The present work is about the existence of best proximity points for Prešić type nonself operators in b-metric spaces. In order to elaborate the results an example is presented. Moreover, some interesting formulations of Prešić fixed point results are also established. In addition a result in double controlled metric type space is also formulated.
Citation: Samina Batul, Dur-e-Shehwar Sagheer, Hassen Aydi, Aiman Mukheimer, Suhad Subhi Aiadi. Best proximity point results for Prešić type nonself operators in b-metric spaces[J]. AIMS Mathematics, 2022, 7(6): 10711-10730. doi: 10.3934/math.2022598
[1] | Müzeyyen Sangurlu Sezen . Interpolative best proximity point results via $ \mathbf{\gamma } $-contraction with applications. AIMS Mathematics, 2025, 10(1): 1350-1366. doi: 10.3934/math.2025062 |
[2] | Umar Ishtiaq, Fahad Jahangeer, Doha A. Kattan, Manuel De la Sen . Generalized common best proximity point results in fuzzy multiplicative metric spaces. AIMS Mathematics, 2023, 8(11): 25454-25476. doi: 10.3934/math.20231299 |
[3] | Naeem Saleem, Hüseyin Işık, Sana Khaleeq, Choonkil Park . Interpolative Ćirić-Reich-Rus-type best proximity point results with applications. AIMS Mathematics, 2022, 7(6): 9731-9747. doi: 10.3934/math.2022542 |
[4] | Arshad Ali Khan, Basit Ali, Talat Nazir, Manuel de la Sen . Completeness of metric spaces and existence of best proximity points. AIMS Mathematics, 2022, 7(5): 7318-7336. doi: 10.3934/math.2022408 |
[5] | Basit Ali, Muzammil Ali, Azhar Hussain, Reny George, Talat Nazir . Best proximity points in non-Archimedean fuzzy metric spaces with application to domain of words. AIMS Mathematics, 2022, 7(9): 16590-16611. doi: 10.3934/math.2022909 |
[6] | Seher Sultan Yeşilkaya, Cafer Aydın, Adem Eroǧlu . Fixed point results on ordered Prešić type mappings. AIMS Mathematics, 2020, 5(5): 5140-5156. doi: 10.3934/math.2020330 |
[7] | Khalil Javed, Muhammad Nazam, Fahad Jahangeer, Muhammad Arshad, Manuel De La Sen . A new approach to generalized interpolative proximal contractions in non archimedean fuzzy metric spaces. AIMS Mathematics, 2023, 8(2): 2891-2909. doi: 10.3934/math.2023151 |
[8] | Abdullah Shoaib, Poom Kumam, Shaif Saleh Alshoraify, Muhammad Arshad . Fixed point results in double controlled quasi metric type spaces. AIMS Mathematics, 2021, 6(2): 1851-1864. doi: 10.3934/math.2021112 |
[9] | Arshad Ali Khan, Basit Ali, Reny George . On semi best proximity points for multivalued mappings in quasi metric spaces. AIMS Mathematics, 2023, 8(10): 23835-23849. doi: 10.3934/math.20231215 |
[10] | Mustafa Aslantas, Hakan Sahin, Raghad Jabbar Sabir Al-Okbi . Some best proximity point results on best orbitally complete quasi metric spaces. AIMS Mathematics, 2023, 8(4): 7967-7980. doi: 10.3934/math.2023401 |
The present work is about the existence of best proximity points for Prešić type nonself operators in b-metric spaces. In order to elaborate the results an example is presented. Moreover, some interesting formulations of Prešić fixed point results are also established. In addition a result in double controlled metric type space is also formulated.
In 1922, Banach proved his famous result known as the Banach contraction principle, which is a simple and powerful result with a wide range of applications [12]. Many generalizations of Banach contraction principle can be seen in the literature, see e.g., [5,9,16,17,19,21,22,23,24].
Consider the kth order nonlinear difference equation
xn=f(xn−1,xn−2,…,xn−k),n=k,k+1,… | (1.1) |
with initial values x0,…,xk−1∈X, where k≥1 is a positive integer and f:Xk→X. This difference equation can be discussed with the perspective of fixed point theory by considering the fact that x∗ is a fixed point of f if and only if it is the solution of (1.1) exist, that is,
x∗=f(x∗,x∗,…,x∗). |
The first step in this direction is taken by Preˆsiˊc in 1965 by establishing a generalization of the Banach contraction principle in the following manner:
Theorem 1.1. [28] Let (X,d) be a complete metric space. Given k≥1 a positive integer and f:Xk→X. Assume also that
d(f(x0,…,xk−1),f(x1,…xk))≤k∑i=0ηid(xi−1,xi),∀x0,…,xk−1∈X, |
where η1,η2,…,ηk are positive constants such that k∑i=1ηi∈(0,1). Then there exists a unique x∗∈X such that x∗=f(x∗,x∗,…x∗), that is, f has a unique fixed point x∗∈X. Moreover, for any initial values x0,…,xk−1∈X the iterative sequence given by (6.1) converges to x∗.
Note that for k=1, the map f:X→X becomes a self map and hence the above Theorem is the generalization of Banach contraction principle (for contractions defined on Xk). In [14], Theorem 1.1 is generalized by ˊCiriˊc and Preˆsiˊc in the following way:
Theorem 1.2. [14]Let (X,d) be a complete metric space. Given k≥1 a positive integer and f:Xk→X. Suppose that
d(f(x0,…,xk−1),f(x1,…xk))≤μmax{d(x0,x1),…,d(xk−1,xk)},∀x0,…,xk−1∈X, |
where μ∈(0,1) is a constant. Then there exists a unique x∗∈X such that x∗=f(x∗,x∗,…x∗), that is, f has a unique fixed point x∗∈X. Moreover, for any initial values x0,…,xk−1∈X, the iterative sequence given by (6.1) converges to x∗.
Note that a fixed point of the operator f:Xk→X can be considered as the equilibrium point of the kth order nonlinear difference Eq (6.1). Therefore, the above theorems can be taken as a tool to ensure the existence and uniqueness of the kth order nonlinear difference equation. Some other generalizations are obtained by Pˆacurar in [13,27]. Recently, Ali et al. [3] studied the existence of an approximate solution of the equation x=f(x,x,…,x), where f:Hk→K. This equation has a solution if H and K have some common element, but has no solution otherwise. Hence in that case we can only get the approximate solution of the equation. The approximate solution of x=f(x,x,…,x), with the error term d(H,K) is called a best proximity point of f:Hk→K. The classical result of approximation theory given by Fan [18] is a great source of inspiration for various researchers in study of approximate solutions of x=f(x). This result is given as follows:
Theorem 1.3. Let H be a nonempty compact convex subset of a normed linear space X and f:H→X be a continuous function. Then there exists x∈X such that
‖x−f(x)‖=infa∈A{‖f(x)−a‖}. |
Recently Altun et al. [7,8] investigated certain best proximity points results on KW-type nonlinear contractions and fractals. Furthermore Ali et al. [3] used the metric space (X,d) endowed with a graph and proved some best proximity results. These results are the generalizations of already existing results which are stated earlier.
Czerwik [15] gave a generalization of the famous Banach fixed point theorem in so-called b-metric spaces. For some important results on b-metric spaces, we refer the reader to [2,4,10,11,25,26].
The purpose of present research is to extend the results of Ali et al.[3], in the setting of b-metric spaces equipped with an order. Hence, many results in literature become special cases of results presented in this article. Our paper also contains some examples for the validation of presented results and an application for further authentication.
We include the following definitions before giving the main results.
