Next Article in Journal
Analysis and Applications of Bonferroni Mean Operators and TOPSIS Method in Complete Cubic Intuitionistic Complex Fuzzy Information Systems
Next Article in Special Issue
2D and 3D Visualization for the Static Bifurcations and Nonlinear Oscillations of a Self-Excited System with Time-Delayed Controller
Previous Article in Journal
Fast Treetops Counting Using Mathematical Image Symmetry, Segmentation, and Fast k-Means Classification Algorithms
Previous Article in Special Issue
A Modified Krasnosel’skiǐ–Mann Iterative Algorithm for Approximating Fixed Points of Enriched Nonexpansive Mappings
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Two Generic Convergence Results for Infinite Products of Generalized Nonexpansive Mappings

by
Simeon Reich
* and
Alexander J. Zaslavski
*
Department of Mathematics, The Technion—Israel Institute of Technology, Haifa 32000, Israel
*
Authors to whom correspondence should be addressed.
Symmetry 2022, 14(3), 534; https://doi.org/10.3390/sym14030534
Submission received: 7 February 2022 / Revised: 23 February 2022 / Accepted: 1 March 2022 / Published: 5 March 2022
(This article belongs to the Special Issue Nonlinear Analysis and Its Applications in Symmetry)

Abstract

:
In our 2014 work with M. Gabour, we introduced a metric space of generalized nonexpansive self-mappings of bounded and closed subsets of a Banach space and studied, using the Baire category approach, the asymptotic behavior of iterates of a generic operator belonging to this class. In the definition of a generalized nonexpansive mapping the norm is replaced by a general function which can be symmetric as a particular case. In this paper, we prove the convergence of infinite products of generalized nonexpansive self-mappings to a common fixed point in a generic setting.

1. Introduction

Nonexpansive mappings (in other words, those Lipschitz mappings the Lipschitz constant of which is equal to one) have been studied intensively in recent decades. See, for example, refs. [1,2,3,4,5,6,7,8,9,10] and the references cited therein. This research activity has its origins in Banach’s classical theorem [11] regarding the existence of a unique fixed point for strict contractions (that is, those mappings the Lipschitz constant of which is strictly less than one). It also concerns the asymptotic behavior of both exact and inexact iterates of a nonexpansive mapping, and their possible convergence. These theoretical investigations have found significant and diverse applications in the engineering, medical and natural sciences [9,10,12,13,14,15]. In this vein, we introduced and studied in [16] a certain class of nonlinear operators, which we now proceed to describe.
Suppose that K X is a bounded, closed and convex set in a Banach space ( X , · ) , and that a continuous function f : X [ 0 , ) satisfies f ( 0 ) = 0 ,
sup { | f ( z ) | : z K K } < ,
and possesses the following properties:
(P1) for every ϵ > 0 , there exists δ > 0 such that if z 1 , z 2 K satisfy f ( z 1 z 2 ) δ , then the inequality z 1 z 2 ϵ is true;
(P2) for every γ ( 0 , 1 ) , there exists ϕ ( γ ) ( 0 , 1 ) for which
f ( γ ( z 1 z 2 ) ) ϕ ( γ ) f ( z 1 z 2 )   for   each   pair   of   points   z 1 , z 2 K ;
(P3) the function ( ξ 1 , ξ 2 ) f ( ξ 1 ξ 2 ) , ξ 1 , ξ 2 K , is uniformly continuous on K × K .
Set
diam ( K ) : = sup { z 1 z 2 : z 1 , z 2 K } .
Denote by B the collection of all continuous operators A : K K which satisfy
f ( A z 1 A z 2 ) f ( z 1 z 2 )   for   every   pair   of   points   z 1 , z 2 K .
For each pair of operators T , S B , define
d ( T , S ) : = sup { T z S z : z K } .
Clearly, ( B , d ) is a complete metric space.
Using the Baire category approach, we showed in [16] that a typical (generic) mapping in the space B has a unique fixed point which attracts all its iterates. It should be mentioned that the classical theorem of De Blasi and Myjak [17] regarding nonexpansive mappings is a special case of this result, where the continuous function f = · . As a matter of fact, the mappings defined above can be considered generalized nonexpansive mappings with respect to f which can be symmetric as a particular case. Such a viewpoint, where in some problems of functional analysis the norm is replaced by a general function, was employed, for instance, in [18,19] in the study of generalized best approximation problems. At this point, it is worth noting that generalized nonexpansive mappings were also studied in [20,21]. In the present paper, we continue our studies of genericity in nonlinear analysis and establish the convergence of infinite products of generalized nonexpansive self-mappings to a common fixed point in a generic setting. In this connection, we remark in passing that infinite products of operators and studies of their convergence properties arise naturally in many areas of mathematics and its applications such as approximation theory, computed tomography, image recovery, optimization theory and population biology.

