KYUNGPOOK Math. J. 2019; 59(4): 665-678
An Ishikawa Iteration Scheme for two Nonlinear Mappings in CAT(0) Spaces
Kritsana Sokhuma
Department of Mathematics, Faculty of Science and Technology, Phranakhon Rajabhat University, Bangkok 10220, Thailand
e-mail : k_sokhuma@yahoo.co.th
Received: December 9, 2017; Revised: December 10, 2018; Accepted: December 11, 2018; Published online: December 23, 2019.

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract

We construct an iteration scheme involving a hybrid pair of mappings, one a single-valued asymptotically nonexpansive mapping t and the other a multivalued nonexpansive mapping T, in a complete CAT(0) space. In the process, we remove a restricted condition (called the end-point condition) from results of Akkasriworn and Sokhuma [1] and and use this to prove some convergence theorems. The results concur with analogues for Banach spaces from Uddin et al. [16].

Keywords: Ishikawa iteration, CAT(0) spaces, multivalued mapping, asymptotically nonexpansive mapping.
1. Introduction

Fixed point theory in CAT(0) spaces was first studied by Kirk [6, 8] who showed that every nonexpansive mapping defined on a bounded closed convex subset of a complete CAT(0) space always has a fixed point. Since then, the existence problem of fixed point and the Δ–convergence problem of iterative sequences to a fixed point for nonexpansive mappings and asymptotically nonexpansive mappings in a CAT(0) space have been extensively developed with many papers published.

Let (X, d) be a geodesic metric space. 2K is denoted as the family of nonempty subsets of K, FB(K) is the collection of all nonempty closed bounded subsets of K and KC(K) is the collection of all nonempty compact convex subsets of K. A subset K of X is called proximinal if for each xX there exists an element kK such that

$d(x,k)=dist(x,K)=inf{d(x,y):y∈K}.$

The notation PB(K) is the collection of all nonempty bounded proximinal subsets of K.

Let H be the Hausdorff metric with respect to d such that

$H(A,B)=max{supx∈Adist(x,B),supy∈Bdist(y,A)}, A,B∈FB(X),$

where dist(x,B) = inf{d(x, y) : yB} is the distance from the point x to the subset B.

A mapping t : KK is said to be nonexpansive if

$d(tx,ty)≤d(x,y) for all x,y∈K.$

A point x is called a fixed point of t if tx = x.

A mapping t : KK is called asymptotically nonexpansive if there is a sequence {kn} of positive numbers with the property $limn→∞kn=1$ such that

$d(tnx,tny)≤knd(x,y) for all n≥1,x,y∈K.$

A multivalued mapping T : KFB(K) is said to be nonexpansive if

$H(Tx,Ty)≤d(x,y) for all x,y∈K.$

A multivalued mapping T : KFB(K) is said to satisfy condition(E) if there exists μ ≥ 1 such that for each x, yK

$dist(x,Ty)≤μ dist(x,Tx)+d(x,y).$

Let T : KPB(K) be a multivalued mapping and define the mapping PT for each x by

$PT(x):={y∈Tx:d(x,y)=dist(x,Tx)}.$

A point x is called a fixed point for a multivalued mapping T if xTx.

Then, IT is strongly demiclosed if for every sequence {xn} in K which converges to xK and such that limn→∞d(xn, Txn) = 0, then xT(x).

For every continuous mapping T : K → 2K, IT is strongly demiclosed but the converse is not true. Notice also that if T satisfies condition (E), then IT is strongly demiclosed.

The notation Fix(T) stands for the set of fixed points of a mapping T and Fix(t) ∩ Fix(T) stands for the set of common fixed points of t and T. A precise point x is called a common fixed point of t and T if x = txTx.

In 2009, Laokul and Panyanak [9] defined the iterative and proved the Δ–convergence for nonexpansive mapping in CAT(0) spaces as follows:

Let C be a nonempty closed convex subset of a complete CAT(0) space and t : CC be a nonexpansive mapping with Fix(t) := {xC : tx = x} ≠ ∅︀. Suppose {xn} is generated iteratively by x1C,

$yn=βnxn⊕(1−βn)xn,xn+1=αntyn⊕(1−αn)xn.$

for all n ∈ ℕ, where {αn} and {βn} are real sequences in [0, 1] such that one of the following two conditions is satisfied:

(i) αn ∈ [a, b] and βn ∈ [0, b] for some a, b with 0 < ab < 1,

(ii) αn ∈ [a, 1] and βn ∈ [a, b] for some a, b with 0 < ab < 1,

Then the sequence {xn} is Δ–convergent to a fixed point of t.

In 2010, Sokhuma and Kaewkhao [15] proved the convergence theorem for a common fixed point in Banach spaces as follows.

Let E be a nonempty compact convex subset of a uniformly convex Banach space X, and t : EE and T : EKC(E) be a single-valued nonexpansive mapping and a multivalued nonexpansive mapping, respectively. Assume in addition that Fix(t) ∩ Fix(T) ≠ ∅︀ and Tw = {w} for all w ∈ Fix(t) ∩ Fix(T). Suppose {xn} is generated iterative by x1E,

$yn=(1−βn)xn+βnzn,xn+1=(1−αn)xn+αntyn,$

for all n ∈ ℕ where znTxn and {αn}, {βn} are sequences of positive numbers satisfying 0 < aαn, βnb < 1. Then the sequence {xn} converges strongly to a common fixed point of t and T.

In 2013, Sokhuma [14] proved the convergence theorem for a common fixed point in CAT(0) spaces as follows.

Let K be a nonempty compact convex subset of a complete CAT(0) space X and t : KK and T : KFC(K) a single-valued nonexpansive mapping and a multivalued nonexpansive mapping respectively and Fix(t) ∩ Fix(T) ≠ ∅︀ satisfying Tw = {w} for all w ∈ Fix(t) ∩ Fix(T). Let {xn} is generated iterative by x1K,

$yn=(1−βn)xn⊕βnzn,xn+1=(1−αn)xn⊕αntyn,$

for all n ∈ ℕ where znTxn and {αn}, {βn} are sequences of positive numbers satisfying 0 < aαn, βnb < 1. Then the sequence {xn} converges strongly to a common fixed point of t and T.

In 2013, Laowang and Panyanak proved the convergence theorem for a common fixed point in CAT(0) spaces as follows.

### Theorem 1.1.([10])

Let C be a nonempty bounded closed convex subset of a complete CAT(0) space X. Let f : CC be a pointwise asymtotically nonexpansive mapping, and g : CC a quasi-nonexpansive mapping, and let T : CKC(C) be a multivalued mapping satisfying conditions (E) and Cλ for some λ ∈ (0, 1). If f, g and T are pairwise commuting, then there exists a point zC such that z = f(z) = g(z) ∈ T(z).

In 2015, Akkasriworn and Sokhuma [1] proved the convergence theorem for a common fixed point in a complete CAT(0) space as follows.

### Theorem 1.2

Let E be a nonempty bounded closed convex subset of a complete CAT(0) space X, t : EE and T : EFB(E) an asymptotically nonexpansive mapping and a multivalued nonexpansive mapping, respectively. Assume that t and T are commuting and Fix(t) ∩ Fix(T) ≠ ∅︀ satisfying Tw = {w} for all wFix(t) ∩ Fix(T) and $∑n=1∞(kn−1)<∞$. Let {xn} be the sequence of the modified Ishikawa iterates defined by

$yn=(1−βn)xn⊕βnzn,xn+1=(1−αn)xn⊕αntnyn,$

for all n ∈ ℕ where znTtnxn and {αn}, {βn} ∈ [0, 1]. Then {xn} is Δ–convergent to a common fixed point of t and T.

In 2016, Uddin and Imdad [17] introduced the following iteration scheme:

Let K be a nonempty closed, bounded and convex subset of Banach space X, let f : KK be a single-valued nonexpansive mapping and let T : KFB(K) be a multivalued nonexpansive mapping with Fix(f) ∩ Fix(T) ≠ ∅︀ such that PT is a nonexpansive mapping. The sequence {xn} of the modified Ishikawa iteration is defined by

$yn=αnzn+(1−αn)xn,xn+1=βnfyn+(1−βn)xn,$

where x0K, znPT (xn) and 0 < aαn, βnb < 1. Then {xn} converges strongly to a common fixed point of f and T.

The Ishikawa iteration method was studied with respect to a pair of single-valued asymptotically nonexpansive mapping and a multivalued nonexpansive mapping. It also established the convergence theorem of a sequence from such process in a nonempty bounded closed convex subset of a complete CAT(0) space. A restricted condition (called end-point condition) in Akkasriworn and Sokhuma’s results was removed [1].

Here, an iteration method modifying the above was introduced and called the Ishikawa iteration method

### Definition 1.3

Let K be a nonempty bounded closed convex subset of a complete CAT(0) space X, t : KK be a single-valued asymptotically nonexpansive mapping and T : KPB(K) be a multivalued nonexpansive mapping where PT (x) = {yTx : d(x, y) = dist(x, Tx)}. For fixed x1K the sequence {xn} of the Ishikawa iteration is defined by

$yn=(1−βn)xn⊕βnzn,xn+1=(1−αn)xn⊕αntnyn,$

for all n ∈ ℕ where znPT (tnxn) and {αn}, {βn} ∈ (0, 1).

2. Preliminaries

Relevant basic definitions followed previous research results and iterative methods were used frequently.

Let (X, d) be a metric space. A geodesic path joining xX to yX is a map c from a closed interval [0, s] ⊂ ℝ to X such that c(0) = x, c(s) = y, and d(c(t), c(u)) = |tu| for all t, u ∈ [0, s]. In particular, c is an isometry and d(x, y) = s. The image α of c is called a geodesic (or metric) segment joining x and y. When it is unique this geodesic segment is denoted by [x, y]. The space (X, d) is said to be a geodesic space if every two points of X are joined by a geodesic, and X is said to be uniquely geodesic if there is exactly one geodesic joining x and y for each x, yX. A subset YX is said to be convex if Y includes every geodesic segment joining any two of its points.

A geodesic triangle Δ(x1, x2, x3) in a geodesic metric space (X, d) consists of three points x1, x2, x3 in X (the vertices of Δ) and a geodesic segment between each pair of vertices (the edges of Δ). A comparison triangle for the geodesic triangle Δ(x1, x2, x3) in (X, d) is a triangle Δ̄(x1, x2, x3) := Δ(χ̄1, χ̄2, χ̄3) in the Euclidean plane such that for i, j ∈ {1, 2, 3}.

A geodesic space is said to be a CAT(0) space if all geodesic triangles of appropriate size satisfy the following comparison axiom.

CAT(0): Let Δ be a geodesic triangle in X and let Δ̄ be a comparison triangle for Δ. Then Δ is said to satisfy the CAT(0) inequality if for all x, y ∈ Δ and all comparison points χ̄, ȳ ∈ Δ̄, .

If x, y1, y2 are points in a CAT(0) space and

$y0=12y1⊕12y2,$

then the CAT(0) inequality implies that

$d(x,y0)2≤12d(x,y1)2+12d(x,y2)2−14d(y1,y2)2.$

This is the (CN) inequality of Bruhat and Tits [3]. A geodesic space is a CAT(0) space if and only if it satisfies the (CN) inequality [2].

The following results and methods deal with the concept of asymptotic centres. Let K be a nonempty closed convex subset of a CAT(0) space X and {xn} be a bounded sequence in X. For xX, define the asymptotic radius of {xn} at x as the number

$r(x,{xn})=lim supn→∞ d(xn,x).$

Let

$r≡r(K,{xn}):=inf{r(x,{xn}):x∈K}$

and

$A≡A(K,{xn}):={x∈K:r(x,{xn})=r}.$

The number r and the set A are called the asymptotic radius and asymptotic centre of {xn} relative to K respectively.

If X is a complete CAT(0) space and K is a closed convex subset of X, then A(K, {xn}) consists of exactly one point. A sequence {xn} in CAT(0) space X is said to be Δ–convergent to xX if x is the unique asymptotic centre of every subsequence of {xn}. A bounded sequence {xn} is said to be regular with respect to K if for every subsequence {$xn′$}, we get

$r(K,{xn})=r(K,{xn′}).$

The definition of Δ–convergence is presented below.

### Definition 2.1.([12, 8])

A sequence {xn} in a CAT(0) space X is said to be Δ–convergent to xX if x is the unique asymptotic centre of {un} for every subsequence {un} of {xn}. In this case,$Δ−limn→∞xn=x$and x is the Δ – limit of {xn}.

Some elementary facts about CAT(0) spaces which will be used in the proofs of the main results are stated. The following lemma can be found in [4, 5, 8].

### Lemma 2.2.([8])

Every bounded sequence in a complete CAT(0) space has a Δ–convergent subsequence.

### Lemma 2.3.([4])

If K is a closed convex subset of a complete CAT(0) space and {xn} is a bounded sequence in K, then the asymptotic centre of {xn} is in K.

### Lemma 2.4.([5])

Let (X, d) be a CAT(0) space.

(i) For x, yX and u ∈ [0, 1], there exists a unique point z ∈ [x, y] such that$d(x,z)=ud(x,y) and d(y,z)=(1−u)d(x,y).$

he notation (1 − u)xty is used for the unique point z satisfying (2.2).

(ii) For x, y, zX and u ∈ [0, 1], $d((1−u)x⊕uy,z)≤(1−u)d(x,z)+ud(y,z).$

The existence of fixed points for asymptotically nonexpansive mappings in CAT(0) spaces was proved by Kirk [7] as the following result.

### Theorem 2.5

Let K be a nonempty bounded closed and convex subset of a complete CAT(0) space X and let t : KK be asymptotically nonexpansive. Then t has a fixed point.

### Theorem 2.6.([13])

Let X be a complete CAT(0) space and K be a nonempty bounded closed and convex subset of X and t : KK be an asymptotically nonexpansive mapping. Then It is demiclosed at 0.

### Corollary 2.7.([5])

Let K be a closed and convex subset of a complete CAT(0) space X and let t : KX be an asymptotically nonexpansive mapping. Let {xn} be a bounded sequence in K such that $limn→∞ d(txn,xn)=0$and $Δ−limn→∞xn=w$. Then tw = w.

### Lemma 2.8.([11])

Let X be a complete CAT(0) space and let xX. Suppose {αn} is a sequence in [a, b] for some a, b ∈ (0, 1) and {xn}, {yn} are sequences in X such that $lim supn→∞ d(xn,x)≤r,lim supn→∞ d(yn,x)≤r$, and $limn→∞ d((1−αn)xn⊕αnyn,x)=r$ for some r ≥ 0. Then $limn→∞ d(xn,yn)=0$.

The following fact is well-known.

### Lemma 2.9

Let X be a CAT(0) space, K be a nonempty compact convex subset of X and {xn} be a sequence in K. Then,

$dist(y,Ty)≤d(y,xn)+dist(xn,Txn)+H(Txn,Ty)$

where yK and T be a multivalued mapping from K in to FB(K).

The important property can be found in [18].

### Lemma 2.10

Let {an} and {bn} be two sequences of nonnegative numbers such that

$an+1≤(1+bn)an,$

for all n ≥ 1. If $∑n=1∞bn$ converges, then $limn→∞an$ exists. In particular, if there is a subsequence of {an} which converges to 0 then $limn→∞an=0$.

3. Main Results

The following lemmas play very important roles in this section.

### Lemma 3.1

Let K be a nonempty bounded closed convex subset of a complete CAT(0) space X, T : KPB(K) be a multivalued mapping, and PT (x) = {yTx : d(x, y) = dist(x, Tx)}. Then the following are equivalent

(1) xFix(T), that is xTx;

(2) PT (x) = {x}, that is x = y for each yPT (x);

(3) xFix(PT), that is xPT (x).

Further, Fix(T) = Fix(PT).

Proof

(1) implies (2). Since xTx, then d(x, Tx) = 0. Therefore, for any yPT (x), d(x, y) = dist(x, Tx) = 0 and so x = y. That is, PT (x) = {x}.

(2) implies (3). Since PT (x) = {x}, then x ∈ Fix(PT) and hence xPT (x).

(3) implies (1). Since x ∈ Fix(PT), then xPT (x). Therefore, d(x, x) = dist(x, Tx) = 0 and so xTx by the closedness of Tx.

This implies that Fix(T) = Fix(PT).

### Lemma 3.2

Let K be a nonempty bounded closed convex subset of a complete CAT(0) space X, t : KK and T : KPB(K) a single-valued asymptotically nonexpansive mapping and a multivalued nonexpansive mapping respectively with Fix(t) ∩ Fix(T) ≠ ∅︀ such that PT is nonexpansive and $∑n=1∞(kn−1)<∞$. Let {xn} be the sequence of Ishikawa iterates defined by (1.2). Then $limn→∞ d(xn,w)$ exists for all wFix(t) ∩ Fix(T).

Proof

Let x1K and w ∈ Fix(t)∩Fix(T), in the view of Lemma 3.1, wPT(w) = {w}. Now consider,

$d(xn+1,w)=d((1−αn)xn⊕αntnyn,w)≤(1−αn)d(xn,w)+αnd(tnyn,tnw)≤(1−αn)d(xn,w)+αnknd(yn,w)=(1−αn)d(xn,w)+αnknd((1−βn)xn⊕βnzn,w)≤(1−αn)d(xn,w)+αnkn(1−βn)d(xn,w)+αnknβnd(zn,w)≤(1−αn)d(xn,w)+αnkn(1−βn)d(xn,w)+αnknβndist(zn,PT(w))≤(1−αn)d(xn,w)+αnkn(1−βn)d(xn,w)+αnknβnH(PT(tnxn),PT(w))≤(1−αn)d(xn,w)+αnkn(1−βn)d(xn,w)+αnknβnd(tnxn,w)≤(1−αn)d(xn,w)+αnkn(1−βn)d(xn,w)+αnβnkn2d(xn,w)=[1+αn(1+βnkn) (kn−1)]d(xn,w).$

By the convergence of kn and αn, βn ∈ (0, 1), there exists some M >0 such that

$d(xn+1,w)≤[1+M(kn−1)]d(xn,w).$

By condition $∑n=1∞(kn−1)<∞$ and Lemma 2.10, which implies that $limn→∞ d(xn,w)$ exists.

### Lemma 3.3

Let K be a nonempty bounded closed convex subset of a complete CAT(0) space X, t : KK and T : KPB(K) a single-valued asymptotically nonexpansive mapping and a multivalued nonexpansive mapping respectively with Fix(t) ∩ Fix(T) ≠ ∅︀ such that PT is nonexpansive and $∑n=1∞(kn−1)<∞$. Let {xn} be the sequence of Ishikawa iterates defined by (1.2). Then $limn→∞ d(tnyn,xn)=0$.

Proof

Let x1K and w ∈ Fix(t) ∩ Fix(T), in view of Lemma 3.1, wPT (w) = {w}. From Lemma 3.2, $limn→∞ d(xn,w)=c$ is set. Now consider,

$d(yn,w)=d((1−βn)xn⊕βnzn,w)≤(1−βn)d(xn,w)+βnd(zn,w)=(1−βn)d(xn,w)+βndist(zn,PT(w))≤(1−βn)d(xn,w)+βnH(PT(tnxn),PT(w))≤(1−βn)d(xn,w)+βnd(tnxn,w)≤(1−βn)d(xn,w)+βnknd(xn,w).$

Notice that

$d(tnyn,w)≤knd(yn,w)≤kn[(1−βn)d(xn,w)+βnknd(xn,w)]=kn(1−βn)d(xn,w)+βnkn2d(xn,w)=(kn−knβn+βnkn2)d(xn,w)=[kn+βnkn(kn−1)]d(xn,w)≤[1+βnkn(kn−1)]d(xn,w).$

Then,

$lim supn→∞ d(tnyn,w)≤lim supn→∞knd(yn,w)≤lim supn→∞[1+βnkn(kn−1)]d(xn,w).$

By kn → 1 as n→∞ and αn, βn ∈ (0, 1), which implies that

$lim supn→∞ d(tnyn,w)≤lim supn→∞ d(yn,w)≤lim supn→∞ d(xn,w)=c.$

Since $c=limn→∞ d(xn+1,w)=limn→∞ d((1−αn)xn⊕αntnyn,w)$, it implies by Lemma 2.8 that $limn→∞ d(tnyn,xn)=0$.

### Lemma 3.4

Let K be a nonempty bounded closed convex subset of a complete CAT(0) space X, t : KK and T : KPB(K) a single-valued asymptotically nonexpansive mapping and a multivalued nonexpansive mapping respectively with Fix(t) ∩ Fix(T) ≠ ∅︀ such that PT is nonexpansive and $∑n=1∞(kn−1)<∞$. Let {xn} be the sequence of Ishikawa iterates defined by (1.2). Then $limn→∞ d(xn,zn)=0$.

Proof

Let x1K and w ∈ Fix(t) ∩ Fix(T), in view of Lemma 3.1, wPT (w) = {w}. Consider,

$d(xn+1,w)=d((1−αn)xn⊕αntnyn,w)≤(1−αn)d(xn,w)+αnd(tnyn,w)≤(1−αn)d(xn,w)+αnknd(yn,w)$

and hence

$d(xn+1,w)−d(xn,w)αn≤knd(yn,w)−d(xn,w).$

Therefore, by 0 < aαnb < 1, it follows that

$(d(xn+1,w)−d(xn,w)αn)+d(xn,w)≤knd(yn,w).$

Thus,

$lim infn→∞{(d(xn+1,w)−d(xn,w)αn)+d(xn,w)}≤lim infn→∞knd(yn,w).$

It follows that

$c≤lim infn→∞ d(yn,w).$

Since $lim supn→∞ d(yn,w)≤c$, it follows that

$c=limn→∞ d(yn,w)=limn→∞ d((1−βn)xn⊕βnzn,w).$

Recall that

$d(zn,w)=dist(zn,PT(w))≤H(PT(tnxn),PT(w))≤d(tnxn,w)≤knd(xn,w).$

Hence,

$lim supn→∞ d(zn,w)≤lim supn→∞knd(xn,w)≤lim supn→∞ d(xn,w)=c.$

Thus,

$limn→∞ d(xn,zn)=0.$

### Lemma 3.5

Let K be a nonempty bounded closed convex subset of a complete CAT(0) space X, t : KK and T : KPB(K) a single-valued asymptotically nonexpansive mapping and a multivalued nonexpansive mapping respectively with Fix(t) ∩ Fix(T) ≠ ∅︀ such that PT is nonexpansive and $∑n=1∞(kn−1)<∞$. Let {xn} be the sequence of Ishikawa iterates defined by (1.2). Then $limn→∞ d(tnxn,xn)=0$.

Proof

Consider,

$d(tnxn,xn)≤d(tnxn,tnyn)+d(tnyn,xn)≤knd(xn,yn)+d(tnyn,xn)=knd(xn,(1−βn)xn⊕βnzn)+d(tnyn,xn)≤kn[(1−βn)d(xn,xn)+βnd(xn,zn)]+d(tnyn,xn)=knβnd(xn,zn)+d(tnyn,xn).$

Then,

$limn→∞ d(tnxn,xn)≤limn→∞knβnd(zn,xn)+limn→∞ d(tnyn,xn).$

Hence, by Lemmas 3.3 and 3.4,$limn→∞ d(tnxn,xn)=0$.

### Lemma 3.6

Let K be a nonempty bounded closed convex subset of a complete CAT(0) space X, t : KK and T : KPB(K) a single-valued asymptotically nonexpansive mapping and a multivalued nonexpansive mapping respectively with Fix(t) ∩ Fix(T) ≠ ∅︀ such that PT is nonexpansive and $∑n=1∞(kn−1)<∞$. Let {xn} be the sequence of Ishikawa iterates defined by (1.2). Then $limn→∞ d(txn,xn)=0$.

Proof

Consider,

$d(txn,xn)=d(xn,txn)≤d(xn,tnxn)+d(tnxn,txn)≤d(xn,tnxn)+k1[d(tn−1xn,tn−1xn−1)+d(tn−1xn−1,xn)]≤d(xn,tnxn)+k1kn−1d(xn,xn−1)+k1d(tn−1xn−1,xn)≤d(xn,tnxn)+k1kn−1αn−1d(tn−1yn−1,xn−1)+k1(1−αn−1)d(xn−1,tn−1xn−1)+k1kn−1αn−1d(yn−1,xn−1)≤d(xn,tnxn)+k1kn−1αn−1d(tn−1yn−1,xn−1)+k1(1−αn−1)d(xn−1,tn−1xn−1)+k1kn−1αn−1βn−1d(zn−1,xn−1).$

It follows from Lemmas 3.2 – 3.4 that,

$limn→∞ d(txn,xn)=0.$

### Theorem 3.7

Let K be a nonempty bounded closed convex subset of a complete CAT(0) space X, t : KK and T : KPB(K) a single-valued asymptotically nonexpansive mapping and a multivalued nonexpansive mapping respectively with Fix(t) ∩ Fix(T) ≠ ∅︀ such that PT is nonexpansive and $∑n=1∞(kn−1)<∞$. Let {xn} be the sequence of Ishikawa iterates defined by (1.2). Then {xn} is Δ–convergent to y implies yFix(t) ∩ Fix(T).

Proof

Since {xn} is Δ–convergent to y. From Lemma 3.6,

$limn→∞ d(txn,xn)=0.$

By Corollary 1.6, yK and ty = y, it follows that y ∈ Fix(t). It follows from Lemma 2.9 that,

$dist(y,PT(y))≤d(y,xn)+dist(xn,PT(xn))+H(PT(xn),PT(y))≤d(y,xn)+d(xn,zn)+d(xn,y)→0 as n→∞.$

It follows that y ∈ Fix(PT) then y ∈ Fix(T). Therefore y ∈ Fix(t) ∩ Fix(T) as desired.

### Theorem 3.8

Let K be a nonempty bounded closed convex subset of a complete CAT(0) space X, t : KK and T : KPB(K) a single-valued asymptotically nonexpansive mapping and a multivalued nonexpansive mapping respectively with Fix(t) ∩ Fix(T) ≠ ∅︀ such that PT is nonexpansive and $∑n=1∞(kn−1)<∞$. Let {xn} be the sequence of Ishikawa iterates defined by (1.2). Then {xn} is Δ–convergent to a common fixed point of t and T.

Proof

Since Lemma 3.6 guarantees that {xn} is bounded and $limn→∞ d(txn,xn)=0$. So, let ωw(xn) := ∪A({un}) where the union is taken over all subsequences {un} of {xn}. If ωw(xn) ⊂ Fix(t) ∩ Fix(T), then there exists a subsequence {un} of {xn} such that A({un}) = {u}. By Lemmas 1.2 and 1.3 there exists a subsequence {vn} of {un} such that $Δ−limn→∞vn=v∈K$. Since

$limn→∞ d(tvn,vn)=0$, it follows that v ∈ Fix(t). Since,

$dist(v,PT(v))≤dist(v,PT(vn))+H(PT(vn),PT(v))≤d(v,zn)+d(vn,v)≤d(v,vn)+d(vn,zn)+d(vn,v)→0 as n→∞.$

It follows that v ∈ Fix(PT) and v ∈ Fix(T) by Lemma 3.1. Therefore v ∈ Fix(t) ∩ Fix(T) as desired. Suppose that uv, since t is a single-valued asymptotically nonexpansive mapping and v ∈ Fix(t) ∩ Fix(T),$limn→∞ d(xn,v)$ exists by Lemma 3.2. Then by the uniqueness of asymptotic centres,

$lim supn→∞ d(vn,v)

a contradiction, and hence u = v ∈ Fix(t) ∩ Fix(T).

To show that {xn} is Δ–convergent to a common fixed point of t and T, it suffices to show that ωw(xn) consists of exactly one point. Let {un} be a subsequence of {xn}. By Lemmas 1.2 and 1.3 there exists a subsequence {vn} of {un} such that $Δ−limn→∞vn=v∈K$. Let A({un}) = {u} and A({xn}) = {x}. It has seen that u = v and v ∈ Fix(t) ∩ Fix(T).

It can complete the proof by showing that x = v. Suppose not, since $limn→∞ d(xn,v)$ exists, then by the uniqueness of asymptotic centres,

$lim supn→∞ d(vn,v)

a contradiction, and hence the conclusion follows.

Acknowledgements

I would like to thank the Institute for Research and Development, Phranakhon Rajabhat University, for financial support.

References
1. N. Akkasriworn, and K. Sokhuma. Convergence theorems for a pair of asymptotically and multivalued nonexpansive mapping in CAT(0) spaces. Commun Korean Math Soc., 30(3)(2015), 177-189.
2. MR. Bridson, and A. Haefliger. Metric spaces of non-positive curvature, , Springer-Verlag, Berlin, 1999.
3. F. Bruhat, and J. Tits. Groupes reductifs sur un corps local. Inst Hautes Etudes Sci Publ Math., 41(1972), 5-251.
4. S. Dhompongsa, WA. Kirk, and B. Panyanak. Nonexpansive set-valued mappings in metric and Banach spaces. J Nonlinear Convex Anal., 8(2007), 35-45.
5. S. Dhompongsa, WA. Kirk, and B. Sims. Fixed points of uniformly lipschitzian mappings. Nonlinear Anal., 65(2006), 762-772.
6. WA. Kirk. Geodesic geometry and fixed point theory, Seminar of Mathematical Analysis(Malaga/Seville, 2002/2003), Colecc. Abierta, Univ. Sevilla Secr. Publ, Seville, 64(2003), 195-225.
7. WA. Kirk. Geodesic geometry and fixed point theory II, International Conference on Fixed Point Theory and Applications, Yokohama Publ, Yokohama, (2004), 113-142.
8. WA. Kirk, and B. Panyanak. A concept of convergence in geodesic spaces. Nonlinear Anal., 68(2008), 3689-3696.
9. T. Laokul, and B. Panyanak. Approximating fixed points of nonexpansive mappings in CAT(0) spaces. Int J Math Anal., 3(2009), 1305-1315.
10. W. Laowang, and B. Panyanak. A note on common fixed point results in uniformly convex hyperbolic spaces. J Math., (2013) Art. ID 503731, 5 pp.
11. W. Laowang, and B. Panyanak. Approximating fixed points of nonexpansive nonself mappings in CAT(0) spaces. Fixed Point Theory Appl., (2010) Art. ID 367274, 11 pp.
12. TC. Lim. Remarks on some fixed point theorems. Proc Amer Math Soc., 60(1976), 179-182.
13. B. Nanjaras, and B. Panyanak. Demiclosed principle for asymptotically nonexpansive mappings in CAT(0) spaces. Fixed Point Theory Appl., (2010) Art. ID 268780, 14 pp.
14. K. Sokhuma. Δ–convergence Theorems for a pair of single valued and multivalued nonexpansive mappings in CAT(0) spaces. J Math Anal., 4(2013), 23-31.
15. K. Sokhuma, and A. Kaewkhao. Ishikawa iterative process for a pair of single-valued and multivalued nonexpansive mappings in Banach spaces. Fixed Point Theory Appl., (2010) Art. ID 618767, 9 pp.
16. I. Uddin, AA. Abdou, and M. Imdad. A new iteration scheme for a hybrid pair of generalized nonexpansive mappings. Fixed Point Theory Appl., 2014(2014), 13 pp.
17. I. Uddin, and M. Imdad. A new iteration scheme for a hybrid pair of nonexpansive mappings. Honam Math J., 38(1)(2016), 127-139.
18. H. Zhou, RP. Agarwal, YJ. Cho, and YS. Kim. Nonexpansive mappings and iterative methods in uniformly convex Banach spaces. Georgian Math J., 9(2002), 591-600.