 Research
 Open access
 Published:
On new generalizations of Smarzewski’s fixed point theorem
Journal of Inequalities and Applications volume 2014, Article number: 493 (2014)
Abstract
In this work, we prove some new generalizations of Smarzewski’s fixed point theorem and some new fixed point theorems which are original and quite different from the wellknown results in the literature.
MSC:46B20, 47H09, 54H25.
1 Introduction and preliminaries
Let (X,\parallel \cdot \parallel ) be a normed space with its zero vector θ. We use B(X) and S(X) to denote respectively the closed unit ball and unit sphere centered at θ with radius 1, that is,
and
The notion of uniformly convex (UC, for short) Banach space was introduced by Clarkson [1], and the research of geometric properties of the Banach space started from 1936. The function {\delta}_{X}:[0,2]\to [0,1], defined by
is called the modulus of convexity of X. The normed space X is called uniformly convex if {\delta}_{X}(\epsilon )>0 for every \epsilon \in (0,2]. It is well known that a uniformly convex Banach space is reflexive and all Hilbert spaces and Banach spaces {\ell}^{p} and {L}^{p} (1<p<\mathrm{\infty}) all are uniformly convex; see, e.g., [2–7] for more details. The normed space X is said to be strictly convex if \parallel x+y\parallel <2 whenever x,y\in S(X) with \parallel xy\parallel >0. It is obvious that a Banach space X is strictly convex if and only if {\delta}_{X}(2)=1. It is well known that the strict convexity of a normed space X can be characterized by the properties: for any nonzero vectors x,y\in X, if \parallel x+y\parallel =\parallel x\parallel +\parallel y\parallel, then y=cx for some real c>0. For each \epsilon >0, the modulus of convexity of X in the direction z\in S(X) is defined by
Clearly, {\delta}_{X}(\epsilon )=inf\{{\delta}_{X}(\epsilon ,z):z\in S(X)\}. The Banach space X is called uniformly convex in every direction (UCED, for short) if for any z\in S(X) and \epsilon >0, \delta (\epsilon ,z)>0. Some characterizations of UCED Banach spaces were proved by Day et al. [7]; see also [4].
Fact 1.1 (see, e.g., [2–4, 7])

(a)
Every UC Banach space is UCED.

(b)
Every UCED Banach space is strictly convex.
Let (X,\parallel \cdot \parallel ) be a Banach space and K be a given nonempty closed subset of X. For x\in X and a bounded sequence \{{x}_{n}\}\subset X, define the asymptotic radius of \{{x}_{n}\} at x as the number
The asymptotic radius of \{{x}_{n}\} with respect to K is defined by
and the set
is called the asymptotic center of \{{x}_{n}\} with respect to K. For any bounded sequence \{{x}_{n}\} in X, r(x,\{{x}_{n}\}) is easily seen to be a nonnegative, continuous and convex functional of x\in X. Moreover, if K is a nonempty convex subset of X, then A(K,\{{x}_{n}\}) is also convex.
Fact 1.2 [[8], Lemma 2.2]
Every bounded sequence in a UCED Banach space X has a unique asymptotic center with respect to any nonempty weakly compact convex subset of X.
Definition 1.1 A normed space (X,\parallel \cdot \parallel ) is said to have the (UAC)property if every bounded sequence in X has a unique asymptotic center with respect to any nonempty weakly compact convex subset of X.
According to Facts 1.1 and 1.2, it is easy to know that Hilbert spaces, UC Banach spaces and UCED Banach spaces all have the (UAC)property.
Let C be a nonempty subset of a normed space (X,\parallel \cdot \parallel ) and T:C\to X be a mapping. T is said to be nonexpansive if
The concept of firmly nonexpansive mappings was introduced by Bruck [9]. Let \lambda \in (0,1). The mapping T is said to be λfirmly nonexpansive [9] if
It is obvious that every λfirmly nonexpansive mapping is nonexpansive, but the converse is not true. The following example shows that there exists a nonexpansive mapping which is not a λfirmly nonexpansive mapping for some \lambda \in (0,1).
Example A Let X=\mathbb{R} with the absolutevalue norm \cdot  and C=[2,10]. Let T:C\to X be defined by Tx=x. Then T is a nonexpansive mapping. For x=6, y=4 and \lambda =\frac{1}{2}, we have
which deduces that T is not a \frac{1}{2}firmly nonexpansive mapping. In fact, T is not λfirmly nonexpansive for all \lambda \in (0,1).
In 1965, Browder [10], Kirk [11] and Göhde [12] proved respectively that every nonexpansive mapping T from a nonempty weakly compact convex subset K of a uniformly convex Banach space X into itself has a fixed point. It is know that the convexity of sets and mappings plays an important role in fixed point theory and the union of convex sets does not ensure that it is convex. In 1991, Smarzewski [13] proved the following interesting theorem.
Theorem 1.1 (Smarzewski [13])
Let X be a uniformly convex Banach space and C={\bigcup}_{k=1}^{n}{C}_{k} be a finite union of nonempty weakly compact convex subsets {C}_{k} of X. If T:C\to C is a λfirmly nonexpansive mapping for some \lambda \in (0,1), then T has a fixed point in C.
Smarzewski’s fixed point theorem (i.e., Theorem 1.1) is not always true if T is merely nonexpansive, even in X=\mathbb{R}.
Example B [13]
Let X=\mathbb{R} with the absolutevalue norm \cdot  and C=[2,1]\cup [2,1]. Then the mapping T:C\to C defined by Tx=x is nonexpansive and fixed point free.
In this paper, in order to promote Smarzewski’s fixed point theorem, we first introduce the concept of reactive firmly nonexpansive mappings.
Definition 1.2 Let C be a nonempty subset of a normed space (X,\parallel \cdot \parallel ) and \phi :C\times C\to [0,1) be a function. A mapping T:C\to X is said to be reactive firmly nonexpansive with respect to φ if
Remark 1.1

(a)
Every reactive firmly nonexpansive mapping is nonexpansive.

(b)
It is obvious that any λfirmly nonexpansive mapping is reactive firmly nonexpansive with respect to the function φ defined by \phi (s,t)=\lambda for all (s,t)\in C\times C.
Example C Let X=\mathbb{R} with the absolutevalue norm \cdot  and C=[2,5]\cup [10,20]. Let T:C\to X be defined by
Then the following statements hold.

(a)
T is a nonexpansive mapping.

(b)
T is not \frac{1}{3}firmly nonexpansive.

(c)
Define \phi :C\times C\to [0,1) by
\phi (s,t):=\{\begin{array}{ll}\frac{1}{10},& \text{if}s,t\in [2,5],\\ \frac{1}{3},& \text{otherwise.}\end{array}
Then T is reactive firmly nonexpansive with respect to φ.
Proof Obviously, statement (a) holds. To see (b), let x=4 and y=3. Since
we show that T is not \frac{1}{3}firmly nonexpansive. Finally, we prove (c). We consider the following four possible cases to verify
for all x,y\in C.
Case 1. If x,y\in [2,5], then
and
So (1.1) holds for all x,y\in [2,5].
Case 2. If x\in [2,5] and y\in [10,20], then
and
Since
we prove that (1.1) holds for all x\in [2,5] and y\in [10,20].
Case 3. If x\in [10,20] and y\in [2,5], then
and
Since
we prove that (1.1) holds for all x\in [10,20] and y\in [2,5].
Case 4. If x,y\in [10,20], then
and
So (1.1) holds for all x,y\in [10,20].
By Cases 14, we verify that inequality (1.1) holds for all x,y\in C. Hence T is reactive firmly nonexpansive with respect to φ and (c) is proved. □
In this paper, we establish some generalizations of Smarzewski’s fixed point theorem for reactive firmly nonexpansive mappings and some new fixed point theorems which are original and quite different from the wellknown results in the literature.
2 New generalizations of Smarzewski’s fixed point theorem and applications to fixed point theory
In this section, we first establish a new fixed point theorem for reactive firmly nonexpansive mappings which is generalized Smarzewski’s fixed point theorem. We assume 0<\phi (s,t)<1 for all (s,t)\in C\times C in the following main theorem.
Theorem 2.1 Let X be a strictly convex Banach space with its zero vector θ and C={\bigcup}_{k=1}^{n}{C}_{k} be a finite union of nonempty weakly compact convex subsets {C}_{k} of X. Let \phi :C\times C\to (0,1) be a function and T:C\to C be a mapping. Suppose that

(a)
X has the (UAC)property,

(b)
T is reactive firmly nonexpansive with respect to φ.
Then T has a fixed point in C.
Proof Let z\in C be given. Since C is Tinvariant and C is bounded, the sequence {\{{T}^{j}z\}}_{j=1}^{\mathrm{\infty}}\subset C is bounded. Define the functional r:X\to [0,\mathrm{\infty}), the asymptotic radius of \{{T}^{j}z\} at x\in X, by
For each 1\le k\le n, let the number
and the set
be respectively the asymptotic radius and the asymptotic center of the sequence \{{T}^{j}z\} with respect to {C}_{k}. Since X has the (UAC)property, let {x}_{k}\in {C}_{k} be the unique asymptotic center of \{{T}^{j}z\} with respect to {C}_{k} for 1\le k\le n. So
For any k and j, since T is nonexpansive, we have
which implies
Let
Clearly, m<\mathrm{\infty}. For arbitrary x\in C={\bigcup}_{k=1}^{n}{C}_{k}, x\in {C}_{{k}_{x}} for some {k}_{x}, 1\le {k}_{x}\le n. Thus we have
Taking the infimum for x over C yields r(C,\{{T}^{j}z\})\ge m. Conversely, suppose r({x}_{{k}_{0}})=m for some {k}_{0}, 1\le {k}_{0}\le n. Then
By (2.2) and (2.3), we prove
Put
It is obvious that \mathrm{\u266f}(L)\le n, where \mathrm{\u266f}(L) is the cardinal number of L. We claim that for u\in C with r(u)=m if and only if u\in L. Indeed, it suffices to show that if u\in C with r(u)=m, then u\in L. Let u\in C={\bigcup}_{k=1}^{n}{C}_{k}. Then there is some {k}_{u}, 1\le {k}_{u}\le n, such that u\in {C}_{{k}_{u}}. So we obtain
and hence
By the uniqueness of asymptotic center {x}_{{k}_{u}}, we have u={x}_{{k}_{u}}, which means that u\in L.
Next, we will prove TL\subset L (i.e., L is Tinvariant). For any u\in L, since r(u)=m, from (2.1) and (2.4) it follows that
and hence r(Tu)=m. According to our claim, we get Tu\in L, which completes the assertion. Now, we show that T has a fixed point in L. Suppose to the contrary that T has no fixed point in L, that is, Tu\ne u for all u\in L. Then T{x}_{k}\notin {C}_{k} for all {x}_{k}\in L. Indeed, if T{x}_{k}\in {C}_{k} for some {x}_{k}\in L, since
and by the uniqueness of asymptotic center {x}_{k}, we obtain T{x}_{k}={x}_{k}. This means that {x}_{k} is a fixed point for T, contradicting our assumption, hence T{x}_{k}\notin {C}_{k} for all {x}_{k}\in L. For TL\subset L, there exists w\in L such that {T}^{\alpha}w=w for some 1\le \alpha \le \mathrm{\u266f}(L). Since
it follows that
Since T is reactive firmly nonexpansive with respect to φ, we have
for 1\le i\le \alpha, where {T}^{0}=I (the identity mapping). Therefore, in view of strict convexity of the norm, there is {t}_{i}>0 such that
From (2.5) and (2.6), we have
which implies that {t}_{i}=1 for all i. So, by (2.6) again, we get
Since
we obtain
which implies v=\theta, contradicting the fact that v\ne \theta. Therefore T must have a fixed point in L\subset C and this completes the proof. □
The following results are immediate consequences of Theorem 2.1.
Corollary 2.1 Let X be a strictly convex Banach space with its zero vector θ and C={\bigcup}_{k=1}^{n}{C}_{k} be a finite union of nonempty weakly compact convex subsets {C}_{k} of X. Let T:C\to C be a mapping. Suppose that

(a)
X has the (UAC)property,

(b)
T is λfirmly nonexpansive for some \lambda \in (0,1).
Then T has a fixed point in C.
Corollary 2.2 Let C={\bigcup}_{k=1}^{n}{C}_{k} be a finite union of nonempty weakly compact convex subsets {C}_{k} of a UCED Banach space X and \phi :C\times C\to (0,1) be a function. If T:C\to C is T is reactive firmly nonexpansive with respect to φ, then T has a fixed point in C.
Corollary 2.3 [[8], Theorem 2.8]
Let C={\bigcup}_{k=1}^{n}{C}_{k} be a finite union of nonempty weakly compact convex subsets {C}_{k} of a UCED Banach space X. If T:C\to C is λfirmly nonexpansive for some \lambda \in (0,1), then T has a fixed point in C.
Remark 2.1 Theorem 2.1 and Corollaries 2.1 and 2.2 all generalize and improve Smarzewski’s fixed point theorem, [[8], Theorem 2.8] and [[14], Theorems 2.8, 2.9].
In Theorem 2.1, if {C}_{1}={C}_{2}=\cdots ={C}_{n}:=C, then we obtain the following new fixed point theorem.
Theorem 2.2 Let X be a strictly convex Banach space with its zero vector θ and C be a nonempty weakly compact convex subset of X. Let \phi :C\times C\to (0,1) be a function and T:C\to C be a mapping. Suppose that

(a)
X has the (UAC)property,

(b)
T is reactive firmly nonexpansive with respect to φ.
Then T has a fixed point in C.
The following results are immediate from Theorem 2.2.
Corollary 2.4 Let X be a strictly convex Banach space with its zero vector θ and C be a nonempty weakly compact convex subset of X. Let \phi :C\times C\to (0,1) be a function and T:C\to C be a mapping. Suppose that

(a)
X has the (UAC)property,

(b)
T is λfirmly nonexpansive for some \lambda \in (0,1).
Then T has a fixed point in C.
Corollary 2.5 Let C be a nonempty weakly compact convex subset of a UCED Banach space X. Let \phi :C\times C\to (0,1) be a function. If T:C\to C is reactive firmly nonexpansive with respect to φ, then T has a fixed point in C.
Corollary 2.6 Let C be a nonempty weakly compact convex subset of a UCED Banach space X. If T:C\to C is λfirmly nonexpansive for some \lambda \in (0,1), then T has a fixed point in C.
Definition 2.1 Let C be a nonempty subset of a normed space (X,\parallel \cdot \parallel ) and \alpha ,\beta \in (0,1). A mapping T:C\to X is said to be (\alpha ,\beta )firmly nonexpansive if
Finally, by applying Theorem 2.1, we give some new fixed point theorems for (\alpha ,\beta )firmly nonexpansive mappings.
Theorem 2.3 Let X be a strictly convex Banach space with its zero vector θ and C={\bigcup}_{k=1}^{n}{C}_{k} be a finite union of nonempty weakly compact convex subsets {C}_{k} of X. Let T:C\to C be a mapping and α, β be positive real numbers satisfying 0<(1\alpha )(1\beta )<1. Suppose that

(a)
X has the (UAC)property,

(b)
T is (\alpha ,\beta )firmly nonexpansive.
Then T has a fixed point in C.
Proof Since \alpha ,\beta >0 and 0<(1\alpha )(1\beta )<1, we have \alpha +\beta \alpha \beta \in (0,1). Define \phi :C\times C\to (0,1) by
Due to T is (\alpha ,\beta )firmly nonexpansive, we obtain
for all x,y\in C. So, T is reactive firmly nonexpansive with respect to φ. Therefore the conclusion follows from Theorem 2.1. □
Corollary 2.7 Let C be a nonempty weakly compact convex subset of a UCED Banach space X. Let α, β be positive real numbers satisfying 0<(1\alpha )(1\beta )<1. If T:C\to C is (\alpha ,\beta )firmly nonexpansive, then T has a fixed point in C.
Corollary 2.8 Let C be a nonempty weakly compact convex subset of a uniformly convex Banach space X. Let α, β be positive real numbers satisfying 0<(1\alpha )(1\beta )<1. If T:C\to C is (\alpha ,\beta )firmly nonexpansive, then T has a fixed point in C.
References
Clarkson JA: Uniformly convex spaces. Trans. Am. Math. Soc. 1936, 40: 296–414.
Goebel K, Reich S: Uniform Convexity, Hyperbolic Geometry, and Nonexpansive Mappings. Dekker, New York; 1984.
Goebel K, Kirk WA: Topics in Metric Fixed Point Theory. Cambridge University Press, Cambridge; 1990.
Megginson RE: An Introduction to Banach Space Theory. Springer, New York; 1998.
Takahashi W: Nonlinear Functional Analysis. Fixed Point Theory and Its Applications. Yokohama Publishers, Yokohama; 2000.
Khamsi MA, Kirk WA: An Introduction to Metric Spaces and Fixed Point Theory. Wiley, New York; 2001.
Day MM, James RC, Swaminathan S: Normed linear spaces that are uniformly convex in every direction. Can. J. Math. 1971, 23: 1051–1059. 10.4153/CJM19711095
Du WS, Huang YY, Yen CL:Fixed point theorems for nonexpansive mappings on nonconvex sets in UCED Banach spaces. Int. J. Math. Math. Sci. 2002, 31: 251–257. 10.1155/S0161171202107113
Bruck RE: Nonexpansive projections on subsets of Banach spaces. Pac. J. Math. 1973, 47: 341–355. 10.2140/pjm.1973.47.341
Browder FE: Nonexpansive nonlinear operators in a Banach space. Proc. Natl. Acad. Sci. USA 1965, 54: 1041–1044. 10.1073/pnas.54.4.1041
Kirk WA: A fixed point theorem for mappings which do not increase distances. Am. Math. Mon. 1965, 72: 1004–1006. 10.2307/2313345
Göhde D: Zum Prinzip der kontraktiven Abbildung. Math. Nachr. 1965, 30: 251–258. (in German) 10.1002/mana.19650300312
Smarzewski R: On firmly nonexpansive mappings. Proc. Am. Math. Soc. 1991, 113: 723–725. 10.1090/S00029939199110500231
Kaczor W: Fixed points of asymptotically regular nonexpansive mappings on nonconvex sets. Abstr. Appl. Anal. 2003,2003(2):83–91. 10.1155/S1085337503205054
Acknowledgements
The author was supported by Grant No. MOST 1032115M017001 of the Ministry of Science and Technology of the Republic of China.
Author information
Authors and Affiliations
Corresponding author
Additional information
Competing interests
The author declares that he has no competing interests.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0), which permits use, duplication, adaptation, distribution, and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Cite this article
Du, WS. On new generalizations of Smarzewski’s fixed point theorem. J Inequal Appl 2014, 493 (2014). https://doi.org/10.1186/1029242X2014493
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/1029242X2014493