Definition 2.1. [6] Consider a metric space (X,d). Suppose H and K are two non-empty subsets of X. An element x∈H is said to be a best proximity point of the mapping T:H→K if
d(x,Tx)=d(H,K). |
Remark 2.1. From the above definition, it is obvious that a best proximity point reduces to a fixed point for self-mappings.
Basha and Shahzad [29] have presented the following definition:
Definition 2.2. Consider a complete metric space (X,d). Suppose that H,K are non empty subsets of X. If each sequence {kn} in K with d(h,kn)→d(h,K), for some h∈H, has a convergent subsequence. Then, K is called approximately compact with respect to H.
Ali et al. [3] introduced path admissible mappings as follows:
Definition 2.3. Suppose that H,K are nonempty subsets of a metric space (X,d) endowed with a binary relation R. Then T:H×H→K is said to be path admissible, if
{d(w1,T(h1,h2))=d(H,K),d(w2,T(h2,h3))=d(H,K),h1Ph3,⇒w1Rw2 |
where h1,h2,h3,w1,w2∈H.
Here, by w1Rw2 we mean that w1 and w2 are related with each other under the binary relation R and h1Ph3, we mean that for above mentioned h1,h2,h3∈Xwehaveh1Rh2 and h2Rh3.
Definition 2.4. Suppose H,K are non empty subsets of a metric space (X,d). An element h∗∈H is said to be a best proximity point of T:H×H→K if
d(h∗,T(h∗,h∗))=d(H,K), | (2.1) |
where
d(H,K)=inf{d(h,k):h∈H,k∈K}. |
First, we recall some definitions which are used in the sequel. Let (X,db) be a b-metric space with coefficient b≥1. Suppose that H and K are two nonempty subsets of X, then define the following sets:
db(H,K)=inf{db(h,k):h∈H,k∈K},db(x0,K)=inf{db(x0,k):k∈K},H0={h∈H:db(h,k)=db(H,K)forsomek∈K},K0={k∈K:db(h,k)=db(H,K)forsomeh∈H}. |
Definition 3.1. Consider a b-metric space (X,db) with coefficient b≥1. Suppose that H and K are nonempty subsets of X. The element h∗∈H is said to be a best proximity point of the mapping T:H→K if
db(h∗,T(h∗))=db(H,K). | (3.1) |
Definition 3.2. Consider a b-metric space (X,db) with coefficient b≥1 and let H and K be two nonempty subsets of X. Then K is said to be approximately compact with respect to H, if each sequence {kn}⊆K with db(h,kn)→db(h,K) for some h∈H, has a convergent subsequence.
Definition 3.3. Let (X,db) be a b-metric space with coefficient b≥1 and R is the binary relation on X. Suppose H, K are nonempty subsets of X. A mapping T:H×H→K is called path admissible, whenever ∀h1,h2,h3,w1,w2∈H we have
{db(w1,T(h1,h2))=db(H,K),db(w2,T(h2,h3))=db(H,K),h1Ph3,⇒w1Rw2, |
here, by w1Rw2 mean that w1 and w2 are related with each other under the binary relation R and h1Ph3 we mean that for above mentioned h1,h2,h3∈H, we have h1Rh2 and h2Rh3.
Theorem 3.1. Suppose that (X,db) is a complete b-metric space with coefficient b≥1 endowed with a binary relation R, where db is a continuous functional. Assume that H and K are nonempty closed subsets of X. Consider a mapping T:H×H→K such that for each h1,h2,h3,w1,w2∈H with h1Ph3 that is h1Rh2,h2Rh3 and db(w1,T(h1,h2))=db(H,K)=db(w2,T(h2,h3)), we have:
db(w1,w2)≤Γmax{db(h1,h2),db(h2,h3)}, | (3.2) |
where Γ∈[0,1) such that bΓ<1. Furthermore, suppose that the subsequent conditions are true:
(1) T is path admissible;
(2) ∃h0,h1,h2∈H which satisfy db(h2,T(h0,h1))=db(H,K) and h0Ph2;
(3) T(H×H0)⊆K0;
(4) K is approximately compact with respect to H;
(5) If {hj}⊆X such that hjPhj+2 for each j∈N and hj→x∗ as j→∞, then hjRx∗ for all j∈N and x∗Rx∗.
Then T has a best proximity point.
Proof. Using condition (ii), we have h0,h1,h2∈H satisfying
db(h2,T(h0,h1)=db(H,K),andh0Ph2, |
that is, h0Rh1,h1Rh2. From condition (iv), T(h1,h2)∈K0, and by the definition of K0, we have h3∈H which satisfies
db(h3,T(h1,h2))=db(H,K). |
Due to condition (i), we have h2Rh3. Hence, h1Ph3. By continuing same process, we build a sequence {hj}≥2⊆H which satisfies
db(hj+1,T(hj−1,hj))=db(H,K)foreachj∈N, | (3.3) |
and hj−1Phj+1. That is, hj−1Rhj,hjRhj+1∀j∈N. From (3.2), we have
db(hj,hj+1)≤Γmax{db(hj−2,hj−1),db(hj−1,hj)}foreachj=2,3,4,…. | (3.4) |
For convenience, we take cj=db(hj,hj+1) for each j∈N∪{0}. Then we can rewrite (3.4) as
cj≤Γmax{cj−2,cj−1}foreachj=2,3,4,…. |
By using induction, we can get cn−1≤Zψnwhereψ=Γ1/2. It is obviously true for j=0,1 by considering
Z=max{c0/ψ,c1/ψ2}, |
since Z is max{c0/ψ,c1/ψ2}, one writes
c0≤Zψandc1≤Zψ2. |
We obtain
c2≤Γmax{c0,c1}≤Γmax{Zψ,Zψ2}≤ΓZψ=Zψ3,⋮cj≤Γmax{cj−1,cj−2}≤Γmax{Zψj,Zψj−1}≤ΓZψj−1=Zψj+1. |
Therefore, we have
cj−1≤Zψj∀j∈N. |
Hence,
db(hj−1,hj)≤Zψj∀j∈N. | (3.5) |
By using triangle inequality, we get
db(hj,hj+q)≤b{db(hj,hj+1)+db(hj+1,hj+q)},=bdb(hj,hj+1)+bdb(hj+1,hj+q),≤bdb(hj,hj+1)+bb{db(hj+1,hj+2)+db(hj+2,hj+q)},=bdb(hj,hj+1)+b2{db(hj+1,hj+2)+db(hj+2,hj+q)},=bdb(hj,hj+1)+b2db(hj+1,hj+2)+b2db(hj+2,hj+q),≤bdb(hj,hj+1)+b2db(hj+1,hj+2)+⋯+bqdb(hj+q−1,hj+q),≤bZψj+1+b2Zψj+2+b3Zψj+3+⋯+bqZψj+q,≤bψj+1{1+bψ+b2ψ2+⋯+bq−1ψq−1}Z,≤1−(bψ)q1−bψ Zbψj+1,<bψj+11−bψ Z. |
Thus, {hj} is a Cauchy sequence in H, so there is an element h∗∈H such that hj→h∗ and hj∈H0 which satisfies
db(H,K)=db(h∗,T(hj−1,hj)), |
that is, hj−1Rh∗.
