18: 925–927.
He B, He X, Liu H, Wu T: Self-adaptive projection method for co-coercive variational inequalities. Eur. J. Oper. Res. 2009, 196: 43–48. 10.1016/j.ejor.2008.03.004
Facchinei F, Pang JS I. In Finite-Dimensional Variational Inequality and Complementarity Problems. Springer, New York; 2003.
Facchinei F, Pang JS II. In Finite-Dimensional Variational Inequality and Complementarity Problems. Springer, New York; 2003.
Yao YH, Postolache M, Liou YC: Strong convergence of a self-adaptive method for the split feasibility problem. Fixed Point Theory Appl. 2013., 2013: Article ID 201
Yao YH, Yang PX, Kang SM: Composite projection algorithms for the split feasibility problem. Math. Comput. Model. 2013, 57: 693–700. 10.1016/j.mcm.2012.07.026
Yao YH, Liou YC, Shahzad N: A strongly convergent method for the split feasibility problem. Abstr. Appl. Anal. 2012., 2012: Article ID 125046
Mohammad E, Abdul L: General split feasibility problems in Hilbert spaces. Abstr. Appl. Anal. 2013., 2013: Article ID 805104
Censor Y, Elfving T: A multiprojection algorithm using Bregman projections in a product space. Numer. Algorithms 1994, 8: 221–239. 10.1007/BF02142692
Byrne C: Iterative oblique projection onto convex sets and the split feasibility problem. Inverse Probl. 2002, 18: 441–453. 10.1088/0266-5611/18/2/310
Wang PY, Zhou HY: A preconditioning method of the CQ algorithm for solving the extended split feasibility problem. J. Inequal. Appl. 2014., 2014: Article ID 163 10.1186/1029-242X-2014-163
Yang Q: On variable-step relaxed projection algorithm for variational inequalities. J. Math. Anal. Appl. 2005, 302: 166–179. 10.1016/j.jmaa.2004.07.048
López G, Martín-Márquez M, Wang F, Xu H-K: Solving the split feasibility problem without prior knowledge of matrix norms. Inverse Probl. 2012., 28: Article ID 085004
Wang Z, Yang Q, Yang Y: The relaxed inexact projection methods for the split feasibility problem. Appl. Math. Comput. 2011, 217: 5347–5359. 10.1016/j.amc.2010.11.058
Yang Q: The relaxed CQ algorithm solving the split feasibility problem. Inverse Probl. 2004, 20: 1261–1266. 10.1088/0266-5611/20/4/014
The authors would like to thank the associate editor and the referees for their comments and suggestions. This research was supported by the National Natural Science Foundation of China (11071053).
The authors declare that they have no competing interests.
The authors take equal roles in deriving results and writing of this paper. All authors read and approved the final manuscript.
About this article
Cite this article
Wang, P., Zhou, H. & Zhou, Y. Preconditioning methods for solving a general split feasibility problem. J Inequal Appl 2014, 435 (2014). https://doi.org/10.1186/1029-242X-2014-435