2. Main Results

We continue to use the definitions, notations and assumptions introduced in the Introduction. In particular, we suppose that f : X [ 0 , ) is a continuous function for which the equality f ( 0 ) = 0 is true and f ( K K ) is a bounded set, and which has properties (P1)–(P3).
Denote by A the collection of all continuous operators A : K K and by M the collection of all sequences { A t } t = 1 A . We denote the identity operator by A 0 for every mapping A A .
The set M is equipped with the metric ρ : M × M [ 0 , ) which is defined by
ρ ( { T t } t = 1 , { S t } t = 1 ) = sup { T t z S t z : z K , t = 1 , 2 , } ,
{ T t } t = 1 , { S t } t = 1 M .
It is not difficult to see that ( M , ρ ) is a complete metric space. We use the topology which is induced in M by ρ .
Denote by M n e the collection of all sequences { A t } t = 1 M such that for every integer t 1 , we have
f ( A t z 1 A t z 2 ) f ( z 1 z 2 ) , z 1 , z 2 K .
In view of property (P3), M n e is a closed subset of M . In the sequel the topological subspace M n e M is equipped with the relative topology.
A collection E of mappings A : K K is called uniformly equicontinuous (ue, for short) if for every positive number ϵ , there is a positive number δ such that T z 1 T z 2 ϵ for every operator T E and every pair of points z 1 , z 2 K which satisfies z 1 z 2 δ . Define
M u e : = { { A t } t = 1 M : { A t } t = 1   is   a   ue   set } .
It is clear that M u e is a closed subset of the metric space M . We equip M u e M with the relative topology. Properties (P1)–(P3) imply that M n e M u e .
Denote by M n e * the collection of all { A t } t = 1 M n e satisfying
t = 1 { x K : A t x = x }
and denote by M ¯ n e * the closure of M n e * in the complete metric space M n e . We equip the topological subspace M ¯ n e * M n e with the relative topology.
Denote by M u e * the collection of all sequences of operators { A t } t = 1 M u e for which there is a point x ^ K such that for every natural number t,
A t x ^ = x ^
and
f ( A t ξ x ^ ) f ( ξ x ^ )   for   every   ξ K ,
and denote by M ¯ u e * its closure in the complete metric space M u e . We equip M ¯ u e * M u e with the relative topology.
In this paper, we prove the following two generic convergence results. Note that in the case where f = · they are proved in [8].
Theorem 1.
There is a set F M ¯ n e * , which is a countable intersection of open and everywhere dense subsets of the space M ¯ n e * , such that for every sequence of operators { B t } t = 1 F , there is a point x * K for which the following two assertions are true.
1. B t x * = x * for all positive integers t.
2. For every positive number ϵ, there is an open neighborhood U of { B t } t = 1 in the topological space M ¯ n e * and an integer N 1 so that for every sequence of mappings { C t } t = 1 U , every natural number T N , every mapping r : { 1 , , T } and every point x K , we have
C r ( T ) · · C r ( 1 ) x x * ϵ .
Theorem 2.
There is a set F M ¯ u e * , which is a countable intersection of open and everywhere dense subsets of the topological space M ¯ u e * , such that for every sequence of operators { B t } t = 1 F , there is a point x * K for which the following two assertions are true.
1. B t x * = x * for all natural numbers t and
f ( B t ξ x * ) f ( ξ x * ) , ξ K , t = 1 , 2 , .
2. For every positive number ϵ, there is an open neighborhood U of the sequence of operators { B t } t = 1 in the topological space M ¯ u e * and an integer N 1 so that for every sequence of operators { C t } t = 1 U , every natural number T N , every mapping r : { 1 , , T } and every point x K , we have
C r ( T ) · · C r ( 1 ) x x * ϵ .

