1. Introduction
Following the famous theorem known as the Contraction Mapping Principle, established by Banach in 1922, fixed point theory has become a fundamental area of research in modern mathematical analysis due to its wide range of applications in nonlinear analysis, differential equations, topology, optimization, computer science, etc. This foundational result has served as a basis for a wide range of extensions and generalizations in fixed point theory. In particular, various authors have pursued different research directions by relaxing the contraction conditions, modifying the completeness assumptions, or by introducing more general structures, such as partial metric spaces, fuzzy metric spaces, and spaces endowed with graph structures. These efforts have broadened the applicability of the original result to more abstract or complex mathematical settings.
Fixed point theorems, particularly generalizations developed for multivalued mappings defined on two different metric spaces, are powerful tools in modeling real-world systems across various disciplines. Such results can serve as a cornerstone in many applications, ranging from differential equations to artificial intelligence, game theory, and network analysis.
Fisher [
1,
2] established a remarkable result concerning the fixed points of compositions of two mappings on two complete metric spaces, providing a relationship between the fixed points of these mappings. Since then, numerous researchers have extended this result in various directions, exploring different types of theorem for two or more mappings. Examples of such works include [
2,
3,
4,
5,
6,
7,
8,
9] and others.
In 2000, Fisher and Türkoğlu [
10], by considering the multivalued version of the related fixed point theorem for single-valued mappings in [
2], established several related fixed point theorems for multivalued mappings on two complete and compact metric spaces. In addition, Chourasia and Fisher [
11], Jain and Fisher [
12], Popa [
13], Rohen and Murthy [
14], and Biçer et al. [
15] have also proved some related fixed point theorems for multivalued mappings under certain contractive conditions.
In this paper, by considering the multivalued version of the related fixed point theorem for single-valued mappings, we present a new type of related fixed point theorem for multivalued mappings in two related complete metric spaces.
The following are some characteristics of the present work.
The concept of the related orbital completeness of two metric spaces for multivalued mappings is introduced.
A new related fixed point theorem for multivalued mappings is established.
While in existing related fixed point theorems for multivalued mappings, at least two or more contraction conditions are used, in the present main theorem, only one contraction condition is used.
Unlike other existing theorems for multivalued mappings, only the classical metric is used in the contraction conditions in the main results presented.
A multivalued version of the Bollenbacher and Hicks’s result [
16] is obtained as a corollary of the present main theorem.
A single-valued version of the present main theorem is obtained like a simple corollary.
Two illustrative examples are given.
2. Preliminaries
Let be a metric space. Throughout the paper we denote by the family of all nonempty subsets of Z, by the family of all nonempty closed subsets of Z, and by the family of all nonempty closed bounded subsets of Z.
Let and let F be a mapping of Z into . We shall use the following definitions.
Definition 1 ([
17])
. An orbit of the multivalued mapping F at the point of is a sequence Definition 2 ([
17])
. A metric space is said to be F-orbitally complete if every Cauchy sequence of the form converges in Z. Definition 3. Let and be two metric spaces. Let F be a mapping of Z into and G be a mapping of Y into . Let and
Consider the following sets,Then, the metric spaces and are called related -orbitally complete for if every Cauchy sequence in and converges to a point in Z and converges to a point in Y, respectively. Note that the two metric spaces and are related -orbitally complete. However, the related -orbitally complete and metric spaces are not necessarily complete, as is shown by the following example.
Example 1. Let with the Euclidean metric ρ. Let the mappings be defined by for all . Then, for and , we have Therefore, and are related -orbitally complete for , but is not complete.
In a recent paper [
18], Romaguera introduced the definition of 0-lower semicontinuity as a generalization of lower semicontinuity.
Similarly, we use the following definition.
Definition 4. Let and be two metric spaces. We shall say that a real-valued function is -orbitally 0-lower semicontinuous (briefly 0-lsc) at with respect to if and , then , where and are sequences in and , respectively.
Definition 5. Let F be a mapping of Z into and G be a mapping of Y into . Then, the composition of the mappings F and G is defined by 3. Main Results
In this section, firstly we give the following related fixed point theorem in two related orbitally complete metric spaces, and .
Theorem 1. Let and be two metric spaces and let F be a mapping of Z into and G be a mapping of Y into . Suppose there exist and such thatfor all and , where . If and are related -orbitally complete for some , then we have - (a)
There exist two sequences in and in such that - (b)
for all ,
- (c)
If F is a mapping of Z into and G is a mapping of Y into , then the following statements are equivalent:
- (i)
and .
- (ii)
, and , are -orbitally 0-lsc at with respect to , where and .
- (iii)
and .
Further, if and , then and .
Proof. Suppose that
and
are related
-orbitally complete for
and
. Then, from Inequality (
2), there exist
and
such that
Similarly, there exist
and
such that
and continuing in this way, we obtain two sequences
in
and
in
such that
and
, and
for all
.
Now we shall show that the sequences and are the Cauchy sequences.
Using Inequality (
3), we get
Therefore,
is bounded and also non-decreasing. Thus,
is convergent. Let
be any two positive integers with
. From the triangle inequality property of the metrics
and
, we have
Since
is convergent, for any
, we can choose a positive integer
such that
for all
. Thus, based on Inequality (
4), we get
for all
, and thus
and
are two Cauchy sequences in
and
, respectively. Since
and
are related
-orbitally complete, the sequence
has a limit
u in
Z and the sequence
has a limit
v in
Y. Thus, the proof of (a) is complete.
To prove (b), let
. Then, from Inequalities (
2) and (
3), we get
Letting
m tend to infinity, it follows that
Thus the proof of (b) is complete.
Now suppose that F is a mapping of Z into and G is a mapping of Y into .
(i) ⇒ (ii): Assume that
and
. Clearly,
Let
,
be two sequences in
and
, respectively, with
,
. Then, we get
and
and so
and
be
-orbitally 0-lsc at
with respect to
since
(ii) ⇒ (iii): Let
and
are
-orbitally 0-lsc at
with respect to
. From (a), there exist two sequences
in
and
in
such that
. We also have
and
Since
and
are
-orbitally 0-lsc at
with respect to
,
(iii) ⇒ (i): Now let . Then we have . Since is a closed subset of Z, , and so . Similarly, if , then .
We now assume that and . Then, , since . Therefore, . Similarly, , since and so , which completes the proof. □
Note that since the inequality
holds for all
and
, we obtain the following result.
Corollary 1. Let and be two metric spaces and let F be a mapping of Z into and G be a mapping of Y into . Suppose there exist and such thatfor all and , where . If and are related -orbitally complete for some , then - (a)
There exist two sequences in and in such that - (b)
for all ,
- (c)
If F is a mapping of Z into and G is a mapping of Y into , then the following statements are equivalent:
- (i)
and .
- (ii)
, and , are -orbitally 0-lsc at with respect to , where and .
- (iii)
and .
Further, if and , then and .
Proof. We have
Then the results (a) and (c) follow immediately from Theorem 1.
To prove (b), let
. Similarly to in the proof of (b) in Theorem 1, using Inequality (
5), we get
Letting
m tend to infinity, it follows that
□
If we let
where
, then from Corollary 1, we have the following multivalued version of Bollenbacher and Hicks’s result [
16], which is a version of Caristi’s famous fixed point theorem [
19].
Corollary 2. Let be a metric space and let T be a mapping of Z into . Suppose there exists such thatfor each , where . If is T-orbitally complete for some , then - (a)
There exists a sequence in such that ,
- (b)
for all ,
- (c)
If T is a mapping of Z into , then the following statements are equivalent:
- (i)
.
- (ii)
, is T-orbitally 0-lsc at u with respect to , where .
- (iii)
.
Corollary 3. Let be a metric space and let T be a mapping of Z into . Suppose there exists such thatfor each and for all , where . If is T-orbitally complete for some , then - (a)
There exists a sequence in such that ,
- (b)
for all ,
- (c)
The following statements are equivalent:
- (i)
.
- (ii)
, is T-orbitally 0-lsc at u with respect to , where .
- (iii)
.
Proof. Define the function
on
Z by
. Since
is bounded,
is a mapping of
Z into
. Then, from Inequality (
7), we get
Thus, from Inequality (
8), we have
and
and so
Hence, the results follow, since all the conditions of Corollary 2 are satisfied. □
We need the following definition for the next corollary.
Definition 6. If we let F be a single-valued mapping f of Z into Y and G be a single-valued mapping g of Y into Z, then from (1) we getwhere and . Then, the metric spaces and are called related -orbitally complete for if every Cauchy sequence in and converges to a point in Z and converges to a point in Y, respectively.
We finally give the following corollary for single-valued mappings.
Corollary 4. Let and be two metric spaces and let f be a mapping of Z into Y and g be a mapping of Y into Z satisfying the inequalityfor all and , where . If and are related -orbitally complete for some , then - (a)
for and , exist.
- (b)
for all ,
- (c)
and if and only if , and , are -orbitally 0-lsc at with respect to .
Further, if and , then and .
Proof. Define two mappings
f of
Z into
and
g of
Y into
by putting
for all
and
for all
, respectively. It follows that
F and
G satisfy Inequality (
2). Then, the results (a) and (b) follow, since all the conditions of Theorem 1 are satisfied.
Now we prove (c). Suppose that
,
and
and
are sequences in
and
, respectively, with
. Then, we get
and also
.
Similarly we have and . Thus, and are 0-lsc at .
Now
are 0-lsc at
and let
. It follows from (a) that
and
. Then,
Since
are 0-lsc at
, we have
and
and so
and
. □
4. Examples
We finally give two examples which support our main result.
Example 2. Let and with the Euclidean metrics ρ and ϱ, respectively. Define the mappings and byfor all and for all . Then, for and , we haveTherefore, and are related -orbitally complete. If and are taken for each and , then we getwhere . Thus, Inequality (2) is satisfied. The sequencesin and converge to 0. Also, and and so and . Example 3. Let and with the Euclidean metrics ρ and ϱ, respectively. Define the mappings and byfor all and for all . If and are taken for each and , then we getwhere . Thus, Inequality (2) is satisfied. Take the points and . If we choose as and as , then we obtain the following sequences in and , respectively.and so the sequences converge to 0. Also, and and so and . Note that the closedness of the set and , for all and for all , is a necessary condition in (c) of Theoremn 1. For example, if we take in the example above, then .
5. Conclusions
In this research article, by introducing the concept of the related orbital completeness of two metric spaces for multivalued mappings, a theorem concerning the fixed points of the compositions of two multivalued mappings defined on two orbitally complete metric spaces is presented. The relationship between the fixed points of these mappings is also investigated. Some important results are derived as corollaries of the main theorem. Finally, two concrete examples are provided to illustrate the significance of the main result.
We believe that the introduced concept and the corresponding fixed point theorem open up new directions for future research, including potential extensions to more general spaces, other types of contractions, and applications in various mathematical and applied fields. In particular, it would be interesting to investigate how the findings of this paper behave in the setting of b-metric spaces, as considered in [
20].