Open Access

Convergence theorems of a new iteration for two nonexpansive mappings

Journal of Inequalities and Applications20142014:82

https://doi.org/10.1186/1029-242X-2014-82

Received: 29 November 2013

Accepted: 7 February 2014

Published: 19 February 2014

Abstract

The purpose of this paper is to introduce the following new general implicit iteration scheme for approximating the common fixed points of a pair of nonexpansive mappings in a uniformly convex Banach space: for any x 0 C , the iterative process { x n } defined by x n = a n x n 1 + b n T y n + c n S x n , y n = a n x n 1 + b n x n + c n S x n 1 + d n T x n , where { a n } , { b n } , { c n } , { a n } , { b n } , { c n } , { d n } are seven sequences of real numbers satisfying a n + b n + c n = 1 , a n + b n + c n + d n = 1 , and T , S : C C are two nonexpansive mappings. We approximate the common fixed points of these two mappings by weak and strong convergence of the scheme.

Keywords

nonexpansive mapping common fixed points weak and strong convergence semicompact condition ( A )

1 Introduction

Let C be a nonempty subset of a real Banach space E. A mapping T of C into itself is called nonexpansive if T x T y x y holds for all x , y C . We first recall the following two iterative processes due to Ishikawa [1] and Mann [2], respectively.

(I) Let C be a nonempty convex subset of E and let T : C C be a mapping. For any given x 0 C the sequence { x n } defined by
x n + 1 = ( 1 a n ) x n + a n T y n , y n = ( 1 b n ) x n + b n T x n , n 0 ,
is called the Ishikawa iteration sequence, where { a n } and { b n } are two real sequences in [ 0 , 1 ] satisfying some conditions. In particular, if b n = 0 for all n 0 , then { x n } defined by
x 0 C , x n + 1 = ( 1 a n ) x n + a n T x n , n 0 ,

is called the Mann iteration sequence.

In [3], Liu introduced the concepts of Ishikawa and Mann iterative processes with errors as follows.

(II) For a nonempty subset C of a Banach space E and a mapping T : C C , the sequence { x n } defined by
x 0 C , x n + 1 = ( 1 a n ) x n + a n T y n + u n , y n = ( 1 b n ) x n + b n T x n + v n , n 0 ,

where { u n } and { v n } are two summable sequences in E. In particular, if b n = 0 , v n = 0 , the sequence { x n } is called the Mann iteration sequence with errors.

Unfortunately, the definitions of Liu, which depend on the convergence of the error terms, are against the randomness of errors. Xu [4] studied the following new iteration process.

(III) Let C be a nonempty convex subset of E and let T : C C be a mapping. For any given x 0 C the sequence { x n } defined by
x n + 1 = a n x n + b n T y n + c n u n , y n = a n x n + b n T x n + c n v n , n 0 ,

is called the Ishikawa iteration sequence with errors. Here { u n } and { v n } are two bounded sequences in C and { a n } , { b n } , { c n } , { a n } , { b n } , { c n } are six sequences in [ 0 , 1 ] satisfying the following conditions: a n + b n + c n = 1 , a n + b n + c n = 1 , n 1 . In particular, if b n = 0 , c n = 0 , the sequence { x n } is called the Mann iteration sequence with errors. Chidume and Moore [5] studied the above schemes in 1999.

A generalization of Mann and Ishikawa iterative schemes was given by Das and Debata [6] and Takahashi and Tamura [7]. This scheme dealt with two mappings: x 0 C
x n = a n x n 1 + ( 1 a n ) T y n 1 , y n 1 = b n x n 1 + ( 1 b n ) S x n 1 , n 1 .

Recently Khan and Fukhar [8] considered the above iterative process with bounded errors.

In the present paper, we consider the following scheme:
x 0 C , x n = a n x n 1 + b n T y n + c n S x n , y n = a n x n 1 + b n x n + c n S x n 1 + d n T x n , n 1 ,
(1)

where { a n } , { b n } , { c n } , { a n } , { b n } , { c n } , { d n } are seven sequences of real numbers in [ 0 , 1 ] satisfying a n + b n + c n = 1 , a n + b n + c n + d n = 1 , and T , S : C C are two nonexpansive mappings. Recently, some authors discuss similar issues (for example, please refer to [911]).

Approximating fixed points is an important subject in the theory of nonexpansive mappings and its applications in numerous applied areas. One is the convergence of iteration schemes constructed through nonexpansive mappings. In this paper, we study the iterative scheme given in (1) for weak and strong convergence for a pair of nonexpansive mappings in a uniformly convex Banach space. Before our discussions, we first recall the following definitions.

A Banach space E is said to satisfy Opial’s condition if whenever { x n } is a sequence in E which converges weakly to x, then
lim inf n x n x < lim inf n x n y for all  y E , y x .

It is well known that every Hilbert space satisfies the Opial condition (see for example [12]).

A mapping T is said to be semicompact (see, e.g., [13]) if for any sequence { x n } n = 1 in C such that lim n x n T x n = 0 , there exists a subsequence { x n j } j = 1 of { x n } n = 1 such that { x n j } j = 1 converges strongly to some u C .

A mapping T with domain D ( T ) and range R ( T ) in E is said to be demiclosed at a point p E if whenever { x n } is as sequence in D ( T ) such that { x n } converges weakly to x D ( T ) and { T x n } converges strongly to p, then T x = p .

We shall make use of the following results.

Lemma 1.1 [14]

Let { s n } , { t n } be two sequences of nonnegative real numbers satisfying
s n + 1 s n + t n n 1 .
  1. (a)

    If n = 1 t n < , then lim n s n exists.

     
  2. (b)

    If n = 1 t n < and { s n } has a subsequence converging to zero, then lim n s n = 0 .

     

Lemma 1.2 [15]

Let E be a uniformly convex Banach space satisfying Opial’s condition and let C be a nonempty closed convex subset of E. Let T be a nonexpansive mapping of C into itself. Then I T is demiclosed with respect to zero.

Lemma 1.3 [16]

Suppose that E is a uniformly convex Banach space and 0 < p t n q < 1 for all positive integers n. Also suppose that { x n } and { y n } are two sequences of E such that lim sup n x n r , lim sup n y n r and lim sup n t n x n + ( 1 t n ) y n = r hold for some r 0 . Then lim n x n y n = 0 .

2 Main results

We first give the following key lemma.

Lemma 2.1 Let E be a real uniformly convex Banach space and C its nonempty convex subset. Let T , S : C C be nonexpansive mappings. Let { x n } be the sequence as defined in (1) with the following conditions:
  1. (1)

    a n 0 , a n 0 , b n 0 , as n ;

     
  2. (2)

    b n , c n , c n , d n [ δ , 1 δ ] for some δ ( 0 , 1 ) ;

     
  3. (3)

    c n + d n γ for some γ ( 0 , 1 ) .

     
If F : = F ( T ) F ( S ) , then we have
  1. (i)

    lim n x n p exists for all p F and { x n } , { T x n } and { S x n } are all bounded;

     
  2. (ii)

    lim n x n T x n = 0 = lim n x n S x n .