3. An Auxiliary Result

For every bounded mapping T : K X define
T : = sup { T ξ : ξ K } .
In our study, we will use the next result (Lemma 6.9 of [8]).
Lemma 1.
Assume that F is a nonempty uniformly equicontinuous collection of mappings T : K K , N 1 is an integer and ϵ > 0 . Then there is δ > 0 such that for every sequence of mappings { T t } t = 1 N F , every sequence of mappings { S t } t = 1 N , where the (not necessarily continuous) mappings S t : K K , t = 1 , N , satisfy
S t T t δ , t = 1 , N ,
and every point ξ K , the following inequality is true:
T N · · T 1 ξ S N · · S 1 ξ ϵ .

4. Proofs of Theorems

We prove Theorems 1 and 2 simultaneously.
For every sequence of mappings A = { A t } t = 1 M u e * , there is x ( A ) K for which
A t x ( A ) = x ( A ) , t = 1 , 2 , ,
f ( A t y x ( A ) ) f ( y x ( A ) ) , y K , t = 1 , 2 , .
Let A = { A t } t = 1 M u e * and γ ( 0 , 1 ) . For all natural numbers t , define a mapping A t γ : K K by
A t γ x = ( 1 γ ) A t x + γ x ( A ) , x K .
Evidently, { A t γ } t = 1 M u e , if { A t } t = 1 M n e , then { A t γ } t = 1 M n e , and
A t γ x ( A ) = x ( A ) , t = 1 , 2 , .
By property (P2), (6) and (7), for every point y K and every integer t 1 ,
f ( A t γ y x ( A ) ) = f ( ( 1 γ ) ( A t y x ( A ) ) )
ϕ ( 1 γ ) f ( A t y x ( A ) ) ϕ ( 1 γ ) f ( y x ( A ) ) .
Thus { A t γ } t = 1 M u e * , and if { A t } t = 1 M n e , then { A t γ } t = 1 M n e * . Clearly, the collection
{ { A t γ } t = 1 : { A t } t = 1 M u e * , γ ( 0 , 1 ) }
is an everywhere dense subset of the topological space M ¯ u e * and
{ { A t γ } t = 1 : { A t } t = 1 M n e * , γ ( 0 , 1 ) }
is an everywhere dense subset of the topological space M ¯ n e * .
In view of property (P3) for every positive integer q, there is
δ q ( 0 , ( 4 q ) 1 )
such that the following property is true:
(a) if the points z 1 , z 2 , ξ 1 , ξ 2 K satisfy z i ξ i 2 δ q , i = 1 , 2 , then
| f ( z 1 z 2 ) f ( ξ 1 ξ 2 ) | ( 4 q ) 1 .
Let { A t } t = 1 M u e * , γ ( 0 , 1 ) be given and let q be a natural number. In view of property (P1) there is
λ q ( 0 , δ q )
for which the following implication holds:
if   the   points   z 1 , z 2 K   and   f ( z 1 z 2 ) 2 λ q ,   then   z 1 z 2 δ q .
Fix a natural number n ( γ , q ) for which
ϕ ( 1 γ ) n ( γ , q ) sup { f ( z z ) : z , z K } < 4 1 λ q .
By (9) and (11), the following property holds:
(b) For every natural number T n ( γ , q ) , every mapping r : { 1 , , T } { 1 , 2 , } and every point x K , we have
f ( A r ( T ) γ A r ( 1 ) γ x x ( A ) ) ϕ ( 1 γ ) T f ( x x ( A ) ) 4 1 λ q .
Property (b) and Equation (10) imply that the following property is true:
(c) For each integer T n ( γ , q ) , each mapping r : { 1 , , T } { 1 , 2 , } and each point x K , we have
A r ( T ) γ A r ( 1 ) γ x x ( A ) δ q .
Lemma 1 implies that there is an open neighborhood
U ( { A t } t = 1 , γ , q )
of { A t γ } t = 1 in the topological space M ¯ u e * (respectively, in M ¯ n e * if { A t } t = 1 M n e * ) such that the following property is true:
(d) For each { C t } t = 1 U ( { A t } t = 1 , γ , q ) , each x K and each
r : { 1 , , n ( γ , q ) } { 1 , 2 , } ,
we have
ρ ( { C t } t = 1 , { A t γ } t = 1 ) δ q
and
A r ( n ( γ , q ) ) γ A r ( 1 ) γ x C r ( n ( γ , q ) ) C r ( 1 ) x δ q .
Assume that
{ C t } t = 1 U ( { A t } t = 1 , γ , q ) , r : { 1 , , n ( γ , q ) } { 1 , 2 , } .
Property (d) and (14) imply (13) for each x K . Property (c) and inequality (13) imply that for each x K ,
C r ( n ( γ , q ) ) C r ( 1 ) x x ( A ) 2 δ q .
Hence the next property is true:
(e) For every point x K , every sequence of mappings
{ C t } t = 1 U ( { A t } t = 1 , γ , q )
and every mapping
r : { 1 , , n ( γ , q ) } { 1 , 2 , } ,
inequality (15) is true.
Set
F = q = 1 { U ( { A t } t = 1 , γ , i ) : { A t } t = 1 M u e * , γ ( 0 , 1 ) , i = q , q + 1 , }
(respectively,
F = q = 1 { U ( { A t } t = 1 , γ , i ) : { A t } t = 1 M n e * , γ ( 0 , 1 ) , i = q , q + 1 , } ) .
Evidently, F is a countable intersection of open and everywhere dense subsets of the topological space M ¯ u e * (respectively, M ¯ n e * ).
Suppose that
{ B t } t = 1 F , ϵ > 0 .
Fix an integer q 1 such that
q 1 < ϵ .
By (16) and (17), there exist A = { A t } t = 1 M u e * (respectively, M n e * ), a positive number γ < 1 and a natural number i q for which
{ B t } t = 1 U ( { A t } t = 1 , γ , i ) .
It follows from (15), (18), (19) and property (e) that for every natural number t, every natural number p n ( γ , i ) and every point x K , we have
B t p x x ( A ) 2 δ i < ( 2 i ) 1 ( 2 q ) 1 < ϵ .
Since ϵ is an arbitrary positive number, the inequality above implies that for every point x K and every natural number t, there exists the limit
lim p B t p x .
When combined with inequality (20), the above relation implies that for each x K and each natural number t,
lim p B t p x x ( A ) 2 δ i < ( 2 i ) 1 < ϵ .
Since ϵ is an arbitrary positive number, the above inequality implies that there exists a point x * K for which
lim p B t p x = x * , x K , t = 1 , 2 , ,
B t x * = x * , t = 1 , 2 , ,
and
x ( A ) x * 2 δ i < ϵ .
Property (e), (15), (18) and (24) imply that for every point x K , every sequence of mappings { C t } t = 1 U ( { A t } t = 1 , γ , i ) , every natural number T n ( γ , i ) and every mapping r : { 1 , , T } { 1 , 2 , } , we have
C r ( T ) C r ( 1 ) x x * x * x ( A ) + C r ( T ) C r ( 1 ) x x ( A )
4 δ i < i 1 q 1 < ϵ .
Assume now that x K and that t is a positive integer. Property (a), (4.8) and (4.15) imply that
B t x A t γ x δ i .
In view of property (a), (24) and (25),
| f ( B t x x * ) f ( A t γ x x ( A ) ) | ( 4 i ) 1 .
When combined with (8), (9), (18), (24) and property (a), the inequality above implies that
f ( B t x x * ) ( 4 i ) 1 + f ( A t γ x x ( A ) ) ϵ / 2 + f ( x x ( A ) ) f ( x x * ) + ϵ
and
f ( B t x x * ) f ( x x * ) + ϵ .
Since ϵ is an arbitrary positive number, we conclude that
f ( B t x x * ) f ( x x * ) .
This completes the proofs of both Theorems 1 and 2.