Furthermore, we have to prove that db(h∗,T(hj−1,hj))→db(h∗,K)asj→∞. Consider,
db(h∗,K)≤db(h∗,T(hj−1,hj))=limn→∞db(hj+1,T(hj−1,hj))=db(H,K)≤db(h∗,K) |
Therefore,
db(h∗,T(hj−1,hj))→db(h∗,K)asj→∞ | (3.6) |
Since T is approximately compact with respect to H, the sequence {T(hj−1,hj)} has a subsequence {T(hjm−1,hjm)}, which converges to a point k∗∈K. That is,
db(h∗,k∗)=limm→∞db(hjm+1,T(hjm−1,hjm)=db(H,K). |
Hence, h∗∈H0. As we know T(hj,h∗)∈K0, we have g∈H satisfying
db(g,T(hj,h∗))=db(H,K). | (3.7) |
By assumption (vi), we have hjRh∗ for all j∈N. Thus, we have
db(h∗,T(hj−1,hj))=db(H,K),anddb(g,T(hj,h∗))=db(H,K)∀j∈N. |
Hence, we get hj−1Ph∗. Also, hj−1Rhj, and hjRh∗ for all j∈N. Hence, from (6.2),
db(hj+1,g)≤Γmax{db(hj−1,hj),db(hj,h∗)}foreachj=2,3,4,…. |
Taking j→∞, we obtain db(h∗,g)=0,thatisg=h∗. Putting g=h∗ in (3.7), we have
db(h∗,T(hj,h∗))=db(H,K). |
That is, h∗Rh∗. Furthermore, we know that T(h∗,h∗)∈K0, and we have an element t∈H which satisfies
db(t,T(h∗,h∗))=db(H,K). | (3.8) |
Condition (vi) implies that h∗Rh∗. Hence,
db(t,T(h∗,h∗))=db(H,K),anddb(h∗,T(hj,h∗))=db(H,K)foreachj∈N. |
Therefore, hjPh∗foreachj∈N, that is, hjRh∗,h∗Rh∗ for each j∈N. Thus, from (3.2),
db(h∗,t)≤Γmax{db(hj,h∗),db(h∗,h∗)}foreachj∈N, |
Taking limit as j→∞, we have db(h∗,t)=0,thatist=h∗. Putting t=h∗ in (3.8), we have
db(h∗,T(h∗,h∗))=db(H,K). |
Theorem 3.2. Let H and K be nonempty subsets of a complete b-metric space (X,db) endowed with binary relation R, where b-metric is a continuous functional. Consider a mapping T:H×H→K such that for each h1,h2,h3,w1,w2∈H with h1Ph3, that is, h1Rh2,h2Rh3 and
db(w1,T(h1,h2))=db(H,K)=db(w2,T(h2,h3)), we have
db(h3,w2)≤Γmax{db(h1,h2),db(h2,w1)}, | (3.9) |
where Γ∈[0,1) such that bΓ<1.
Furthermore, suppose that the subsequent conditions are true
(1) T is path admissible;
(2) ∃h0,h1,h2∈H which satisfy db(h2,T(h0,h1))=db(H,K) and h0Ph2;
(3) T(H×H0)⊆K0;
(4) K is approximately compact with respect to H;
(5) When {hj}⊆X such that hjPhj+2 for each j∈N and hj→x∗ as n→∞, then hjRx∗ for all j∈N and x∗Rx∗.
Then there exists a point h∗∈H which satisfies
db(h∗,T(h∗,h∗))=db(H,K), |
that is, T has a best proximity point.
Proof. Proceeding as in Theorem 3.1, we obtain a sequence {hj:j∈N−1} in H0 satisfying
db(hj+1,T(hj−1,hj))=db(H,K)foreachj∈N, |
and hj−1Phj+1, that is hj−1Rhj,hjRhj+1∀j∈N.
From (3.9), we have
db(hj,hj+1)≤Γmax{db(hj−2,hj−1),db(hj−1,hj)}foreachj=2,3,4,…. |
Following the proof of Theorem 3.1 and above inequality, {hj} is a Cauchy sequence in H such that hj→h∗ and h∗∈H0. As T(hj,h∗)∈K0, we have w∈H satisfying
db(w,T(hj,h∗))=db(H,K). | (3.10) |
From assumption (vi), we get hjRh∗ for all j∈N. We already have
db(h∗,T(hj−1,hj))=db(H,K). |
Thus, we get hj−1Ph∗, that is hj−1Rhj and hjRh∗ for all j∈N. Hence, from (3.9), we get
db(h∗,w)≤Γmax{db(hj−1,hj),db(hj,hj+1)}foreachj∈N. |
Taking limit as j→∞ in above inequality, we get db(h∗,w)=0, that is, h∗=w. Using w=h∗ in (3.10),
db(h∗,T(hj,h∗))=db(H,K). |
Further, note that T(h∗,h∗)∈K0, and there is q∈H which satisfies
db(q,T(h∗,h∗))=db(H,K). |
Hypothesis (vi) implies h∗Rh∗. Hence, we have
db(h∗,T(hj,h∗))=db(H,K),anddb(q,T(h∗,h∗))=db(H,K), |
andhjPh∗,thatishjRh∗andh∗Rh∗. |
Thus, from (3.9),
db(h∗,q)≤Γmax{db(hj,h∗),d(h∗,h∗)}foreachj∈N. |
Letting j→∞, we have q=h∗. Thus, we have
db(h∗,T(h∗,h∗))=db(H,K). |
Example 3.1. Consider X=R2 endowed with the b-metric given by
db((s1,s2),(c1,c2))=|s1−c1|2+|s2−c2|2foreachs=(s1,s2),c=(c1,c2)∈R2. |
Define a binary relation R on R2 as sRc if and only if s1≤c1ands2≤c2. Take
H={(0,s):s∈[−2,2]},andK={(1,s):s∈[−2,2]}. |
Define
T:H×H→K,T((0,s),(0,c))=(1,c)∀(0,s),(0,c)∈H. |
Let ¯h1=(0,h1),¯h2=(0,h2),¯h3=(0,h3)∈[−2,2]. To find w1 and w2, we have
db(¯w1,T(¯h1,¯h2))=db(H,K)=db(¯w2,T(¯h2,¯h3)). | (3.11) |
For this, consider
db(H,K)=inf{db(¯h,¯k):¯h∈H,¯k∈K},=inf{db((0,s),(1,s)):wheres∈[−2,2]},=inf{∣0−1∣2+∣s−s∣2:wheres∈[−2,2]},=1. |
That is,
db(H,K)=1. | (3.12) |
db(¯w1,T(¯h1,¯h2)=db((0,w1),T((0,h1),(0,h2))),=db((0,w1),(1,h2)),=∣0−1∣2+∣w1−h2∣2,=1+(w1−h2)2. |
Then
db(¯w1,T(¯h1,¯h2))=1+(w1−h2)2. | (3.13) |
Using (3.12) and (3.13) in (3.11), we obtain
1=1+(w1−h2)2. |
That is,
w1=h2. |
Similarly,
db(¯w2,T(¯h2,¯h3))=db((0,w2),T((0,h2),(0,h3))),=db((0,w2),(1,h3)),=∣0−1∣2+∣w2−h3∣2,=1+(w2−h3)2. |
From (3.11), we obtain
w2=h3. |
¯w1=(0,w1)=(0,h2),¯w2=(0,w2)=(0,h3). |
Thus, ¯h1,¯h2,¯h3,¯w1,¯w2∈H, with ¯h1P¯h3.
Also, we have
db(¯h3,¯w2)≤Γmax{db(¯h1,¯h2),d(¯h2,¯w1)}, | (3.14) |
where
db(¯h3,¯w2)=db((0,h3),(0,w2)),=∣0−0∣+∣h3−w2∣2,=∣h3−h3∣2,=0. |
Using above equation in (3.14), we get
db(¯h3,¯w2)=0=ψmax{db(¯h1,¯h2),db(¯h2,¯w1)}. |
Here, we say ψ=Γ12=12∈[0,1). Now, we will prove condition (i) of Theorem 3.2. Consider
¯h1=(0,h1),¯h2=(0,h2),¯h3=(0,h3)∈Hsuchthat¯h1P¯h3. |
Since ¯w1=(0,w1)=(0,h2) and w2=(0,w2)=(0,h3), we now prove
db((0,w1),T((0,h1),(0,h2)))=db(H,K)anddb(H,K)=db((0,w2),T((0,h2),(0,h3))), |
d(¯w1,T(¯h1,¯h2))=db((0,w1),T((0,h1),(0,h2))),=db((0,h2),(1,h2)),=∣0−1∣2+∣h2−h2∣2,=1=db(H,K). |
Similarly,
db(¯w2,T(¯h2,¯h3))=db((0,w2),T((0,h2),(0,h3))),=db((0,h2),(1,h3)),=∣0−1∣2+∣h3−h3∣2,=1=db(H,K). |
This implies that ¯w1R¯w2. Thus, T is path admissible. Now, we will prove condition (ii):
db(h2,T(h0,h1))=db(H,K),andh0Ph2. |
We need to consider
¯h1=(0,0),¯h2=(0,12),¯h3=(0,58)∈H, |
such that
db((0,58),T((0,0),(0,12)))=db((0,58)−(1,0+12+24)),=∣(0−1∣2+∣58−58∣2,=1,=db(H,K), |
and (0,0)P(0,58). Moreover, assumption (v) holds, that is, hjPhj+2 for all j∈N, and hj→a as j→∞, then hjRa for each j∈N and aRa. Therefore, all axioms are true. Hence, T has a best proximity point.
Theorem 3.3. Let H and K be nonempty closed subsets of a complete b-metric space (X,db) with coefficient b≥1 endowed with a binary relation R, where the b-metric is a continuous functional. Consider a mapping T:H×H→K such that for each h1,h2,h3,w1,w2∈H with h1Ph3, that is, h1Rh2,h2Rh3, and
db(w1,T(h1,h2))=db(H,K)=db(w2,T(h2,h3)), we have
db(T(h2,w1),T(h3,w2))≤Γ{db(T(h1,h2),T(h2,h3))}, | (3.15) |
where Γ∈[0,1) such that bΓ<1.
Furthermore, suppose that the subsequent conditions are true
(1) T is path admissible;
(2) ∃h0,h1,h2∈H which satisfy db(h2,T(h0,h1))=db(H,K) and h0Ph2;
(3) T(H×H0)⊆K0;
(4) K is approximately compact with respect to H;
(5) if {hj} and {¯hj} are in X such that hj→h and ¯hj→¯h, then T(hj,¯hj)→T(h,¯h).
Then there exists a point h∗∈H so that
db(h∗,T(h∗,h∗))=db(H,K), |
that is, T has a best proximity point.
Proof. By using a similar argument as in Theorem 3.1, we build a sequence {hj≥2}⊆H which satisfies
db(hj+1,T(hj−1,hj))=db(H,K)∀j∈N, |
and hj−1Phj+1, that is, hj−1Rhj,hjRhj+1∀j∈N.
From (3.15), we have
db(T(hj−1,hj),T(hj,hj+1))≤Γmax{db(T(hj−2,hj−1)),db(T(hj−1,hj))}foreachj=2,3,4,…. |
Inductively, we get
db(T(hj−1,hj),T(hj,hj+1))≤Γj−1max{db(T(h0,h1),db(h1,h2))}. |
By using triangle inequality and above inequality for each j∈N, we have
db(T(hj,hj+1),T(hj+1,hj+p))≤j+p−1∑i=jdb(T(hi,hi+1),T(hi+1,hi+2)). |
This proves that T(hj−1,hj) is a Cauchy sequence in the closed subset K. Since X is complete, there exists k∗∈K such that T(hj−1,hj)→k∗.
Moreover,
db(h∗,K)≤db(h∗,T(hj−1,hj))=limn→∞db(hj+1,T(hj−1,hj))=db(H,K)≤db(h∗,K). |
Therefore, db(k∗,hj+1)→db(k∗,H) as j→∞.
Using hypothesis (v), {hj} has a subsequence {hjl} that converges to an element h∗⊆H such that
db(h∗,T(h∗,h∗))=liml→∞db(hjl+1,T(hjl−1,hjl))=db(H,K). |
Then
db(h∗,T(h∗,h∗))=db(H,K). |
Theorem 3.4. Let H and K be closed nonempty subsets of a complete b-metric space (X,db) endowed with a binary relation R where the b-metric is a continuous functional. Consider a mapping T:H×H→K such that for each h1,h2,h3,w1,w2∈H with h1Ph3, that is, h1Rh2 and h2Rh3, and
db(w1,T(h1,h2))=db(H,K)=db(w2,T(h2,h3)), we have
db(T(h2,w1),T(h3,w2))≤Γmax{db(T(h1,h2),T(h2,h3)),db(T(h2,h3),T(w1,w2))}, | (3.16) |
where Γ∈[0,1) such that bΓ<1.
Furthermore, suppose that the subsequent conditions are true:
(1) T is path admissible;
(2) ∃h0,h1,h2∈H which satisfy db(h2,T(h0,h1))=db(H,K) and h0Ph2;
(3) T(H×H0)⊆K0;
(4) K is approximately compact with respect to H;
(5) if {hj}, {¯hj} in X such that hj→h and ¯hj→¯h, then T(hj,¯hj)→T(h,¯h).
Then there exists a point h∗∈H which satisfies
d(h∗,T(h∗,h∗))=d(H,K), |
that is, T has a best proximity point.
Proof. Using the assumptions, we can build a sequence {hj}≥2 in H0 which satisfies
db(hj+1,T(hj−1,hj))=db(H,K)∀j∈N, | (3.17) |
and hj−1Phj+1, that is, hj−1Rhj and hjRhj+1 for all j∈N. From (3.16), we have
db(T(hj−1,hj),T(hj,hj+1))≤Γmax{db(T(hj−2,hj−1)),T(hj−1,hj)),db(T(hj−1,hj),T(hj,hj+1))}=db(T(hj−2,hj−1)),T(hj−1,hj))foreachj=2,3,…. |
Therefore, we have
db(Tj−1,T(hj)≤Γdb(Tj−2,Tj−1)foreachj=2,3,4,…. |
By using induction, we get
db(Tj−1,Tj)≤Γdb(Tj−2,Tj−1),≤Γ(Γdb(Tj−3,Tj−2)),=Γ2db(Tj−3,Tj−2),≤Γ2Γdb(Tj−4,Tj−3),=Γ3db(Tj−4,Tj−3),⋮≤Γj−1db(T1,T0)forj=2,3,4,…. |
Hence,
\begin{equation} d_b( T_{j}, T_{j+1}) \leq \Gamma^{j} d_b( T_{0}, T_{1})\; \; \; \mathrm{for}\; \; \; j = 1, 2, 3, \ldots . \end{equation} | (3.18) |
By using triangle inequality,
\begin{equation} \begin{split} d_b( T_{j}, T_{j+p}) \leq&\; b\lbrace d_b( T_{j}, T_{j+1}) + d_b( T_{j+1}, T_{j+p})\rbrace, \\ \; = &\; b d_b( T_{j}, T_{j+1}) +b d_b( T_{j+1}, T_{j+p}), \\ \leq&\; b d_b( T_{j}, T_{j+1}) +bb \lbrace d_b( T_{j+1}, T_{j+2})+ d_b( T_{j+2} T_{j+p})\rbrace, \\ = &\; b d_b( T_{j}, T_{j+1}) +b^{2} d_b( T_{j+1}, T_{j+2})+b^{2} d_b( T_{j+2} T_{j+p}), \\ \leq& \; b d_b( T_{j}, T_{j+1}) +b^{2} d_b( T_{j+1}, T_{j+2})+\cdots+ d_b( T_{j+p-1} T_{j+p}). \\ \end{split} \end{equation} | (3.19) |
By using (3.18) in (3.19) , we get
\begin{equation*} \begin{split} d_b( T_{j}, T_{j+p}) \leq& \; b\Gamma^{j} d_b( T_{0}, T_{1}) +b^{2}\Gamma^{j+1} d_b( T_{0}, T_{1})+b^{3}\Gamma^{j+2} \\ & d_b( T_{0}, T_{1})+\cdots+ b^{p} \Gamma^{j+p-1} d_b( T_{0}, T_{1}), \\ = & b\Gamma^{j} d_b( T_{0}, T_{1})(1+b \Gamma +b^{2}\Gamma^{2}+\cdots+b^{p-1}\Gamma^{p-1}), \\ \leq& b\Gamma^{j+1} d_b( T_{0}, T_{1}) \frac{1-(b\Gamma)^{p}}{1-\Gamma} , \\ < & b\Gamma^{j+1} d_b( T_{0}, T_{1}) \frac{1}{1-\Gamma}. \end{split} \end{equation*} |
Letting j\rightarrow \infty in above inequality, we have
\begin{equation*} \lim\limits_{j\rightarrow \infty} d_b( T(h_{j}, T_{j+1}), T(h_{j+p}, h_{j+p+1})\leq 0. \end{equation*} |
That is,
\begin{equation*} \lim\limits_{j\rightarrow \infty} d_b( T(h_{j}, h_{j+1}), T(h_{j+p}, h_{j+p+1}) = 0. \end{equation*} |
We get a Cauchy sequence T(h_{j-1}, h_{j}) in the closed subset K . Since X is complete, consider k_{\ast} \; \in \; K such that T(h_{j-1}, h_{j}) \rightarrow k_{\ast} . Moreover, Consider,
\begin{align*} d_b(h_{\ast}, K )&\leq d_b(h_{\ast}, T(h_{j-1}, h_{j}))\\ & = \lim\limits_{n\rightarrow \infty}d_b(h_{j+1}, T(h_{j-1}, h_{j}))\\ & = d_b(H, K)\\ &\leq d_b(h_{\ast}, K). \end{align*} |
Therefore, d_b(k_{\ast}, h_{j+1})\; \; \rightarrow \; \; d_b(k_{\ast}, H) as j \rightarrow \infty .
Condition (iv) implies, \lbrace h_{j} \rbrace has a subsequence \lbrace h_{j_{l}} \rbrace that converges to an element h_{\ast} \in H such that
\begin{equation*} d_b(h_{\ast}, T(h_{\ast}, h_{\ast})) = \lim\limits_{l \rightarrow \infty} d_b(h_{j_{l+1}}, T(h_{j_{l-1}}, h_{j_{l}})) = d_b(H, K ). \end{equation*} |
Hence,
\begin{equation*} d_b(h_{\ast}, T(h_{\ast}, h_{\ast})) = d_b(H, K ). \end{equation*} |
Theorem 3.5. Consider a complete b -metric space (X, d_b) with a coefficient b \geq 1 endowed with a binary relation \mathcal{R} , where b -metric is continuous. Suppose that H and K are non empty closed subsets of \ X . Consider a mapping T:H \times H \rightarrow K such that for each h_{1}, h_{2}, h_{3}, w_{1}, w_{2} \in H with h_{1}Ph_{3} such that h_{1}\mathcal{R}h_{2}, and h_{2}\mathcal{R}h_{3} , and d_b(w_{1}, T(h_{1}, h_{2})) = d_b(H, K) = d_b(w_{2}, T(h_{2}, h_{3})) , we have
\begin{equation*} d_b( T(h_{2}, h_{3}), T(w_{1}, w_{2}))\leq \Gamma \max \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3})), d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2})) \rbrace , \label{eq:21a} \end{equation*} |
where \Gamma \in[0, 1) with b \Gamma < 1 .
Furthermore, suppose that the subsequent conditions are true:
(1) T is path admissible;
(2) There exist h_{0}, h_{1}, h_{2} \in H which satisfy d_b(h_{2}, T(h_{0}, h_{1})) = d_b(H, K) and h_{0}Ph_{2} ;
(3) T(H \times H_{0})\subseteq K _{0} ;
(4) K is approximately compact with respect to H ;
(5) When \lbrace h_{j} \rbrace , \lbrace \overline{h_{j}} \rbrace \subseteq X such that h_{j} \rightarrow h and \overline{h_{j}} \rightarrow \overline{h} , then T(h_{j}, \overline{h_{j}}) \rightarrow T(h, \overline{h}) .
Then there exists a point h_{\ast} \in \; H which satisfies
\begin{equation*} d (h_{\ast}, T(h_{\ast}, h_{\ast})) = d (H, K ), \end{equation*} |
that is, T has a best proximity point.
Proof. This theorem can be proved by using similar argument as in Theorem 3.4.
In order to generalize the idea of partial ordering in metric spaces and partially ordered metric spaces, Jachymski [20] in 2008 has introduced the idea of a metric space endowed with a graph. This section is about a consequence of our results in the setting of metric spaces endowed with a graph.
Theorem 4.1. Let (X, d_b) be a complete b -metric space endowed with a graph G , where d_b is a continuous functional. Suppose that H and K are non empty closed subsets of X . Consider a mapping T: H \times H \rightarrow K such that for each h_{1}, h_{2}, h_{3}, w_{1}, w_{2} \in H with h_{1}Ph_{3} , that is, h_1 \mathcal{R}h_2, \; h_2\mathcal{R} h_3 and\\ d_b(w_{1}, T(h_{1}, h_{2})) = d_b(H, K) = d_b(w_{2}, T(h_{2}, h_{3})) , we have either
\begin{equation*} d_b(w_{1}, w_{2})\leq \Gamma \max \left\lbrace d_b(h_{1}, h_{2}), d_b(h_{2}, h_{3}) \right\rbrace , \end{equation*} |
or
\begin{equation*} d_b(h_{3}, w_{2})\leq \Gamma \max \left\lbrace d_b(h_{1}, h_{2}), d_b(h_{2}, w_{1}) \right\rbrace , \end{equation*} |
where \Gamma \in[0, 1) such that b \Gamma < 1 . Furthermore, assume that all the conditions of Theorem 3.1 are satisfied. Then T has a best proximity point.
Proof. It follows by using the same procedure as in Theorems 3.1 and 3.2.
Theorem 4.2. Let H and K be closed nonempty subsets of a complete b -metric space (X, d_b) endowed with a graph G = (V(G), E) where the b -metric is a continuous functional. Consider a mapping T:H \times H \rightarrow K such that for each h_{1}, h_{2}, h_{3}, w_{1}, w_{2} \in H with h_{1}Ph_{3} , that is, (h_{1}, h_{2})\in E and (h_{2}, h_{3}) \in E , and d_b(w_{1}, T(h_{1}, h_{2})) = d_b(H, K) = d_b(w_{2}, T(h_{2}, h_{3})) , we have either
\begin{equation*} d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2}))\leq \Gamma \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3}))\rbrace , \end{equation*} |
or
\begin{equation*} d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2}))\leq \Gamma \max \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3})), d_b( T(h_{2}, h_{3}), T(w_{1}, w_{2})) \rbrace , \label{eq:16a} \end{equation*} |
or
d_b( T(h_{2}, h_{3}), T(w_{1}, w_{2}))\leq \Gamma \max \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3})), d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2})) \rbrace , |
where \Gamma \in[0, 1) such that b \Gamma < 1 . Furthermore, assume that all the conditions of Theorem 3.3 are satisfied. Then T has a best proximity point.
Proof. It follows by using the same arguments given in Theorem 3.3, Theorem 3.4 and ref{thm5a}.
Taking A = B = X in Theorems 4.1 and 4.2, we obtain the following results, which guarantee the existence of a fixed point of the mapping T:X\times X\rightarrow X.
Theorem 5.1. Let (X, d_b) be a complete b -metric space endowed with a graph G , where d_b is a continuous functional. Let T: X \times X \rightarrow X be a mapping such that for each h_{1}, h_{2}, h_{3}, w_{1}, w_{2} \in X with h_{1}Ph_{3} that is (h_1, h_2), \; (h_2, h_3) \in E satisfies one of the following inequalities
\begin{equation*} d_b(w_{1}, w_{2})\leq \Gamma \max \left\lbrace d_b(h_{1}, h_{2}), d_b(h_{2}, h_{3}) \right\rbrace , \end{equation*} |
or
\begin{equation*} d_b(h_{3}, w_{2})\leq \Gamma \max \left\lbrace d_b(h_{1}, h_{2}), d_b(h_{2}, w_{1}) \right\rbrace , \end{equation*} |
where \Gamma \in[0, 1) such that b \Gamma < 1 . Furthermore, assume that the following conditions are satisfied:
(1) T is path admissible;
(2) There exist a_0, a_1, a_3 \in X with a_3 = T(a_0, a_1) and a_0Pa_3 ;
(3) If \lbrace h_{j}\rbrace\subseteq X such that h_{j}Ph_{j+2} for each j \in \mathbb{N} and h_{j} \rightarrow x_{\ast} as j \rightarrow \infty , then (h_{j}, x_{\ast})\in E for all j \in \mathbb{N} and (x_{\ast}, x_{\ast})\in E.
Then T has a fixed point in X.
Theorem 5.2. Let (X, d_b) be a complete b -metric space endowed with a graph G , where d_b is a continuous functional. Let T: X \times X \rightarrow X be a mapping such that for each h_{1}, h_{2}, h_{3}, w_{1}, w_{2} \in X with h_{1}Ph_{3} that is (h_1, h_2), \; (h_2, h_3) \in E satisfies one of the following inequalities:
\begin{equation*} d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2}))\leq \Gamma \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3}))\rbrace , \end{equation*} |
d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2}))\leq \Gamma \max \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3})), d_b( T(h_{2}, h_{3}), T(w_{1}, w_{2})) \rbrace , |
d_b( T(h_{2}, h_{3}), T(w_{1}, w_{2}))\leq \Gamma \max \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3})), d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2})) \rbrace , |
where \Gamma \in[0, 1) such that b \Gamma < 1 . Furthermore, assume that the following conditions are satisfied:
(1) T is path admissible;
(2) There exist a_0, a_1, a_3 \in X with a_3 = T(a_0, a_1) and a_0Pa_3 ;
(3) T is continuous with respect to each coordinate.
Then, T has a fixed point in X.
Suppose that G = (V, E) where V = X and E = X \times X , then Theorems 5.1 and 5.2 give rise to the following corollaries, respectively.
Corollary 5.1. Let (X, d_b) be a complete b -metric space, where d_b is a continuous functional and consider T:X \times X\rightarrow X a mapping such that for each h_{1}, h_{2}, h_{3}, w_{1}, w_{2} \in X , we have either
\begin{equation*} d_b(w_{1}, w_{2})\leq \Gamma \max \left\lbrace d_b(h_{1}, h_{2}), d_b(h_{2}, h_{3}) \right\rbrace , \end{equation*} |
or
\begin{equation*} d_b(h_{3}, w_{2})\leq \Gamma \max \left\lbrace d_b(h_{1}, h_{2}), d_b(h_{2}, w_{1}) \right\rbrace , \end{equation*} |
where \Gamma \in[0, 1) such that b \Gamma < 1 . Then T has a fixed point.
Corollary 5.2. Let (X, d_b) be a complete b -metric space, where d_b is a continuous functional and consider T:X \times X\rightarrow X as a mapping such that for each h_{1}, h_{2}, h_{3}, w_{1}, w_{2} \in X we have either
\begin{equation*} d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2}))\leq \Gamma \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3}))\rbrace , \end{equation*} |
or
d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2}))\leq \Gamma \max \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3})), d_b( T(h_{2}, h_{3}), T(w_{1}, w_{2})) \rbrace , |
or
d_b( T(h_{2}, h_{3}), T(w_{1}, w_{2}))\leq \Gamma \max \lbrace d_b( T(h_{1}, h_{2}), T(h_{2}, h_{3})), d_b( T(h_{2}, w_{1}), T(h_{3}, w_{2})) \rbrace , |
where \Gamma \in[0, 1) such that b \Gamma < 1 . Then T has a fixed point.
In this section Theorem 3.1 has been proved in the setting of double controlled metric spaces introduced by Abdeljawad et al.[1]. In [1] the notion of doubled controlled metric type space is given as follows.
Definition 6.1. Given non comparable functions \alpha, \mu : X\times X \rightarrow [1, \infty). If d_b:X \times X \rightarrow (0, \infty) satisfies
(1) d_b(w_1, w_2) = 0 \Longleftrightarrow w_1 = w_2
(2) d_b(w_1, w_2) = d_b(w_2, w_1)
(3) d_b(w_1, w_3)\leq \alpha(w_1, w_2)d_b(w_1, w_2)+\mu(w_2, w_3)d_b(w_2, w_3) for all w_1, w_2, w_3 \in X.
Then (X, d_b) is called a double controlled metric type by \alpha and \mu .
Remark 6.1. The class of double controlled metric is larger than b -metric. If \alpha(w) = \mu (w) = b \geq 1 for all w\in X then, double controlled metric type is a b -metric with coefficient b .
The notion of convergence, Cauchyness and completeness can be extended naturally in the setting of double controlled metric type space as in [1].
Theorem 6.1. Suppose that (X, d_b) be a complete double controlled metric type space by the functions \alpha, \mu : X\times X \rightarrow [1, \infty) such that
\begin{equation} \sup\limits_{m > 1} \lim\limits_{i\rightarrow \infty}\dfrac{\alpha(w_{i+1}, w_{i+2})}{\alpha({w_i}, {w_{i+1}})}\mu (w_i, w_m) < \dfrac{1}{\Gamma^{\frac{1}{2}}} \end{equation} | (6.1) |
and \lim\limits_{n\rightarrow \infty}\alpha(u, u_n) and \lim\limits_{n\rightarrow \infty}\mu(u, u_n) exist and are finite. Let \mathcal{R} be a binary relation on X , where d_ b is a continuous functional. Assume that H and K are nonempty closed subsets of X . Consider a mapping T: H \times H \rightarrow K such that for each h_{1}, h_{2}, h_{3}, w_{1}, w_{2} \in H with h_{1}Ph_{3} that is h_1 \mathcal{R}h_2, \; h_2\mathcal{R} h_3 and
d_b(w_{1}, T(h_{1}, h_{2})) = d_b(H, K) = d_b(w_{2}, T(h_{2}, h_{3})) , we have:
\begin{equation} d_b(w_{1}, w_{2})\leq \Gamma \max \left\lbrace d_b(h_{1}, h_{2}), d_b(h_{2}, h_{3}) \right\rbrace , \end{equation} | (6.2) |
where \Gamma \in[0, 1) such that b \Gamma < 1 . Furthermore, suppose that the subsequent conditions are true:
(1) T is path admissible;
(2) \exists \; h_{0}, h_{1}, h_{2} \in H which satisfy d_b(h_{2}, T(h_{0}, h_{1})) = d_b(H, K) and h_{0}Ph_{2} ;
(3) T(H \times H _{0})\subseteq K _{0} ;
(4) K is approximately compact with respect to H ;
(5) If \lbrace h_{j}\rbrace\subseteq X such that h_{j}Ph_{j+2} for each j \in \mathbb{N} and h_{j} \rightarrow x_{\ast} as j \rightarrow \infty , then h_{j}\mathcal{R} x_{\ast} for all j \in \mathbb{N} and x_{\ast}\mathcal{R} x_{\ast}.
Then T has a best proximity point.
Proof. Proceeding as in Theorem 3.1 till Eq (3.5) we obtain
d_b(h_{j-1}, h_{j})\; \leq \; Z \psi^{j} \; \; \; \forall \; \; \; j \in \mathbb{N}. Now for m > n
\begin{align*} d_b(h_n, h_m)&\leq \alpha(h_n, h_{n+1})d_b(h_n, h_{n+1})+\mu(h_{n+1}, h_m)d_b(h_{n+1}, h_m)\\ & = \alpha(h_n, h_{n+1})d_b(h_n, h_{n+1})+\mu(h_{n+1}, h_m)d_b(h_{n+1}, h_m)\\ &\leq\alpha(h_n, h_{n+1})d_b(h_n, h_{n+1})+\\ &\mu(h_{n+1}, h_m)\left[\alpha((h_{n+1}, (h_{n+2})d_b(h_{n+1}, h_{n+2})+\mu(h_{n+1}, h_m)d_b(h_{n+2}, h_m)\right]\\ & = \alpha(h_n, h_{n+1})d_b(h_n, h_{n+1})+ \mu(h_{n+1}, h_m)\alpha(h_{n+1}, h_{n+2})d_b(h_{n+1}, h_{n+2})\\ &+\mu(h_{n+1}, h_m)\mu(h_{n+2}, h_m)d_b(h_{n+2}, h_m)\\ &\leq \alpha(h_n, h_{n+1})d_b(h_n, h_{n+1})+ \mu(h_{n+1}, h_m)\alpha(h_{n+1}, h_{n+2})d_b(h_{n+1}, h_{n+2})\\ &+\cdots +\mu(h_{n+1}, h_m)\mu(h_{n+2}, h_m) \cdots \mu(h_{m-2}, h_{m-1})\alpha(h_{m-2}, h_{m-1})d_b(h_{m-2}, h_{m-1})\\ &+\mu(h_{n+1}, h_m)\mu(h_{n+2}, h_m) \cdots \mu(h_{m-2}, h_{m-1})\mu(h_{m-1}, h_m)d_b(h_{m-1}, h_{m})\\ &\leq \alpha(h_n, h_{n+1})d_b(h_n, h_{n+1})+ \mu(h_{n+1}, h_m)\alpha(h_{n+1}, h_{n+2})d_b(h_{n+1}, h_{n+2})\\ &+\cdots +\mu(h_{n+1}, h_m)\mu(h_{n+2}, h_m) \cdots \mu(h_{m-2}, h_{m-1})\alpha(h_{m-2}, h_{m-1})d_b(h_{m-2}, h_{m-1})\\ &+\mu(h_{n+1}, h_m)\mu(h_{n+2}, h_m) \cdots \mu(h_{m-2}, h_{m-1})\mu(h_{m-1}, h_m)\alpha(h_{m-1}, h_m)d_b(h_{m-1}, h_{m})\\ &\leq \alpha(h_n, h_{n+1})Z\psi^{n+1}+ \mu(h_{n+1}, h_m)\alpha(h_{n+1}, h_{n+2})Z\psi^{n+2}\\ &+\cdots +\mu(h_{n+1}, h_m)\mu(h_{n+2}, h_m) \cdots \mu(h_{m-2}, h_{m-1})\alpha(h_{m-2}, h_{m-1})Z\psi^{m-1}\\ &+\mu(h_{n+1}, h_m)\mu(h_{n+2}, h_m) \cdots \mu(h_{m-2}, h_{m-1})\mu(h_{m-1}, h_m)\alpha(h_{m-1}, h_m)Z\psi^{m}\\ & = Z\psi^{n+1}\left[\alpha(h_n, h_{n+1})+\sum\limits_{i = n+1}^{m-1}\left(\prod\limits_{j = n+1}^{i}\mu(h_j , h_m) \right)\alpha(h_i , h_{i+1})\psi^{i-n} \right] \end{align*} |
Denoting \mathbb{S}_q = \sum\limits_{i = 0}^{q}\left(\prod\limits_{j = 0}^{i}\mu(h_j, h_m) \right)\alpha(h_i, h_{i+1})\psi^{i}, we have
d_b(h_n, h_m)\leq Z \psi^{n+1}\left[\alpha(h_n, h_{n+1})+(\mathbb{S}_{m-1}-\mathbb{S}_{n}) \right] |
The ratio test combined with (6.1) imply that the limit of the sequence \{\mathbb{S}_n\} exists. Hence
\begin{equation} \lim\limits_{n, m\rightarrow \infty}d_b(h_n, h_m) = 0, \end{equation} | (6.3) |
implies that \{h_n\} is a Cauchy sequence in H. Since H is complete, there exists some h_{\ast}\in H such that h_n\rightarrow h_{\ast}. Hence by (vi) h_n\mathcal{R}h_{\ast}\; \forall n\in \mathbb{N}. Furthermore, we have to prove that d_b(h_{\ast}, T(h_{n-1}, h_{n}))\rightarrow d_b(h_{\ast}, K)\; \; \; as\; \; \; j\rightarrow \infty . Consider,
\begin{align*} d_b(h_{\ast}, K )&\leq d_b(h_{\ast}, T(h_{n-1}, h_{n}))\\ & = \lim\limits_{n\rightarrow \infty}d_b(h_{n+1}, T(h_{n-1}, h_{n}))\\ & = d_b(H, K)\\ &\leq d_b(h_{\ast}, K) \end{align*} |
Therefore d_b(h_{\ast}, T(h_{n-1}, h_{n}))\rightarrow d_b(h_{\ast}, K) as n\rightarrow \infty. The rest of the proof can be carried out in the same way as in Theorem 3.1 after Equation (3.6).
Remark 6.2. Note that Theorem 3.1 becomes a special case of Theorem 6.1 by taking \alpha(w) = \mu (w) = b \geq 1 for all w\in X .
The author Aiman Mukheimer and Suhad Subhi Aiadi would like to thank Prince Sultan University for paying APC and for the support through TAS research LAB.
The authors declare that they have no competing interests.
[1] |
T. Abdeljawad, N. Mlaiki, H. Aydi, N. Souayah, Double controlled metric type spaces and some fixed point results, Mathematics, 6 (2018), 320. https://doi.org/10.3390/math6120320 doi: 10.3390/math6120320
![]() |
[2] |
M. U. Ali, H. Aydi, M. Alansari, New generalizations of set valued interpolative Hardy-Rogers type contractions in b-metric spaces, J. Funct. Space., 2021 (2021), 6641342. https://doi.org/10.1155/2021/6641342 doi: 10.1155/2021/6641342
![]() |
[3] |
M. U. Ali, M. Farheen, T. Kamran, G. Maniu, Pre\hat{\text{s}}i\acute{c} type nonself operators and related best proximity results, Mathematics, 7 (2019), 394. https://doi.org/10.3390/math7050394 doi: 10.3390/math7050394
![]() |
[4] |
M. U. Ali, T. Kamran, M. Postolache, Solution of Volterra integral inclusion in b-metric spaces via new fixed point theorem, Nonlinear Anal. Model., 22 (2017), 17–30. http://dx.doi.org/10.15388/NA.2017.1.2 doi: 10.15388/NA.2017.1.2
![]() |
[5] | M. U. Ali, T. Kamran, Multivalued F-contraction and related fixed point theorems with application, Filomat, 30 (2016), 3779–3793. |
[6] |
H. H. Al-Sulami, N. Hussain, J. Ahmad, Best proximity results with applications to nonlinear dynamical systems, Mathematics, 7 (2019), 900. https://doi.org/10.3390/math7100900 doi: 10.3390/math7100900
![]() |
[7] |
I. Altun, M. Aslantas, H. Sahin, KW-type nonlinear contractions and their best proximity points, Numer. Func. Anal. Opt., 42 (2021), 935–954. https://doi.org/10.1080/01630563.2021.1933526 doi: 10.1080/01630563.2021.1933526
![]() |
[8] |
I. Altun, H. Sahin, M. Aslantas, A new approach to fractals via best proximity point, Chaos Soliton. Fract., 146 (2021), 110850. https://doi.org/10.1016/j.chaos.2021.110850 doi: 10.1016/j.chaos.2021.110850
![]() |
[9] |
I. Altun, H. Sahin, D. Turkoglu, Caristi-type fixed point theorems and some generalizations on M-metric space, Bull. Malays. Math. Sci. Soc., 43 (2020), 2647–2657. https://doi.org/10.1007/s40840-019-00823-8 doi: 10.1007/s40840-019-00823-8
![]() |
[10] | H. Aydi, M. F. Bota, E. Karapinar, S. Moradi, A common fixed point for weak phi-contractions on b-metric spaces, Fixed Point Theor., 13 (2012), 337–346. |
[11] |
H. Aydi, M. F. Bota, E. Karapinar, S. Mitrović, A fixed point theorem for set-valued quasi-contractions in b-metric spaces, Fixed Point Theory Appl., 2012 (2012), 88. https://doi.org/10.1186/1687-1812-2012-88 doi: 10.1186/1687-1812-2012-88
![]() |
[12] | S. Banach, Sur les opr\acute{\text{e}}ations dans les ensembles abstraits et leurs applications aux \acute{\text{e}}quations int\acute{\text{e}}gals, Fund. Math., 3 (1922), 133–181. |
[13] |
V. Bernide, M. P\hat{\text{a}}curar, Stability of k-step Fixed Point Iterative Sequence for some Pre\hat{\text{s}}i\acute{c} type Contractive Mappings, J. Inequal. Appl., 2014 (2014), 149. https://doi.org/10.1186/1029-242X-2014-149 doi: 10.1186/1029-242X-2014-149
![]() |
[14] | L. B. \acute{\text{C}}iri\acute{\text{c}}, S. B. Pre\hat{\text{s}}i\acute{c}, On Pre\hat{\text{s}}i\acute{c} type generalization of the Banach contraction mapping principle, Acta. Math. Univ. Comenianae, LXXVI (2007), 143–147. |
[15] | S. Czerwik, Contraction mappings in b-metric spaces, Acta Mathematica et Informatica Universitatis Ostraviensis, 1 (1993), 1–5. |
[16] |
M. Edelstein, An extension of Banach's contraction principle, P. Am. Math. Soc., 12 (1961), 7–10. https://doi.org/10.2307/2034113 doi: 10.2307/2034113
![]() |
[17] |
M. Edelstein, On fixed and periodic points under contractive mappings, J. Lond. Math. Soc., s1-37 (1962), 74–79. https://doi.org/10.1112/jlms/s1-37.1.74 doi: 10.1112/jlms/s1-37.1.74
![]() |
[18] |
K. Fan, Extension of two fixed point theorems of F. E. Browder, Math. Z., 112 (1969), 234–240. https://doi.org/10.1007/BF01110225 doi: 10.1007/BF01110225
![]() |
[19] |
L. G. Huang, X. Zhang, Cone metric spaces and fixed point theorems of contractive mappings, J. Math. Anal. Appl., 332 (2007), 1468–1476. https://doi.org/10.1016/j.jmaa.2005.03.087 doi: 10.1016/j.jmaa.2005.03.087
![]() |
[20] | J. Jachymski, The contraction principle for mappings on a metric space with a graph, P. Am. Math. Soc., 136 (2008), 1359–1373. |
[21] |
T. Kamran, F. Uddin, M. U. Ali, Common fixed point theorems for a family of multivalued F-contractions with an application to solve a system of integral equations, Glas. Mat., 52 (2017), 163–177. https://doi.org/10.3336/gm.52.1.12 doi: 10.3336/gm.52.1.12
![]() |
[22] | T. Kamran, C. Vetro, M. U. Ali, M. Waheed, A fixed point theorem for G-monotone multivalued mapping with application to nonlinear integral equations, Filomat, 31 (2017), 2045–2052. |
[23] | E. Karapınar, R. Ali, T. Kamran, M. U. Ali, Generalized multivalued rational type contractions, J. Adv. Math. Stud., 9 (2016), 26–36. |
[24] |
E. Karapinar, R. Agarwal, H. Aydi, Interpolative Reich–Rus–Ciric type contractions on partial metric spaces, Mathematics, 6 (2018), 256. https://doi.org/10.3390/math6110256 doi: 10.3390/math6110256
![]() |
[25] |
E. Karapinar, A. Fulga, A. Petrusel, On Istratescu type contractions in b-metric spaces, Mathematics, 8 (2020), 388. https://doi.org/10.3390/math8030388 doi: 10.3390/math8030388
![]() |
[26] |
T. Kamran, M. Postolache, A. Ghiura, S. Batul, R. Ali, The Banach contraction principle in C^{*} -algebra-valued b-metric spaces with application, Fixed Point Theory Appl., 2016 (2016), 10. https://doi.org/10.1186/s13663-015-0486-z doi: 10.1186/s13663-015-0486-z
![]() |
[27] | M. P\hat{\text{a}}curar, Common fixed points for almost Pre\hat{\text{s}}i\acute{c} type operators, Carpathian J. Math., 28 (2012), 117–126. |
[28] | S. B. Pre\hat{\text{s}}i\acute{c}, Sur la convergence des suites, Comptes Rendus de l'Acad. des Sci. de Paris, 260 (1965), 3828–3830. |
[29] |
S. Sadiq Basha, N. Shahzad, Best proximity point theorems for generalized proximal contractions, Fixed Point Theory Appl., 2012 (2012), 42. https://doi.org/10.1186/1687-1812-2012-42 doi: 10.1186/1687-1812-2012-42
![]() |
1. | Mădălina Păcurar, Asymptotic stability of equilibria for difference equations via fixed points of enriched Prešić operators, 2023, 56, 2391-4661, 10.1515/dema-2022-0185 |