R E S E A R C H
Open Access
-convergence for mixed-type total
asymptotically nonexpansive mappings in
hyperbolic spaces
Li-Li Wan
**Correspondence:
[email protected] School of Science, Southwest University of Science and Technology, Mianyang, Sichuan 621010, China
Abstract
In this paper, we prove some-convergence theorems in a hyperbolic space. A mixed Agarwal-O’Regan-Sahu type iterative scheme for approximating a common fixed point of total asymptotically nonexpansive mappings is constructed. Our results extend some results in the literature.
MSC: 47H09; 49M05
Keywords: total asymptotically nonexpansive mappings; hyperbolic space; -convergence; mixed Agarwal-O’Regan-Sahu type iterative scheme
1 Introduction and preliminaries
In this paper, we work in the setting of hyperbolic spaces introduced by Kohlenbach []. Concretely, (X,d,W) is called a hyperbolic space if (X,d) is a metric space andW:X× X×[, ]→Xis a function satisfying
(I) ∀x,y,z∈X,∀λ∈[, ],d(z,W(x,y,λ))≤( –λ)d(z,x) +λd(z,y); (II) ∀x,y∈X,∀λ,λ∈[, ],d(W(x,y,λ),W(x,y,λ)) =|λ–λ| ·d(x,y);
(III) ∀x,y∈X,∀λ∈[, ],W(x,y,λ) =W(y,x, ( –λ));
(IV) ∀x,y,z,w∈X,∀λ∈[, ],d(W(x,z,λ),W(y,w,λ))≤( –λ)d(x,y) +λd(z,w).
If a space satisfies only (I), it coincides with the convex metric space introduced by Taka-hashi []. The concept of hyperbolic spaces in [] is more restrictive than the hyperbolic type introduced by Goebel [] since (I)-(III) together are equivalent to (X,d,W) being a space of hyperbolic type in []. But it is slightly more general than the hyperbolic space defined in Reich [] (see []). This class of metric spaces in [] covers all normed linear spaces,R-trees in the sense of Tits, the Hilbert ball with the hyperbolic metric (see []), Cartesian products of Hilbert balls, Hadamard manifolds (see [, ]) andCAT() spaces in the sense of Gromov (see []). A thorough discussion of hyperbolic spaces and a detailed treatment of examples can be found in [] (see also [–]).
A hyperbolic space isuniformly convex[] if foru,x,y∈X,r> and∈(, ], there existsδ∈(, ] such that
d
W
x,y,
,u
≤( –δ)r,
provided thatd(x,u)≤r,d(y,u)≤randd(x,y)≥r.
A mapη: (,∞)×(, ]→(, ] is calledmodulus of uniform convexityifδ=η(r,) for givenr> . Besides,ηismonotoneif it decreases withr(for a fixed), that is,
η(r,)≤η(r,), ∀r≥r> .
A subsetCof a hyperbolic spaceXisconvexifW(x,y,λ)∈Cfor allx,y∈Candλ∈[, ]. Let (X,d) be a metric space, and letCbe a nonempty subset ofX. Recall thatT:C→C
is said to be a ({νn},{μn},ζ)-total asymptotically nonexpansive mappingif there exist
non-negative sequences{νn},{μn}withνn→,μn→ and a strictly increasing continuous
functionζ: [,∞)→[,∞) withζ() = such that
dTnx,Tny≤d(x,y) +νnζ
d(x,y)+μn, ∀n≥,x,y∈C. ()
It is well known that each nonexpansive mapping is an asymptotically nonexpansive map-ping and each asymptotically nonexpansive mapmap-ping is a ({νn},{μn},ζ)-total
asymptoti-cally nonexpansive mapping.
T:C→Cis said to beuniformly L-Lipschitzianif there exists a constantL> such that
dTnx,Tny≤Ld(x,y), ∀n≥,x,y∈C.
The following iteration process is a translation of the mixed Agarwal-O’Regan-Sahu type iterative scheme introduced in [] from Banach spaces to hyperbolic spaces. The iteration rate of convergence is similar to the Picard iteration process and faster than other fixed point iteration processes. Besides, it is independent of Mann and Ishikawa iteration pro-cesses.
⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩
x∈C,
xn+=W(Snxn,Tnyn,αn), n≥,
yn=W(Snxn,Tnxn,βn),
()
whereCis a nonempty closed and convex subset of a complete uniformly convex hyper-bolic spaceXwith monotone modulus of uniform convexityη.Ti:C→C,i= , , is a uniformlyLi-Lipschitzian and ({νn(i)},{μ(ni)},ζ(i))-total asymptotically nonexpansive
map-ping, andSi:C→C, i= , , is a uniformlyLi-Lipschitzian and ({νn(i)},{μn(i)},ζ(i))-total
asymptotically nonexpansive mapping such that the following conditions are satisfied:
() ∞n=ν(ni)<∞,
∞
n=μ (i)
n <∞,
∞
n=ν (i)
n <∞,
∞
n=μ (i)
n <∞,i= , ;
() There exists a constantM> such thatζ(i)(r)≤Mr,ζ(i)(r)≤Mr,∀r≥,i= , . Remark Without loss of generality, we can assume thatTi:C→CandSi:C→C,i=
, , both are uniformlyL-Lipschitzian and ({νn},{μn},ζ)-total asymptotically
nonexpan-sive mappings satisfying conditions () and (). In fact, lettingνn=max{νn(i),νn(i),i= , }, μn=max{μ(ni),μ(ni),i= , },L=max{Li,Li,i= , } andζ =max{ζ(i),ζ(i),i= , }, thenSi
andTi,i= , , are the required mappings.
in aCAT() space using the mixed Agarwal-O’Regan-Sahu type iterative scheme. More precisely, one of the results is as follows.
Theorem [] Let C be a bounded closed and convex subset of a complete CAT()
space X. Let Ti : C →C and Si :C → C, i= , , be uniformly L-Lipschitzian and
({νn},{μn},ζ)-total asymptotically nonexpansive mappings.IfF:=
i=F(Ti)∩F(Si)=∅ and the following conditions are satisfied:
(i) ∞n=νn<∞and
∞
n=μn<∞;
(ii) there exist constantsa,b∈(, )with <b( –a)≤ such that{αn} ⊂[a,b];
(iii) there exists a constantM> such thatζ(r)≤Mr,r≥; (iv) d(x,Tiy)≤d(Six,Tiy)for allx,y∈Candi= , ,
then the sequence{xn}defined by()-converges to a common fixed point of Tiand Si,
i= , .
Theorem can be viewed as a improvement and extension of several well-known results in Banach spaces andCAT() spaces, such as [] and []. Our purpose of this paper is to extend Theorem from theCAT() spaces setting to the general setup of uniformly convex hyperbolic spaces.
Let{xn}be a bounded sequence in a hyperbolic spaceX. Forx∈X, we define
rx,{xn}=lim sup n→∞
d(x,xn).
Theasymptotic radius r({xn})of {xn}is given by
r{xn}
=infrx,{xn}
:x∈X.
Theasymptotic radius rC({xn})of {xn}with respect to C⊂Xis given by
rC{xn}=infrx,{xn}:x∈C.
Theasymptotic center A({xn})of {xn}is the set
A{xn}
=x∈X:rx,{xn}
=r{xn}
.
Theasymptotic center AC({xn})of {xn}with respect to C⊂Xis the set
AC{xn}=x∈C:rx,{xn}=rC{xn}.
Recall that a sequence{xn}inXis said to-converge to x∈Xifxis the unique asymptotic center of{un}for every subsequence{un}of{xn}. In this case, we callxthe-limit of {xn}.
The following lemmas are important in our paper.
Lemma [] Let(X,d,W)be a uniformly convex hyperbolic space with monotone modu-lus of uniform convexityη.Let x∈X and{αn}be a sequence in[a,b]for some a,b∈(, ).If
{xn}and{yn}are sequences in X such thatlim supn→∞d(xn,x)≤c,lim supn→∞d(yn,x)≤c andlimn→∞d(W(xn,yn,αn),x) =c for some c≥,then
lim
n→∞d(xn,yn) = .
Lemma [] Let C be a nonempty closed convex subset of a uniformly convex hyperbolic space,and let{xn}be a bounded sequence in C such that A({xn}) ={y}and r({xn}) =ρ.If
{ym}is another sequence in C such thatlimm→∞r(ym,{xn}) =ρ,thenlimm→∞ym=y.
Lemma [] Let{an},{bn}and{cn}be sequences of nonnegative numbers such that
an+≤( +bn)an+cn, ∀n≥.
If∞n=bn<∞andn∞=cn<∞,thenlimn→∞anexists.
2 Main results
In this section, we prove our main theorems.
Theorem Let C be a nonempty closed and convex subset of a complete uniformly con-vex hyperbolic space X with monotone modulus of uniform concon-vexity η.Let Ti:C→C and Si:C→C, i= , ,be uniformly L-Lipschitzian and({νn},{μn},ζ)-total asymptoti-cally nonexpansive mappings.IfF:=i=F(Ti)∩F(Si)=∅and the following conditions are satisfied:
(i) ∞n=νn<∞and
∞
n=μn<∞;
(ii) there exist constantsa,b∈(, )such that{αn} ⊂[a,b];
(iii) there exists a constantM> such thatζ(r)≤Mr,r≥; (iv) d(x,Tiy)≤d(Six,Tiy)for allx,y∈Candi= , ,
then the sequence{xn}defined by()-converges to a common fixed point of Tiand Si, i= , .
Proof We divide our proof into three steps.
Step . In the sequel, we shall show that for eachp∈F,
lim
n→∞d(xn,p) and nlim→∞d(xn,F) exist. ()
In fact, by conditions (), (), (I) and (iii), one gets
d(yn,p) =d
WSnxn,Tnxn,βn
,p
≤( –βn)d
Snxn,p+βndTnxn,p
≤( –βn)
d(xn,p) +νnζ
d(xn,p)+μn
+βn
d(xn,p) +νnζ
d(xn,p)+μn
=d(xn,p) +νnζ
d(xn,p)
+μn
and
d(xn+,p) =dWSnxn,Tnyn,αn,p
≤( –αn)d
Snxn,p+αndTnyn,p
≤( –αn)
d(xn,p) +νnζ
d(xn,p)+μn
+αn
d(yn,p) +νnζ
d(yn,p)+μn
≤( –αn)
( +νnM)d(xn,p) +μn
+αn
( +νnM)d(yn,p) +μn
. ()
Combining () and (), one has
d(xn+,p)≤( +σn)d(xn,p) +ξn, ∀n≥ andp∈F, ()
and
d(xn+,F)≤( +σn)d(xn,F) +ξn, ∀n≥, ()
whereσn=νnM(+αn(+νnM)),ξn= (+αn(+νnM))μn. Furthermore, using condition (i),
one has
∞
n=
σn<∞ and
∞
n=
ξn<∞. ()
Consequently, a combination of (), (), () and Lemma shows that () is proved. Step . We claim that
lim
n→∞d(xn,Tixn) = and nlim→∞d(xn,Sixn) = , i= , . ()
In fact, it follows from () thatlimn→∞d(xn,p) exists for each givenp∈F. Without loss
of generality, we assume that
lim
n→∞d(xn,p) =c≥. ()
By () and (), one has
lim inf
n→∞ d(yn,p)≤lim supn→∞ d(yn,p)≤nlim→∞
( +νnM)d(xn,p) +μn
=c. ()
Noting
dTnyn,p=dTnyn,Tnp
≤d(yn,p) +νnζ
d(yn,p)
+μn
≤( +νnM)d(yn,p) +μn, ∀n≥,
and
dSnxn,p
=dSnxn,Snp
≤d(xn,p) +νnζ
d(xn,p)+μn
by () and (), one has
lim sup n→∞
dTnyn,p≤c and lim sup n→∞
dSnxn,p≤c. ()
Besides, by () one gets
d(xn+,p) =d
WSn
xn,Tnyn,αn
,p≤( +σn)d(xn,p) +ξn,
which yields that
lim n→∞d
WSn
xn,Tnyn,αn
,p=c. ()
Now, by (), () and Lemma , we have
lim n→∞d
Sn
xn,Tnyn
= . ()
Using the same method, we can also have that
lim n→∞d
Snxn,Tnxn= . ()
It follows from (), () and condition (iv) that
lim n→∞d
xn,Tnyn≤ lim n→∞d
Snxn,Tnyn= ()
and
lim n→∞d
xn,Tnxn≤ lim n→∞d
Snxn,Tnxn= . ()
By virtue of (), one has
dyn,Snxn=dWSnxn,Tnxn,βn
,Snxn
≤βndTnxn,Snxn→ asn→ ∞. ()
Because we have
d(xn,yn)≤dxn,Tnxn+dTnxn,Snxn+dSnxn,yn,
it follows from (), () and () that
lim
n→∞d(xn,yn) = . ()
Combining () and (), one obtains
dxn,Tnxn
≤dxn,Tnyn
+dTnyn,Tnxn
≤dxn,Tnyn+d(xn,yn) +νnζ
d(xn,yn)+μn
Moreover, it follows from () and () that
dSnxn,Tnxn≤dSnxn,Tnyn+dTnyn,Tnxn
≤dSnxn,Tnyn+Ld(yn,xn)→ asn→ ∞. ()
This jointly with () and () yields that
dSn
xn,xn
≤dSn
xn,Tnxn
+dTn
xn,xn
→ asn→ ∞,
and
d(xn+,xn)≤d
WSnxn,Tnyn,αn
,xn
≤( –αn)dSnxn,xn+αndTnyn,xn→ asn→ ∞. ()
Now by (), () and (), for eachi= , , one gets
d(xn,Tixn)≤d(xn,xn+) +dxn+,Tin+xn+
+dTin+xn+,Tin+xn
+dTin+xn,Tixn
≤( +L)d(xn,xn+) +d
xn+,Tin+xn+
+LdTinxn,xn→ asn→ ∞. ()
For eachi= , , the combination of (), (), (), () and (iv) yields that
d(xn,Sixn)≤d
xn,Tinxn
+dSixn,Tinxn
≤dxn,Tinxn
+dSnixn,Tinxn
→ asn→ ∞. ()
Therefore, () is proved.
Step . Now we are in a position to prove the -convergence of {xn}. Since{xn} is bounded, by Lemma , it has a unique asymptotic centerAC({xn}) ={x∗}. Let{un}be any subsequence of{xn}withAC({un}) ={u}, then by () and (), for eachi= , , we have
lim
n→∞d(un,Tiun) = and d(un,Siun) = . ()
We claim thatu∈F. In fact, we define a sequence{zm}inCbyzm=Tmu. Then one has
d(zm,un)≤dTmu,Tmun+dTmun,Tm–un+· · ·+d(Tun,un)
≤d(u,un) +νnζ
d(u,un)
+μn+Ld(Tun,un) +· · ·+d(Tun,un).
By (), one gets
lim sup
n→∞ d(zm,un)≤lim supn→∞ d(u,un) =r
u,{un}
,
which yields that
Lemma shows thatlimm→∞Tmu=u. BecauseTis uniformly continuous, we have
Tu=T
lim
m→∞T
m
u
= lim
m→∞T
m+ u=u.
Hence,u∈F(T). Using the same method, we can prove thatu∈F. By the uniqueness of asymptotic centers, we get thatx∗=u. It implies thatx∗is the unique asymptotic center of{un}for each subsequence{un}of{xn}, that is,{xn} -converges tox∗∈F. The proof
is completed.
Competing interests
The author declares that they have no competing interests.
Acknowledgements
Supported by General Project of Educational Department in Sichuan (No. 13ZB0182) and Doctor Research Foundation of Southwest University of Science and Technology (No. 11zx7130).
Received: 30 August 2013 Accepted: 22 October 2013 Published:22 Nov 2013
References
1. Kohlenbach, U: Some logical metatheorems with applications in functional analysis. Trans. Am. Math. Soc.357, 89-128 (2005)
2. Takahashi, W: A convexity in metric spaces and nonexpansive mappings. Kodai Math. Semin. Rep.22, 142-149 (1970) 3. Goebel, K, Kirk, WA: Iteration processes for nonexpansive mappings. In: Singh, SP, Thomeier, S, Watson, B (eds.)
Topological Methods in Nonlinear Functional Analysis. Contemporary Mathematics vol. 21, pp. 115-123 Am. Math. Soc., Providence (1983)
4. Reich, S, Shafrir, I: Nonexpansive iterations in hyperbolic spaces. Nonlinear Anal., Theory Methods Appl.15, 537-558 (1990)
5. Goebel, K, Reich, S: Uniform Convexity, Hyperbolic Geometry, and Nonexpansive Mappings. Monographs and Textbooks in Pure and Applied Mathematics, vol. 83, ix+170. Dekker, New York (1984)
6. Reich, S, Zaslavski, AJ: Generic aspects of metric fixed point theory. In: Kirk, WA, Sims, B (eds.) Handbook of Metric Fixed Point Theory, pp. 557-576. Kluwer Academic, Dordrecht (2001)
7. Bridson, M, Haefliger, A: Metric Spaces of Non-Positive Curvature. Springer, Berlin (1999)
8. Shimizu, T, Takahashi, W: Fixed points of multivalued mappings in certain convex metric spaces. Topol. Methods Nonlinear Anal.8, 197-203 (1996)
9. Agarwal, RP, O’Regan, D, Sahu, DR: Iterative construction of fixed points of nearly asymptotically nonexpansive mappings. J. Nonlinear Convex Anal.8, 61-79 (2007)
10. Chang, SS, Wang, L, Lee, HWJ, Chan, C: Strong and–convergence for mixed type total asymptotically nonexpansive mappings in CAT(0) spaces. Fixed Point Theory Appl.2013, Article ID 122 (2013)
11. Sahin, A, Basarir, M: On the strong convergence of modifiedS-iteration process for asymptotically quasi-nonexpansive mappings in CAT(0) space. Fixed Point Theory Appl.2013, Article ID 12 (2013) 12. Khan, AR, Fukhar-ud-din, H, Khan, MAA: An implicit algorithm for two finite families of nonexpansive maps in
hyperbolic spaces. Fixed Point Theory Appl.2012, Article ID 54 (2012)
13. Leustean, L: Nonexpansive iterations in uniformly convexW-hyperbolic spaces. In: Leizarowitz, A, Mordukhovich, BS, Shafrir, I, Zaslavski, A (eds.) Nonlinear Analysis and Optimization I: Nonlinear Analysis. Contemporary Mathematics, vol. 513, pp. 193-209. Am. Math. Soc., Providence (2010)
10.1186/1029-242X-2013-553