     
Proof For any p F , we have
x n p = a n ( x n 1 p ) + b n ( T y n p ) + c n ( S x n p ) a n x n 1 p + b n T y n p + c n S x n p a n x n 1 p + b n y n p + c n x n p .
(2)
However, as the proof of the above inequality, it follows that
y n p = a n ( x n 1 p ) + b n ( x n p ) + c n ( S x n 1 p ) + d n ( T x n p ) a n x n 1 p + b n x n p + c n S x n 1 p + d n T x n p a n x n 1 p + b n x n p + c n x n 1 p + d n x n p .
(3)
Thus from (2) and (3), it is easy to check
x n p [ a n + b n ( a n + c n ) ] x n 1 p + [ c n + b n ( b n + d n ) ] x n p ,
(4)
which implies that
x n p x n 1 p ,
(5)
and the limit lim n x n p exists for all p F . Furthermore { x n } is bounded and { S x n } and { T x n } are both bounded also. Now suppose lim n x n p = c for some c 0 . By the inequalities (3) and (5), we have
lim sup n y n p c .
(6)
From the iterative process (1), we have
x n p = b n [ T y n p + a n ( x n 1 S x n ) ] + ( 1 b n ) [ S x n p + a n ( x n 1 S x n ) ] .
(7)
Since a n 0 , { x n } , { S x n } are both bounded and (6), it follows from Lemma 1.3 that
lim n T y n S x n = 0 .
(8)
Next, by the inequality (2) and (5), we have
x n p a n x n 1 p + b n y n p + c n x n p a n x n 1 p + b n y n p + c n x n 1 p = x n 1 p + b n ( y n p x n 1 p ) ,
(9)
which means
x n p x n 1 p b n + x n 1 p y n p .
Taking liminf on both sides in the above inequality and by (6) and b n [ δ , 1 δ ] for some δ ( 0 , 1 ) , we have
c lim inf n y n p lim sup n y n p c ,
which yields
lim n y n p = c .
(10)
It is easy to see that
y n p = c n [ S x n 1 p + a n ( x n 1 T x n ) + b n ( x n T x n ) ] + ( 1 c n ) [ T x n p + a n ( x n 1 T x n ) + b n ( x n T x n ) ] .
(11)
Since a n 0 , b n 0 , { S x n } and { T x n } are both bounded, then by Lemma 1.3 we have
lim n S x n 1 T x n = 0 .
(12)
Moreover, by the iterative process (1) again, we get
x n T y n a n x n 1 T y n + c n S x n T y n ,
which, by a n 0 as n and (8), implies that
lim n x n T y n = 0 .
(13)
Hence from (8) and (13), we have
x n S x n x n T y n + T y n S x n 0
(14)
as n . Note that
y n x n = a n ( x n 1 x n ) + c n ( S x n 1 x n ) + d n ( T x n x n ) a n x n 1 x n + c n S x n 1 T x n + c n T x n x n + d n T x n x n ;
(15)
then it follows from the above inequality that
x n T x n x n S x n + S x n T y n + T y n T x n x n S x n + S x n T y n + y n x n x n S x n + S x n T y n + a n x n 1 x n + c n S x n 1 T x n + c n T x n x n + d n T x n x n .
(16)
That is to say,
x n T x n 1 1 γ { x n S x n + S x n T y n + a n x n 1 x n + c n S x n 1 T x n } ,
(17)
which, from (8), (12), (14), and a n 0 , means that
lim n x n T x n = 0 .
(18)

The proof of the lemma is completed. □

Now we give the weak convergence first.

Theorem 2.2 Let E be a uniformly convex Banach space satisfying Opial’s condition and C, S, T, and { x n } be as taken in Lemma  2.1. If F : = F ( T ) F ( S ) , then { x n } converges weakly to a common fixed point of S and T.

Proof Let p F . As the proof of Lemma 2.1, lim n x n p exists. Since E is uniformly convex, every bounded subset of E is weakly compact, so that there exists a subsequence { x n k } of the bounded sequence { x n } such that { x n k } converges weakly to a point q C . Therefore, it follows from (ii) in Lemma 2.1 that
lim k T x n k x n k = 0 .

By Lemma 1.2, we know I T is demiclosed, then it is easy to see that q F ( T ) . With a similar proof, it follows that q F ( S ) also. Next we prove uniqueness. Suppose that this is not true, then there must exist a subsequence { x n j } { x n } such that { x n j } converges weakly to another q C and q q . Then by the same method given above, we can also prove q F ( T ) F ( S ) .

Because we have proved that for any p F , the limit lim n x n p exists, we can let
lim n x n q = d 1 , lim n x n q = d 2 .
By the Opial condition of E, we have
d 1 = lim sup k x n k q < lim sup k x n k q = d 2 = lim sup j x n j q < lim sup j x n j q = d 1 .

This is a contradiction, hence q = q . This implies that { x n } converges weakly to a common fixed point of S and T. □

Next we give several strong convergence results.

Theorem 2.3 Let E be a uniformly convex Banach space and C, { x n } be as taken in Lemma  2.1. If one of the nonexpansive mappings T and S is semicompact and F : = F ( T ) F ( S ) , then { x n } converges strongly to a common fixed point of S and T.

Proof Since one of T and S is semicompact and by (ii) in Lemma 2.1, then there exists a subsequence { x n j } j = 1 of the sequence { x n } n = 1 such that { x n j } j = 1 converges strongly to u. Since C is closed, u C . Continuity of S and T gives S x n j S u 0 and T x n j T u 0 as n j . Then by Lemma 2.1,
S u u = 0 = T u u .

This yields u F . By Lemma 2.1 again, lim n x n p exists for all p F , therefore { x n } must itself converge to u F . This completes the proof. □

Recall that a mapping T : C C , where C is a subset of E, is said to satisfy condition (A) if there exists a nondecreasing function f : [ 0 , ) [ 0 , ) with f ( 0 ) = 0 , f ( r ) > 0 for all r ( 0 , ) such that x T x f ( d ( x , F ( T ) ) ) for all x C where d ( x , F ( T ) ) = inf { x x : x F ( T ) } .

In [17], Senter and Dotson approximated fixed points of a nonexpansive mapping T by Mann iteration. Recently Maiti and Ghosh [18] and Tan and Xu [14] considered the approximation of fixed points of a nonexpansive mapping T by Ishikawa iteration under the same condition (A) which is weaker than the requirement that T is semicompact. Khan and Fukhar [8] modified this condition for two mappings S , T : C C as follows.

Let C be a subset of a Banach space E. Two mappings S , T : C C are said to satisfy condition ( A ) if there exists a nondecreasing function f : [ 0 , ) [ 0 , ) with f ( 0 ) = 0 , f ( r ) > 0 for all r ( 0 , ) such that 1 2 ( x T x + x S x ) f ( d ( x , F ) ) for all x K , where d ( x , F ) = inf { x x : x F = F ( T ) F ( S ) } .

Note that condition ( A ) reduces to condition (A) when S = T . We use condition ( A ) to study the strong convergence of { x n } defined in (1).

Theorem 2.4 Let E be a uniformly convex Banach space and C, { x n } be as taken in Lemma  2.1. Let T, S satisfy the condition ( A ) and F : = F ( T ) F ( S ) , then { x n } converges strongly to a common fixed point of S and T.

Proof By Lemma 2.1, lim n x n x exists for all x F = F ( T ) F ( S ) . Let it be c for some c 0 . If c = 0 , there is nothing to prove. Suppose c > 0 . By Lemma 2.1, lim n x n T x n = 0 = lim n x n S x n . Moreover, from (4), we have x n x x n 1 x which gives
inf x F x n x inf x F x n x .

That is to say, d ( x n , F ) d ( x n 1 , F ) shows that lim n d ( x n , F ) exists by virtue of Lemma 1.1. Now by condition ( A ), lim n f ( d ( x n , F ) ) = 0 . By the properties of f, therefore lim n d ( x n , F ) = 0 . Next we can take a subsequence { x n j } of { x n } and { y j } F such that x n j y j < 2 j . Then following the method of proof of Tan and Xu [14], we find that { y j } is a Cauchy sequence in F and so it converges. Let y j y . Since F is closed, therefore y F and then x n j y . As lim n x n x exists, x n y F . This completes the proof. □

Remark 2.5 From the proof of the above results, it is easy to see that we can extend our theorems to the iterative process (1) with errors as follows: let C be a bounded closed convex subset of E and the sequence { x n } be defined by
x 0 C , x n = a n x n 1 + b n T y n + c n S x n + e n u n , y n = a n x n 1 + b n x n + c n S x n 1 + d n T x n + e n v n
(19)

where { a n } , { b n } , { c n } , { e n } , { a n } , { b n } , { c n } , { d n } , { e n } , are sequences in [ 0 , 1 ] with a n + b n + c n + e n = 1 , a n + b n + c n + d n + e n = 1 and T , S : C C are both nonexpansive mappings, { u n } , { v n } C .

Declarations

Acknowledgements

The authors are very grateful to the referees for their critical and valuable comments, which allowed us to improve the presentation of this article.

Authors’ Affiliations

(1)
College of Mathematics and Information Science, Henan Normal University

References

  1. Ishikawa S: Fixed point and iteration of a nonexpansive mapping in a Banach spaces. Proc. Am. Math. Soc. 1976, 73: 65–71.MathSciNetView ArticleGoogle Scholar
  2. Mann WR: Mean value methods in iteration. Proc. Am. Math. Soc. 1953, 4: 506–510. 10.1090/S0002-9939-1953-0054846-3View ArticleGoogle Scholar
  3. Liu LS: Ishikawa and Mann iterative process with errors for nonlinear strongly accretive mappings in Banach spaces. J. Math. Anal. Appl. 1995, 194: 114–125. 10.1006/jmaa.1995.1289MathSciNetView ArticleGoogle Scholar
  4. Xu YG: Ishikawa and Mann iteration process with errors for nonlinear strongly accretive operator equations. J. Math. Anal. Appl. 1998, 224: 91–101. 10.1006/jmaa.1998.5987MathSciNetView ArticleGoogle Scholar
  5. Chidume CE, Moore C: Fixed point iteration for pseudocontractive maps. Proc. Am. Math. Soc. 1999,127(4):1163–1170. 10.1090/S0002-9939-99-05050-9MathSciNetView ArticleGoogle Scholar
  6. Das G, Debata JP: Fixed points of quasi-nonexpansive mappings. Indian J. Pure Appl. Math. 1986, 17: 1263–1269.MathSciNetGoogle Scholar
  7. Takahashi W, Tamura T: Convergence theorems for a pair of nonexpansive mappings. J. Convex Anal. 1998,5(1):45–58.MathSciNetGoogle Scholar
  8. Khan SH, Fukhar-ud-din H: Weak and strong convergence of a scheme with errors for two nonexpansive mappings. Proc. Am. Math. Soc. 2005,61(8):1295–1301.MathSciNetGoogle Scholar
  9. Kim GE: Weak and strong convergence for quasi-nonexpansive mappings in Banach spaces. Bull. Korean Math. Soc. 2012,49(4):799–813. 10.4134/BKMS.2012.49.4.799MathSciNetView ArticleGoogle Scholar
  10. Zhang F, Zhang H, Zhang YL: New iterative algorithm for two infinite families of multivalued quasi-nonexpansive mappings in uniformly convex Banach spaces. J. Appl. Math. 2013., 2013: Article ID 649537Google Scholar
  11. Zhang JL, Su YF, Cheng QQ: Strong convergence theorems for a common fixed point of two countable families of relatively quasi nonexpansive mappings and applications. Abstr. Appl. Anal. 2012., 2012: Article ID 956950Google Scholar
  12. Opial Z: Weak convergence of successive approximations for nonexpansive mappings. Bull. Am. Math. Soc. 1967, 73: 591–597. 10.1090/S0002-9904-1967-11761-0MathSciNetView ArticleGoogle Scholar
  13. Sun ZH: Strong convergence of an implicit iteration process for a finite family of asymptotically quasinonexpansive mappings. J. Math. Anal. Appl. 2003, 286: 351–358. 10.1016/S0022-247X(03)00537-7MathSciNetView ArticleGoogle Scholar
  14. Tan KK, Xu HK: Approximating fixed points of nonexpansive mappings by the Ishikawa iteration process. J. Math. Anal. Appl. 1993, 178: 301–308. 10.1006/jmaa.1993.1309MathSciNetView ArticleGoogle Scholar
  15. Browder FE: Nonlinear operators and nonlinear equations of evolution in Banach spaces. 18. Proceedings of the Symposium on Pure Mathematics 1976.Google Scholar
  16. Schu J: Weak and strong convergence to fixed points of asymptotically nonexpansive mappings. Bull. Aust. Math. Soc. 1991, 43: 153–159. 10.1017/S0004972700028884MathSciNetView ArticleGoogle Scholar
  17. Senter HF, Dotson WG: Approximating fixed points of nonexpansive mappings. Proc. Am. Math. Soc. 1974,44(2):375–380. 10.1090/S0002-9939-1974-0346608-8MathSciNetView ArticleGoogle Scholar
  18. Maiti M, Gosh MK: Approximating fixed points by Ishikawa iterates. Bull. Aust. Math. Soc. 1989, 40: 113–117. 10.1017/S0004972700003555View ArticleGoogle Scholar

Copyright

© Hou and Du; licensee Springer. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.