Author Contributions

Both authors contributed equally to this paper. All authors have read and agreed to the published version of the manuscript.

Funding

The first author was partially supported by the Israel Science Foundation (Grant No. 820/17), by the Fund for the Promotion of Research at the Technion and by the Technion General Research Fund.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Acknowledgments

The first author was partially supported by the Israel Science Foundation (Grant No. 820/17), by the Fund for the Promotion of Research at the Technion and by the Technion General Research Fund.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Betiuk-Pilarska, A.; Benavides, T.D. Fixed points for nonexpansive mappings and generalized nonexpansive mappings on Banach lattices. Pure Appl. Func. Anal. 2016, 1, 343–359. [Google Scholar]
  2. Goebel, K.; Kirk, W.A. Topics in Metric Fixed Point Theory; Cambridge University Press: Cambridge, UK, 1990. [Google Scholar]
  3. Goebel, K.; Reich, S. Uniform Convexity, Hyperbolic Geometry, and Nonexpansive Mappings; Marcel Dekker: New York, NY, USA; Basel, Switzerland, 1984. [Google Scholar]
  4. Jachymski, J. Extensions of the Dugundji-Granas and Nadler’s theorems on the continuity of fixed points. Pure Appl. Funct. Anal. 2017, 2, 657–666. [Google Scholar]
  5. Kirk, W.A. Contraction mappings and extensions. In Handbook of Metric Fixed Point Theory; Kluwer: Dordrecht, The Netherlands, 2001; pp. 1–34. [Google Scholar]
  6. Kubota, R.; Takahashi, W.; Takeuchi, Y. Extensions of Browder’s demiclosedness principle and Reich’s lemma and their applications. Pure Appl. Func. Anal. 2016, 1, 63–84. [Google Scholar]
  7. Rakotch, E. A note on contractive mappings. Proc. Am. Math. Soc. 1962, 13, 459–465. [Google Scholar] [CrossRef]
  8. Reich, S.; Zaslavski, A.J. Genericity in nonlinear analysis. In Developments in Mathematics; Springer: New York, NY, USA, 2014; Volume 34. [Google Scholar]
  9. Zaslavski, A.J. Approximate solutions of common fixed point problems. In Springer Optimization and Its Applications; Springer: Cham, Switzerland, 2016. [Google Scholar]
  10. Zaslavski, A.J. Algorithms for solving common fixed point problems. In Springer Optimization and Its Applications; Springer: Cham, Switzerland, 2018. [Google Scholar]
  11. Banach, S. Sur les opérations dans les ensembles abstraits et leur application aux équations intégrales. Fund. Math. 1922, 3, 133–181. [Google Scholar] [CrossRef]
  12. Censor, Y.; Zaknoon, M. Algorithms and convergence results of projection methods for inconsistent feasibility problems: A review. Pure Appl. Func. Anal. 2018, 3, 565–586. [Google Scholar]
  13. Gibali, A. A new split inverse problem and an application to least intensity feasible solutions. Pure Appl. Funct. Anal. 2017, 2, 243–258. [Google Scholar]
  14. Takahashi, W. The split common fixed point problem and the shrinking projection method for new nonlinear mappings in two Banach spaces. Pure Appl. Funct. Anal. 2017, 2, 685–699. [Google Scholar]
  15. Takahashi, W. A general iterative method for split common fixed point problems in Hilbert spaces and applications. Pure Appl. Funct. Anal. 2018, 3, 349–369. [Google Scholar]
  16. Gabour, M.; Reich, S.; Zaslavski, A.J. A generic fixed point theorem. Indian J. Math. 2014, 56, 25–32. [Google Scholar]
  17. de Blasi, F.S.; Myjak, J. Sur la convergence des approximations successives pour les contractions non linéaires dans un espace de Banach. CR Acad. Sci. Paris 1976, 283, 185–187. [Google Scholar]
  18. Reich, S.; Zaslavski, A.J. Well-posedness of generalized best approximation problems. Nonlinear Func. Anal. Appl. 2002, 7, 115–128. [Google Scholar]
  19. Reich, S.; Zaslavski, A.J. Porous sets and generalized best approximation problems. Nonlinear Anal. Forum 2004, 9, 135–152. [Google Scholar]
  20. Reich, S.; Zaslavski, A.J. Contractivity and genericity results for a class of nonlinear mappings. J. Nonlinear Convex Anal. 2015, 16, 1113–1122. [Google Scholar]
  21. Reich, S.; Zaslavski, A.J. On a class of generalized nonexpansive mappings. Mathematics 2020, 8, 1085. [Google Scholar] [CrossRef]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Reich, S.; Zaslavski, A.J. Two Generic Convergence Results for Infinite Products of Generalized Nonexpansive Mappings. Symmetry 2022, 14, 534. https://doi.org/10.3390/sym14030534

AMA Style

Reich S, Zaslavski AJ. Two Generic Convergence Results for Infinite Products of Generalized Nonexpansive Mappings. Symmetry. 2022; 14(3):534. https://doi.org/10.3390/sym14030534

Chicago/Turabian Style

Reich, Simeon, and Alexander J. Zaslavski. 2022. "Two Generic Convergence Results for Infinite Products of Generalized Nonexpansive Mappings" Symmetry 14, no. 3: 534. https://doi.org/10.3390/sym14030534

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop