Analysis in Theory and Applications DOI: 10.4208/ata.2013.v29.n4.

3
Anal. Theory Appl., Vol. 29, No. 4 (2013), pp. 342-347

Fixed Point Curve for Weakly Inward Contractions

and Approximate Fixed Point Property
P. Riyas1, ∗ and K. T. Ravindran2
1 Department of Mathematics, Krishna Menon Memorial Government Women’s
College, Kannur, India-670004, India
2 Payyanur College, Department of Mathematics, Payyannur College, Kannur,

India-670627, India
Received 19 September 2013; Accepted (in revised version) 8 November 2013
Available online 31 December 2013

Abstract. In this paper, we discuss the concept of fixed point curve for linear interpo-
lations of weakly inward contractions and establish necessary condition for a nonex-
pansive mapping to have approximate fixed point property.

Key Words: Nonexpansive mapping, contraction, weakly inward contraction, approximate fixed
point property.

AMS Subject Classifications: 46B20, 47H07, 47H10, 54E40, 54H99

1 Introduction
The most well known result in the theory of fixed point is Banach contraction principle.
It has been used to develop much of the rest of fixed point theory. Another key result
in the field is a theorem due to Browder, Ghde, and Kirk involving Hilbert spaces and
nonexpansive mappings. In 1988, Sam B. Nadler and K. Ushijima introduced the con-
cept of fixed point curves for linear interpolations of contraction mappings using Banach
contraction principle. Here we shall establish the concept of fixed point curves for linear
interpolations of weakly inward contractions using the result [2, 4], which is an extension
of the Banach contraction principle for non-self contraction mappings.
Before introduce fixed point curve and approximate fixed point property for weakly
inward contractions, we will give some preliminary definitions and theorems.

∗ Corresponding author. Email addresses: riyasmankadavu@gmail.com (P. Riyas), ktravindran@rediffmail.

com (K. T. Ravindran)

http://www.global-sci.org/ata/ 342 c
2013 Global-Science Press
P. Riyas and K. T. Ravindran / Anal. Theory Appl., 29 (2013), pp. 342-347 343

2 Preliminaries
Definition 2.1 (see [1]). Let K be a closed convex subset of a Banach space X and let C ⊂ K.
Then a mapping T : C → X is said to be inward mapping (respectively, weakly inward) on C
relative to K if Tx ∈ IK ( x) (repectively, Tx ∈ IK ( x)) for x ∈ C, where

IK ( x) = (1 − t) x + ty : y ∈ K and t ≥ 0 ,

and IK ( x) is a closure of the IK ( x) in X.

Lemma 2.1 (see [1]). IK ( x) is a closed convex set containing K for each x ∈ K .

Definition 2.2 (see [1]). Let K be a closed convex set. A map T : C ⊆ K → X is said to be
generalized inward on C relative to K if the following condition is satisfied

d( Tx,K ) < k x − Txk, ∀ x ∈ C with Tx ∈

/ K,

where d(y,K ) = inf{ky − uk : u ∈ K }.

Theorem 2.1 (see [3]). Let C be a non empty closed convex subset of a Banach space X and
T : C → X a weakly inward contraction mapping. Then T has a unique fixed point in C.

Proposition 2.1 (see [3]). Let C be a non empty closed convex subset of a Banach space
X and T : C → X a non expansive mapping that is weakly inward. Then for u ∈ C and
t ∈ (0,1) there exist one point xt ∈ C such that xt = (1 − t)u + tTxt . If C is bounded then
xt − Txt → 0 as t → 1.

Corollary 2.1 (see [3]). Let C be a non-empty closed convex (or star shaped) and bounded
subset of a Banach space X and T : C → C a non expansive mapping. Then there exist a
sequence { xn } in C, such that
lim k xn − Txn k = 0.
n→∞

Definition 2.3 (see [2]). Let ( X,d) be a metric space and f : X → X. Then f has the approx-
imate fixed point property (a.f.p.p) if ∀ǫ > 0, Fǫ ( f ) 6= φ.

Definition 2.4 (see [2]). Let ( X,d) be a metric space. A mapping f : X → X is said to be
asymptotically regular if d( f n ( x), f n+1 ( x)) → 0 as n → ∞ and ∀ x ∈ X.

Lemma 2.2 (see [2]). Let ( X,d) be a metric space and f : X → X be such that f is asymptotically
regular. Then f has approximate fixed point property.
344 P. Riyas and K. T. Ravindran / Anal. Theory Appl., 29 (2013), pp. 342-347

3 Main results
Here first of all we shall discuss a necessary condition for a family of weakly inward
contraction to have a fixed point and using this result we define fixed point curve.

Theorem 3.1. Let X be a Banach space, K be a non-empty closed and convex subset of X and let
T1 ,T2 : K → X, be two weakly inward contractions. Then there exists a family of weakly inward
contraction maps {φt : K → X : t ∈ [0,1]}, such that, ∀t ∈ [0,1], φt has a unique fixed point xt in
K.

Proof. Define φ : K ×[0,1] → X by φ( x,t) = (1 − t) T1 ( x)+ tT2 ( x). Then,

kφ( x,t)− φ(y,t)k =k(1 − t)( T1 ( x)− T1 (y))+ t( T2 ( x)− T2 (y)k

≤(1 − t)kT1 ( x)− T1 (y)k+ tkT2 ( x)− T2 (y)k
≤(1 − t) L1 k x − yk+ tL2 k x − yk, where 0 < L1 , L2 < 1,
≤ L((1 − t)k x − yk+ tk x − yk)
= Lk x − yk, ∀ x,y ∈ K and t ∈ [0,1], (3.1)

where L = max{ L1 , L2 }. Hence φ is a contraction uniformly over [0,1]. As T1 and T2 are

weakly inward, T1 ( x) and T2 ( x)∈ IK ( x), ∀ x ∈ K. Then by Lemma 2.1, (1− t) T1 ( x)+ tT2 ( x)∈
(1 − t) IK ( x)+ tIK ( x) = IK ( x), i.e.,

φ( x,t) ∈ IK ( x), ∀ x ∈ K and t ∈ [0,1].

This shows that φ is a weakly inward contraction over [0,1]. Now define φt : K → X by
φt ( x) = φ( x,t) for t ∈ [0,1] and so by (3.1), we have,

kφt ( x)− φt (y)k ≤ Lk x − yk, ∀ x,y ∈ K and t ∈ [0,1].

That is for each t ∈[0,1], φt is a weakly inward contractiion on a closed convex subset K of
X into X. Hence by Theorem 2.1, φt has a unique fixed point xt in K for each t ∈ [0,1].

Definition 3.1 (Fixed point curve). Let T1 and T2 be two weakly inward contraction on
a closed, convex and bounded subset K of X and FW ( T1 ,T2 ) denote the set of all fixed
points of φt , t ∈[0,1]. Then the map G : [0,1]→ FW ( T1 ,T2 ) defined by G (t)= xt is called the
fixed point curve for T1 and T2 .

Lemma 3.1. Let T1 ,T2 : K → X be two weakly inward contractions on a closed, convex and
bounded subset K of X. Suppose that T1 and T2 have no common fixed points. Then the map
G : [0,1] → FW ( T1 ,T2 ) defined by G (t) = xt is one to one.

Proof. Suppose that G is not one to one. Then there exist some t1 and t2 (t1 6= t2 ) such that
G (t1 ) = G (t2 ), i.e.,
xt1 = xt2 = p.
P. Riyas and K. T. Ravindran / Anal. Theory Appl., 29 (2013), pp. 342-347 345

Then,
(1 − t1 ) T1 ( p)+ t1 T2 ( p) = (1 − t2 ) T1 ( p)+ t2 T2 ( p) =⇒ T1 ( p) = T2 ( p).
Hence,
p = xt1 = φt1 ( p) = (1 − t1 ) T1 ( p)+ t1 T2 ( p) = T1 ( p) = T2 ( p),
which is a contradiction.

Corollary 3.1. The fixed point curve G : [0,1] → FW ( T1 ,T2 ) is continuous.

Corollary 3.2. FW ( T1 ,T2 ) is a closed set.
Now we discuss approximate fixed point property for a non expansive map defined
in a real Hilbert space.
Theorem 3.2. Let H be a real Hilbert space, C be a closed, convex subset of H and f : C → C be
a non expansive mapping such that set of all ǫ− fixed points Fǫ ( f ) of f is non empty. Then the
mapping f t defined by f t ( x) = tx +(1 − t) f ( x) is asymptotically regular for all t ∈ (0,1).
Proof. Let x ∈ C. Consider a sequence ( xn ) defined by the relation xn+1 = f t ( xn ), n =
0,1,2, ··· , and x0 = x. Let y be an ǫ− fixed point of f . Then
xn+1 − y = txn +(1 − t) f ( xn )− y
=t( xn − y)+(1 − t)( f ( xn )− y). (3.2)
For any θ (n),
θ (n)( xn − f ( xn )) = θ (n)( xn − y)− θ (n)( f ( xn )− y). (3.3)
Also,
k x n + 1 − y k2
=ht( xn − y)+(1 − t)( f ( xn )− y),t( xn − y)+(1 − t)( f ( xn )− y)i
=ht( xn − y),t( xn − y)i+ht( xn − y), (1 − t)( f ( xn )− y)i
+h(1 − t)( f ( xn )− y),t( xn − y)i+h(1 − t)( f ( xn )− y), (1 − t)( f ( xn )− y)i
=t2 k xn − yk2 + t(1 − t)h( xn − y), ( f ( xn )− y)i
+ t(1 − t)h( f ( xn )− y), ( xn − y)i+ (1 − t)2 k f ( xn )− yk2
=t2 k xn − yk2 + 2t(1 − t)h( xn − y), ( f ( xn )− y)i+(1 − t)2 k f ( xn )− yk2 , (3.4)
and
[θ (n)]2 k xn − f ( xn )k2
=hθ (n)( xn − f ( xn )),θ (n)( xn − f ( xn ))i
=hθ (n)( xn − y)− θ (n)( f ( xn )− y),θ (n)( xn − y)− θ (n)( f ( xn )− y)i
=[θ (n)]]2 k xn − yk2 −[θ (n)]2 h xn − y, f ( xn )− yi
−[θ (n)]2 h f ( xn )− y,xn − yi+[θ (n)]2 k f ( xn )− yk2 . (3.5)
346 P. Riyas and K. T. Ravindran / Anal. Theory Appl., 29 (2013), pp. 342-347

From (3.4) and (3.5),

k xn+1 − yk2 +[θ (n)]2 k xn − f ( xn )k2

=(t2 +[θ (n)]2 )k xn − yk2 +(2t(1 − t)− 2[θ (n)]2 )h xn − y, f ( xn )− yi
+((1 − t)2 +[θ (n)]2 )k f ( xn )− yk2
≤(t2 +[θ (n)]2 )k xn − yk2 +(2t(1 − t)− 2[θ (n)]2 )k xn − ykk f ( xn )− yk
+((1 − t)2 +[θ (n)]2 )k f ( xn )− yk2
≤(t2 +[θ (n)]2 )k xn − yk2 +(2t(1 − t)− 2[θ (n)]2 )k xn − yk(k f ( xn )− f (y)k+k f (y)− yk)
+((1 − t)2 +[θ (n)]2 )(k f ( xn )− f (y)k+k f (y)− yk)2
<(t2 +[θ (n)]2 )k xn − yk2 +(2t(1 − t)− 2[θ (n)]2 )k xn − yk(k xn − yk+ ǫ)
<(t2 +[θ (n)]2 )k xn − yk2 +(2t(1 − t)− 2[θ (n)]2 )k xn − yk(k xn − yk+ ǫ)
+((1 − t)2 +[θ (n)]2 )(k xn − yk+ ǫ)2
<(k xn − yk+ ǫ)2 ,
=⇒[θ (n)]2 k xn − f ( xn )k2 < (k xn − yk+ ǫ)2 −k xn+1 − yk2 . (3.6)
p
If we take θ (n) = n (k xn − yk+ ǫ)2 −k xn+1 − yk2 , then

1
k xn − f ( xn )k2 < , ∀n,
n2
and therefore,

k xn − f t ( xn )k =k xn − txn −(1 − t) f ( xn )k
=k(1 − t) xn −(1 − t) f ( xn )k
=(1 − t)k xn − f ( xn )k, ∀n,
1
=⇒k xn − xn+1 k2 < 2 , ∀n.
n
∞
=⇒ ∑ k xn − xn+1 k2 < ∞.
n =1

This shows that f t is asymptotically regular.

Corollary 3.3. Let H be a Hilbert space, C be a closed convex subset of H and f : C → C

be a non expansive mapping such that set of all ǫ− fixed points Fǫ ( f ) of f is non empty.
Then the mapping f t defined by f t ( x) = tx +(1 − t) f ( x) has the approximate fixed point
property.

Theorem 3.3. Let H be a Hilbert space, C be a closed convex subset of H and f : C → C be a

non expansive mapping, such that Fǫ ( f ) 6= φ for some ǫ. Then f has the approximate fixed point
property.
P. Riyas and K. T. Ravindran / Anal. Theory Appl., 29 (2013), pp. 342-347 347

Proof. For each t ∈ (0,1), define f t ( x) = tx +(1 − t) f ( x). Then by the above corollary f t has
the approximate fixed point property for each t. Therefore, for any ǫ > 0 and t ∈ (0,1),
Fǫ ( f t ) 6= φ, i.e.,

k f t ( x)− xk < ǫ, ∀t and ǫ,

=⇒kt( x − f ( x)+( f ( x)− x)k < ǫ, ∀t and ǫ.

Letting t → 0, we thus obtain k f ( x)− xk < ǫ, ∀ǫ > 0. Hence f has the approximate fixed
point property.

References

[1] Huagui Duan, Shaoyuan Xu and Guozh En Li, Fixed Points of weakly inward 1-set contrac-
tion mappings, J. Korean Math. Soc., 45(6) (2008), 1725–1740.
[2] Madalina Berinde, Approximate Fixed Point Theorems, Studia Univ. Babes-Bolyai, Mathe-
matica, Volume LI, Number 1, March 2006.
[3] R. P. Agarwal, Donal O’Regan and D. R. Sahu, Fixed Point Theory for Lipschitzian Type
Mappings with Applications, Springer Science & Business Media, LLC, 2009.
[4] Sam B. Nadler and K. Ushijima, Fixed point curves for linear interpolations of contraction
maps, Appl. Anal., 29 (1988).
[5] T. Vidali, Fixed point curve generated by nonexpansive mappings, Fixed Point Theory, 6(2)
(2005), 333–339.
[6] W. A. Kirk, Approximate fixed points of nonexpansive maps, Fixed Point Theory, 10(2)
(2009), 275–288.