The aim of this paper is to introduce the notion of Suzuki type multivalued contraction with simulation functions and then to set up some new fixed point and data dependence results for these type of contraction mappings. We produce an example to support our results. Moreover, we present an application to functional equation arising in dynamical system.
Citation: Azhar Hussain, Saman Yaqoob, Thabet Abdeljawad, Habib Ur Rehman. Multivalued weakly Picard operators via simulation functions with application to functional equations[J]. AIMS Mathematics, 2021, 6(3): 2078-2093. doi: 10.3934/math.2021127
[1] | Monairah Alansari, Mohammed Shehu Shagari, Akbar Azam, Nawab Hussain . Admissible multivalued hybrid Z-contractions with applications. AIMS Mathematics, 2021, 6(1): 420-441. doi: 10.3934/math.2021026 |
[2] | A.G. Ibrahim, A.A. Elmandouh . Existence and stability of solutions of ψ-Hilfer fractional functional differential inclusions with non-instantaneous impulses. AIMS Mathematics, 2021, 6(10): 10802-10832. doi: 10.3934/math.2021628 |
[3] | Mustafa Mudhesh, Hasanen A. Hammad, Eskandar Ameer, Muhammad Arshad, Fahd Jarad . Novel results on fixed-point methodologies for hybrid contraction mappings in Mb-metric spaces with an application. AIMS Mathematics, 2023, 8(1): 1530-1549. doi: 10.3934/math.2023077 |
[4] | Tahair Rasham, Najma Noor, Muhammad Safeer, Ravi Prakash Agarwal, Hassen Aydi, Manuel De La Sen . On dominated multivalued operators involving nonlinear contractions and applications. AIMS Mathematics, 2024, 9(1): 1-21. doi: 10.3934/math.2024001 |
[5] | Amjad Ali, Muhammad Arshad, Awais Asif, Ekrem Savas, Choonkil Park, Dong Yun Shin . On multivalued maps for φ-contractions involving orbits with application. AIMS Mathematics, 2021, 6(7): 7532-7554. doi: 10.3934/math.2021440 |
[6] | Nehad Abduallah Alhajaji, Afrah Ahmad Noman Abdou, Jamshaid Ahmad . Application of fixed point theory to synaptic delay differential equations in neural networks. AIMS Mathematics, 2024, 9(11): 30989-31009. doi: 10.3934/math.20241495 |
[7] | Ricardo Almeida, Natália Martins, J. Vanterler da C. Sousa . Fractional tempered differential equations depending on arbitrary kernels. AIMS Mathematics, 2024, 9(4): 9107-9127. doi: 10.3934/math.2024443 |
[8] | Dong Ji, Yao Yu, Chaobo Li . Fixed point and endpoint theorems of multivalued mappings in convex b-metric spaces with an application. AIMS Mathematics, 2024, 9(3): 7589-7609. doi: 10.3934/math.2024368 |
[9] | Godwin Amechi Okeke, Akanimo Victor Udo, Rubayyi T. Alqahtani, Nadiyah Hussain Alharthi . A faster iterative scheme for solving nonlinear fractional differential equations of the Caputo type. AIMS Mathematics, 2023, 8(12): 28488-28516. doi: 10.3934/math.20231458 |
[10] | Gulalai, Shabir Ahmad, Fathalla Ali Rihan, Aman Ullah, Qasem M. Al-Mdallal, Ali Akgül . Nonlinear analysis of a nonlinear modified KdV equation under Atangana Baleanu Caputo derivative. AIMS Mathematics, 2022, 7(5): 7847-7865. doi: 10.3934/math.2022439 |
The aim of this paper is to introduce the notion of Suzuki type multivalued contraction with simulation functions and then to set up some new fixed point and data dependence results for these type of contraction mappings. We produce an example to support our results. Moreover, we present an application to functional equation arising in dynamical system.
We pronounce (N,℘) a metric space, the classes P(N) and CB(N) of all nonempty, closed and bounded subsets of N.
If U,V∈CB(N), review the functionals:
D:P(N)×P(N)→R+, D(W,V)=inf{℘(ω,λ):ω∈W,λ∈V}, W⊂N;
H:P(N)×P(N)→R+,H(U,V):=max{supu∈Uinfv∈V℘(u,v),supv∈Vinfu∈U℘(u,v)} - the Pompeiu- Hausdorff functional.
Lemma 1.1. [12] Suppose U⊆N and ξ>1. Then for μ∈N there is v∈V such that ℘(μ,v)≤ξD(μ,V).
Definition 1.2. [16] A function Q:N→CB(N) is called multi-valued weakly Picard operator (MWP operator) if for all μ∈N and μ∈Qμ, there is a sequence (μm) in N such that (i) μ0=μ,μ1=λ, (ii) μn+1∈Qμm ∀m≥0, and (iii) (μm) tends to the fixed point of Q.
Utilizing the approach of Rus [16], Popescu [13] defined a multivalued operator and named it as (s,r)-contractive multivalued and siglevalued operators and showed that the underlying mapping is an MWP operator. Following Popescu [13], Kamran [9] extended (s,r)-contractive operators to weakly (s,r)-contractive operators and proved that the underlying mappings are still MWP operators.
A new generalization of Banach's theorem was given by Khojasteh et al. [11] by considering a function ζ:[0,∞)×[0,∞)→R, which follows the assertions given below:
(ζa) ζ(0,0)=0,
(ζb) ζ(α,β)<β−α for all α,β>0,
(ζc) The sequences {αn},{βn} in (0,∞) with limn→∞αn=limn→∞βn>0, then
limn→∞supζ(αn,βn)<0. |
They named such functions as simulation function and the contractive map is named as Z-contraction. Later on, Roldˊan-Lˊopez-de-Hierro et al. [15] replaced (ζc) with (ζ′c),
(ζ′c): the sequences (αn),(βn) in (0,∞) with limn→∞αn=limn→∞βn>0 and αn<βn, then
limn→∞supζ(αn,βn)<0. |
The lemma stated below is essential:
Lemma 1.3. [14] Suppose a sequence (μn) in N with
limn→∞℘(μn,μn+1)=0. | (1.1) |
If the sequence is not a Cauchy, then ∃ ε>0 and sequences μm(j) and μn(j) such that n(j)>m(j)>j and the sequences
℘(μm(j),μn(j)),℘(μm(j),μn(j)+1),℘(μm(j)−1,μn(j)), |
℘(μm(j)−1,μn(j)+1),℘(μm(j)+1,μn(j)+1) |
converges to ε+ when k→∞
The purpose of this paper is to introduce the notion of Suzuki type multivalued Z(ϕ,λ)-contraction and to prove the fixed point and data dependence results for such contraction mapping. We give an example to show the validity of our results. Moreover, we present an application to functional equation arising in dynamical system to show the usability of our results.
We now give the definition of Suzuki type multivalued Z(ϕ,λ)-contraction and obtain some fixed point results.
We start this section with the following notion:
Denote Φ the set consisting of all strictly increasing functions φ:[0,1)→(0,1]. We define
φ(s)={1s+1if0≤s<121−sif12≤s<1. |
Definition 2.1. A mapping Q:N→CB(N) is called Suzuki type multivalued Z(φ,λ)-contraction with respect to ζ, if there is a function φ∈Φ such that
φ(s)max{D(μ,Qμ),12D(λ,Qμ)}≤℘(μ,λ) | (2.1) |
implies
ζ(H(Qμ,Qλ),ρR(μ,λ))≥0 | (2.2) |
where ρ∈[0,1), and
R(μ,λ)=max{℘(μ,λ),D(μ,Qμ),D(λ,Qλ),12(D(μ,Qλ)+D(λ,Qμ))}. |
Theorem 2.2. Suppose that Q:N→CB(N) is Suzuki type multivalued Z(φ,ρ)-contraction, where (N,℘) is a complete metric space. Then Q is an MWP operator.
Proof. Choose μ0∈X, μ1∈Qμ0 and ρ1 be real number such that 0≤ρ≤s≤ρ1<1. One can choose μ2∈Qμ1 such that
℘(μ1,μ2)≤1√ρ1H(Qμ0,Qμ1). |
Since
φ(s)max{D(μ0,Qμ0),12D(μ1,Qμ0)}≤φ(s)D(μ0,Qμ0)≤℘(μ0,μ1), | (2.3) |
from (2.1) and (2.2), we have
0≤ζ(H(Qμ0,Qμ1),ρR(μ0,μ1))≤ρR(μ0,μ1)−H(Qμ0,Qμ1) |
this implies
H(Qμ0,Qμ1)≤ρR(μ0,μ1), |
where
R(μ0,μ1)=max{℘(μ0,μ1),D(μ0,Qμ0),D(μ1,Qμ1),D(μ0,Qμ1)+D(μ1,Qμ0)2}=max{℘(μ0,μ1),D(μ1,Qμ1),D(μ0,Qμ1)2}≤max{℘(μ0,μ1),℘(μ1,μ2),℘(μ0,μ1)+℘(μ1,μ2)2}=max{℘(μ0,μ1),℘(μ1,μ2)}. |
If R(μ0,μ1)≤℘(μ1,μ2), then
℘(μ1,μ2)≤1√ρ1H(Qμ0,Qμ1)≤ρρ1R(μ0,μ1)≤√ρ1℘(μ1,μ2). |
This implies ℘(μ1,μ2)=0 or μ1=μ2∈Qμ1, i.e. μ1 is a fixed point. So, we consider the case when R(μ0,μ1)≤℘(μ0,μ1), this gives
℘(μ1,μ2)≤√ρ1℘(μ0,μ1)<℘(μ0,μ1). |
Hence we can recursively define a sequence (μn) such that μn+1∈Qμn and
℘(μn+1,μn+2)<℘(μn,μn+1) |
which implies
limn→∞℘(μn,μn+1)=0. | (2.4) |
To prove that (μn) is Cauchy, contrary suppose that it is not, then by Lemma 1.3, there is ϵ>0 and two subsequences m(j), n(j) with n(j)>m(j)>j such that ℘(μm(j),μn(j))→ϵ+, ℘(μm(j)+1,μn(j)+1)→ϵ+ as j→∞ also
limj→∞ρR(μm(j),μn(j))=ϵ+. |
By (2.4), we have
φ(s)max{D(μn(j),Qμn(j)),12D(μm(j),Qμn(j))} | (2.5) |
≤max{℘(μn(j),μn(j)+1),12℘(μm(j),μn(j)+1)}≤max{℘(μm(j),μn(j))+℘(μn(j),μn(j)+1),℘(μm(j),μn(j)+1)}=℘(μm(j),μn(j))+℘(μn(j),μn(j)+1)<℘(μm(j),μn(j)), | (2.6) |
thus by (2.1), (2.2) and (ζb), we have
0≤ζ(H(Qμm(j),Qμn(j)),ρR(μm(j),μn(j))) | (2.7) |
≤ρR(μm(j),μn(j))−H(Qμm(j),Qμn(j)), | (2.8) |
therefore, we have
℘(μm(j)+1,μn(j)+1)≤H(Qμm(j),Qμn(j))≤ρR(μm(j),μn(j))<℘(μm(j),μn(j)), |
hence
limj→∞H(Qμm(j),Qμn(j))=ϵ+, |
so by (ζc) and (2.2), we have
0≤limj→∞supζ(H(Qμm(j),Qμn(j)),ρR(μm(j),μn(j)))<0, |
this leads to a contradiction. Hence (μn) is a Cauchy sequence. Completeness of N implies that there is u∈N such that limn→∞μn=u. Moreover
D(μn+1,Qu)≤H(Qμn,Qu). |
Our claim is
D(u,Qμ)≤ρmax{℘(u,μ),D(μ,Qμ)} | (2.9) |
for all μ≠u. Since limn→∞℘(un,u)=0, there exists an n0∈N such that ℘(u,un)<13℘(u,μ) for all n≥n0. Note that φ(s)≤1 and
φ(s)max{D(un,Qun),12D(μ,Qun)}≤max{D(un,Qun),12D(μ,Qun)}≤max{℘(un,un+1),12℘(μ,un+1)}≤max{℘(un,u)+℘(u,un+1),12(℘(μ,u)+℘(u,un+1))}≤max{13℘(u,μ)+13℘(u,μ),12[℘(μ,u)+13℘(u,μ)]}≤max{23℘(u,μ),23℘(μ,u)}=23℘(μ,u)=(℘(u,μ)−13℘(u,μ))≤(℘(u,μ)−℘(un,u))≤℘(un,μ). |
Thus φ(s)max{D(un,Qun),12D(μ,Qun)}≤℘(un,μ) holds for all n≥n0. From (2.1) and (2.2), we have
ζ(H(Qun,Qμ),ρR(un+1,μ))≥0, |
this gives
D(un+1,Qμ)≤H(Qun,Qμ)≤ρmax{℘(un,μ),D(un,Qun),D(μ,Qμ),D(un,Qμ)+D(μ,Qun)2}≤ρmax{℘(un,μ),℘(un,un+1),D(μ,Qμ),D(un,Qμ)+℘(μ,un+1)2}. | (2.10) |
Tending n→∞ to inequality (2.10), we arrived at
D(u,Qμ)≤ρmax{℘(u,μ),D(μ,Qμ),D(u,Qμ)+℘(μ,u)2}. |
If
max{℘(u,μ),D(μ,Qμ),D(u,Qμ)+℘(μ,u)2}=D(u,Qμ)+℘(μ,u)2, |
then
D(u,Qμ)≤ρ(D(u,Qμ)+℘(μ,u)2) |
and this implies that
D(u,Qμ)≤(ρ2−ρ)℘(μ,u)<ρ℘(μ,u)≤ρmax{℘(u,μ),D(μ,Qμ)}. |
Hence
D(u,Qμ)≤ρmax{℘(u,μ),D(μ,Qμ)} |
valid ∀ μ≠u. To show that u∈Qu, we discus two cases:
(i) If 0≤r≤s<12. Contrary suppose that u∉Qu. Since Qu≠ϕ, there exists c∈Qu such that c≠u and
℘(c,u)<D(u,Qu)+(1ρ+1−1)D(u,Qu). |
Thus
(ρ+1)℘(c,u)<D(u,Qu). | (2.11) |
As φ(s)max{D(u,Qu),12D(c,Qu)}≤max{℘(u,c),℘(c,c)}≤℘(u,c), so by (2.1) and (2.2), we have
ζ(H(Qu,Qc),ρR(u,c))≥0, |
this gives
H(Qu,Qc)≤ρmax{℘(u,c),D(u,Qu),D(c,Qc),D(u,Qc)+D(c,Qu)2}≤ρmax{℘(u,c),D(c,Qc),[℘(u,c)+D(c,Qc)]2}=ρmax{℘(u,c),D(c,Qc)}. |
Thus
H(Qu,Qc)≤ρmax{℘(u,c),D(c,Qc)}. | (2.12) |
Since D(c,Qc)≤H(Qu,Qc), therefore, we have
H(Qu,Qc)≤ρmax{℘(u,c),H(Qu,Qc)}, |
hence
H(Qu,Qc)≤ρ℘(u,c). | (2.13) |
Since ρ<1, therefore
D(c,Qc)≤℘(u,c). | (2.14) |
By (2.9) and (2.14), we have
D(u,Qc)≤ρmax{℘(u,c),D(c,Qc)}<ρmax{℘(u,c),℘(u,c)}<ρ℘(u,c)<℘(u,c). | (2.15) |
Using (2.9), (2.12), (2.13) and (2.15), we have
D(u,Qu)≤D(u,Qc)+H(Qu,Qc)<℘(u,c)+ρ℘(u,c)=(1+ρ)℘(u,c)<D(u,Qu), |
a contradiction. Hence u∈Qu.
(ii) If 12≤ρ≤s<1, then we first show that
H(Qμ,Qu)≤ρR(μ,u) | (2.16) |
for all μ∈N with μ≠u.For each n∈N, there exists λn∈Qμ such that ℘(u,λn)<D(u,Qμ)+1n℘(μ,u). So we have
max{D(μ,Qμ),12D(u,Qμ)}≤max{℘(μ,λn),12℘(u,λn)}≤max{(℘(μ,u)+℘(u,λn)),℘(u,λn)}=(℘(μ,u)+℘(u,λn))<(℘(μ,u)+D(u,Qμ)+1n℘(μ,u)). |
Now by (2.9), we have
max{D(μ,Qμ),12D(u,Qμ)}<(℘(μ,u)+ρmax{℘(u,μ),D(μ,Qμ)}+1n℘(μ,u)). | (2.17) |
If max{℘(u,μ),D(μ,Qμ)}=℘(μ,u), then from (2.17), we obtain that
max{D(μ,Qμ),12D(u,Qμ)}<(℘(μ,u)+ρ℘(u,μ)+1n℘(μ,u))=(1+ρ+1n)℘(μ,u)<(1+s+1n)℘(μ,u), |
which gives
φ(s)max{D(μ,Qμ),12D(u,Qμ)}=(1−s)max{D(μ,Qμ),12D(u,Qμ)}≤(12)max{D(μ,Qμ),12D(u,Qμ)}≤(11+s)max{D(μ,Qμ),12D(u,Qμ)}<(11+s)(1+s+1n)℘(μ,u),≤(1+s1+s+1(1+s)n)℘(μ,u)≤(1+1(1+s)n)℘(μ,u). |
Tending n→∞, we have
φ(s)max{D(μ,Qμ),12D(u,Qμ)}≤℘(μ,u). |
By (2.1) with λ=u, (2.16) satisfied. If max{℘(u,μ),D(μ,Qμ)}=D(μ,Qμ), then (2.9) gives
D(μ,Qμ)≤℘(μ,u)+D(u,Qμ)≤℘(μ,u)+ρD(μ,Qμ) |
and hence D(μ,Qμ)≤(11−ρ)℘(μ,u). By (2.17), we obtain that
max{D(μ,Qμ),12D(u,Qμ)}≤(℘(μ,u)+ρD(μ,Qμ)+1n℘(μ,u))≤(℘(μ,u)+(ρ1−ρ)℘(μ,u)+1n℘(μ,u))=(11−ρ)℘(μ,u)+1n℘(μ,u). |
As 12≤ρ<1 and ρ≤s, so
φ(s)max{D(μ,Qμ),12D(u,Qμ)}=(1−s)max{D(μ,Qμ),12D(u,Qμ)}≤(1−ρ)max{D(μ,Qμ),12D(u,Qμ)}≤(1−ρ)((11−ρ)℘(μ,u)+1n℘(μ,u))≤℘(u,μ)+(1−ρn)℘(u,μ)≤℘(u,μ)+(1−ρn)℘(u,μ), |
tending n→∞ gives
φ(s)max{D(μ,Qμ),12D(u,Qμ)}≤℘(μ,u) |
and thus (2.16) satisfied. Using μ=un in (2.16), we have
D(un+1,Qu)≤H(Qun,Qu)≤ρmax{℘(un,u),D(un,Qun),D(u,Qu),D(un,Qu)+D(u,Qun)2}≤ρmax{℘(un,u), ℘(un,un+1),D(u,Qu),D(un,Qu)+℘(u,un+1)2}. | (2.18) |
Tending n→∞ to inequality (2.18), we arrived at
D(u,Qu)≤ρD(u,Qu) |
which further gives that D(u,Qu)=0, that is, u∈Qu.
For the case of single-valued Suzuki type Z(φ,ρ)-contraction, we have the following result:
Theorem 2.3. Assume that Q:N→N be Suzuki type Z(φ,ρ)-contraction on a complete metric space (N,℘), then N is an MWP operator and has a unique fixed point.
Proof. By Theorem 2.2, Q is an MWP operator. Let μ≠λ be two fixed points of Q, then
φ(s)max{℘(μ,Qμ),12℘(λ,Qμ)}=φ(s)12℘(μ,λ)<℘(μ,λ) |
this implies
0≤ζ(℘(Qμ,Qλ),ρR(μ,λ))=ζ(℘(μ,λ),ρ℘(μ,λ))≤ρ℘(μ,λ)−℘(μ,λ) |
this gives
℘(μ,λ)≤ρ℘(μ,λ), |
which is not possible, hence ℘(μ,λ)=0. Thus Q possesses a unique fixed point.
Example 2.1. Let N={0,1,2,3} and ℘(μ,λ)=|μ−λ|. Define Q:N→CB(N) by
Qμ={{0,2}if μ≠3,{1}if μ=3. |
Now for μ,λ∈N with μ≠λ, we have
(μ,λ) | ℘(μ,λ) | D(μ,Qμ) | D(λ,Qμ) | H(Qμ,Qλ) | R(μ,λ) |
(0, 1) | 1 | 0 | 1 | 0 | 1 |
(0, 2) | 2 | 0 | 0 | 0 | 2 |
(0, 3) | 3 | 0 | 1 | 1 | 3 |
(1, 0) | 1 | 1 | 0 | 0 | 1 |
(1, 2) | 1 | 1 | 0 | 0 | 1 |
(1, 3) | 2 | 1 | 1 | 1 | 2 |
(2, 0) | 2 | 0 | 0 | 0 | 2 |
(2, 1) | 1 | 0 | 1 | 0 | 1 |
(2, 3) | 1 | 0 | 1 | 1 | 2 |
(3, 0) | 3 | 2 | 1 | 1 | 3 |
(3, 1) | 2 | 2 | 0 | 1 | 2 |
(3, 2) | 1 | 2 | 1 | 1 | 2 |
So max{D(μ,Qμ):μ∈N}=2, max{D(λ,Qμ):μ,λ∈N}=1 and min{℘(μ,λ):μ,λ∈N}=1. Now for s=0.6,φ(s)=0.4 and
φ(s)max{D(μ,Qμ),12D(λ,Qμ)}=(0.4)(2)<1=min{℘(μ,λ);μ,λ∈N} |
holds true for all μ,λ∈N. Let ζ(α,β)=45β−α, then for ρ=0.7, we have
ζ(H(Q0,Q1),ρR(0,1))=ζ(0,0.7)=(0.8)(0.7)−0>0ζ(H(Q0,Q2),ρR(0,2))=ζ(0,1.4)=(0.8)(1.4)−0>0ζ(H(Q0,Q3),ρR(0,3))=ζ(1,2.1)=(0.8)(2.1)−1>0ζ(H(Q1,Q2),ρR(1,2))=ζ(0,0.7)=(0.8)(0.7)−0>0ζ(H(Q1,Q3),ρR(1,3))=ζ(1,1.4)=(0.8)(1.4)−1>0ζ(H(Q2,Q3),ρR(2,3))=ζ(1,1.4)=(0.8)(1.4)−1>0. |
Hence Q is Z(φ,0.7)-contraction. Thus by Theorem 2.2, Q is an MWP operator and Fix(Q)={0,2}.
We now prove a data dependence result.
Theorem 3.1. Ssppose that Q1, Q2 are two multivalued operators on a metric space (N,℘), if
1. Qi is Z(φ,ρi)-contraction for i=1,2,
2. there exist a real number ρ>0 such that H(Q1μ,Q2μ)≤ρ, ∀ μ∈N.
Then
1. Fix(Qi)∈CL(N), for i=1,2,
2. Q1 and Q2 are multivalued operators and
H(Fix(Q1),Fix(Q2))≤ρ1−max{ρ1,ρ2}. |
Proof. Theorem 2.2 gives that Fix(Qi)≠ϕ for i=1,2. To prove that Fix(Q) is closed, assume that (μn)⊂Fix(Q) such that μn→t as n→∞, we have
φ(s)max{D(μn,Qμn),12D(ω,Qμn)}=φ(s)12D(ω,Qμn)≤12D(ω,Qμn)<℘(ω,μn) |
this implies
0≤ζ(H(Qμn,Qω),ρR(μn,ω))≤ρR(μn,ω)−H(Qμn,Qω) |
this gives
H(Qμn,Qω)≤ρR(μn,ω). |
Now
D(ω,Qω)≤℘(ω,μn)+D(μn,Qω)≤℘(ω,μn)+H(Qμn,Qω)≤℘(ω,μn)+ρR(μn,ω)≤℘(ω,μn)+ρmax{℘(ω,μn),D(μn,Qμn),D(ω,Qω),D(ω,Qμn)+D(μn,Qω)2}, |
letting n→∞, we have D(ω,Qω)=0, so ω∈Fix(Q) and hence Fix(Q) is closed.
(b) From Theorem 2.2, it follows from (i) that Q1 and Q2 are MWP operators. Further, let μ0 be fixed point of Q1, then there exist μ1∈Q2μ0 and a real number q>1 such that
℘(μ0,μ1)≤qH(Q1μ0,Q2μ0), |
since μ1∈Q2μ0, we can choose μ2∈Q2μ1 such that
℘(μ1,μ2)≤qH(Q2μ0,Q2μ1) |
also
φ(s)max{D(μ0,Q2μ0),12D(μ1,Q2μ0)}≤φ(s)℘(μ0,μ1)≤℘(μ0,μ1) |
this implies
0≤ζ(H(Q2μ0,Q2μ1),ρ2R(μ0,μ1))≤ρ2R(μ0,μ1)−H(Q2μ0,Q2μ1). |
Hence
H(Q2μ0,Q2μ1)≤ρ2R(μ0,μ1)≤ρ2℘(μ0,μ1). |
(Choosing R(μ0,μ1)≤℘(μ1,μ2) gives fixed point). Hence
℘(μ1,μ2)≤qH(Q2μ0,Q2μ1)≤qρ2℘(μ0,μ1). |
Therefore, for operator Q2, we get an iterative sequence of successive approximations satisfying μn+1∈Q2μn such that
℘(μn,μn+1)≤(qρ2)n℘(μ0,μ1) |
for all n≥0. If there exists a positive integer N such that for n≥N and l≥1,
℘(μn+l,μn)≤℘(μn,μn+1)+℘(μn+1,μn+2)+...+℘(μn+l−1,μn+l)≤(qρ2)n1−qρ2℘(μ0,μ1). | (3.1) |
Choosing 1<q<min{1ρ1,1ρ2} and taking limit as n→∞, we have (μn) is a Cauchy sequence in N. Let u∈N be the limit point of (μn). We show that u∈Q2. Otherwise, let N be the positive integer such that D(u,Q2μn)>0=℘(u,μn), ∀n≥N. Then ℘(μ,μn+1)>℘(u,μn) for all n≥N, a contradiction. Hence there exist a subsequence μn(j) such that for all j∈N, we have
φ(s)max{D(μn(j),Q2μn(j)),12D(u,Q2μn(j))}≤max{℘(μn(j),μn(j)+1),D(u,Q2μn(j))}≤{℘(μn(j),μn(j)+1),℘(u,μn(j)+1)}≤{℘(μn(j),μn(j)+1),℘(u,μn(j))+℘(μn(j),μn(j)+1)}=℘(u,μn(j))+℘(μn(j),μn(j)+1)=℘(u,μn(j)) |
(since ℘(μn(j),μn(j)+1)→0 as j→∞). Above inequality gives
0≤ζ(H(Q2μn(j),Q2u),ρR(μn(j),u)) |
implies
H(Q2μn(j),Q2u)≤ρR(μn(j),u). |
Thus
D(u,Q2u)=limj→∞D(μn(j)+1,Q2u)≤limj→∞H(Q2μn(j),Q2u)≤limj→∞ρ2℘(μn(j),u)=0. |
So u∈Fix(Q2). In (2.1), letting l→∞, we get
℘(μn,u)≤(qρ2)n1−qρ2℘(μ0,μ1) |
for all n∈N. Then
℘(μ0,u)≤11−qρ2℘(μ0,μ1)≤qρ1−qρ2. |
On the same lines, we have that for each u0∈Fix(Q2), there exists μ∈Fix(Q1) such that
℘(u0,μ)≤11−qρ2℘(u0,u1)≤qρ1−qρ2. |
Hence
H(Fix(Q1),Fix(Q2))≤qρ1−max{qρ1,qρ2}. |
Taking limit as q↘1, we get the desired result.
A variety of fixed point theorems have been used to explore the existence and uniqueness of solution of functional equations arising in dynamic programming. Utilizing Theorem 2.3, we explore the existence and uniqueness of solution for a nonlinear functional equation. From now on, W1(R) and W2(R) are assumed to be Banach spaces and U1⊂W1, U2⊂W2. Class of all bounded real-valued functions on U1 is denoted by B(U1). The metric space pair B(U1,℘) where
dB(f,g)=supμ∈U1|f(μ)−g(μ)|,f,g∈U1, |
is complete. Viewing U1 and U2 as the state and decision spaces respectively, the problem of dynamic programming reduces to the problem of solving the functional equation:
p(μ)=supλ∈U2F(μ,λ,p(η(μ,λ))), |
where η:U1×U2→U1 is the transformation of the process and p(μ) represents the optimal return function with initial functional equation:
p(μ)=supλ∈U2{v(μ,λ)+G(μ,λ,p(η(μ,λ)))},μ∈U1, | (4.1) |
where v:U1×U2→R and G:U1×U2×R→R are bounded functions. Let Q:B(U1)→B(U1) be defined by:
Q(f(μ))=supλ∈U2{v(μ,λ)+G(μ,λ,p(η(μ,λ)))},μ∈U1,f∈B(U1). |
Theorem 4.1. Assume that there exists ρ>0 such that for every (μ,λ)∈U1×U2, f,g∈B(U1) and t∈U1 the inequality
φ(s)max{|f(t)−Q(f(t))|,12|g(t)−Q(f(t))|}≤|f(t)−g(t)| | (4.2) |
implies
|G(μ,λ,f(t))−G(μ,λ,g(t))|≤ρR(f(t),g(t))ρR(f(t),g(t))+1 | (4.3) |
where
R(f(t),g(t))=max{|f(t)−g(t)|,|f(t)−Qf(t)|,|g(t)−Q(g(t))|,|f(t)−Q(g(t))|+|g(t)−Q(f(t))|2}. |
Then (4.1) possesses a bounded solution.
Proof. It is obvious that Q is self map on B(U1). Suppose γ is an arbitrary positive real number and f1,f2∈B(U1). Take μ∈U1 and choose λ1,λ2∈U2 such that
Q(f1(μ))<v(μ,λ1)+G(μ,λ1,f1(η1))+γ | (4.4) |
Q(f2(μ))<g(μ,λ2)+G(μ,λ2,f2(η2))+γ, | (4.5) |
where ηi=ηi(μ,λi), i=1,2. By definition of Q, we get
Q(f1(μ))≥v(μ,λ2)+G(μ,λ2,f1(η2))Q(f2(μ))≥v(μ,λ1)+G(μ,λ1,f2(η1)). |
If the inequality (4.2) holds with f=f1,g=f2, then from (4.3), (4.4) and (4.6), we have
Q(f1(μ))−Q(f2(μ))<G(μ,λ1,f1(η1))−G(μ,λ1,f2(η1))+γ≤ρR(f1(t),f2(t))ρR(f1(t),f2(t))+1+γ. | (4.6) |
Similarly, from (4.3), (4.5) and (4.6), we have
Q(f2(μ))−Q(f1(μ))<ρR(f1(t),f2(t))ρR(f1(t),f2(t))+1+γ. | (4.7) |
Hence, from (4.6) and (4.7), we obtain that
|Q(f1(μ))−Q(f2(μ))|<ρR(f1(t),f2(t))ρR(f1(t),f2(t))+1+γ, |
since the inequality holds for any μ∈U1 and γ>0, we get that
φ(s)max{dB(f1,Q(f1)),12dB(f2,Q(f1))}≤dB(f1,f2) |
implies
dB(Q(f1),Q(f2))≤ρmax{dB(f1,f2),db(f1,Qf1),dB(f2,Qf2),dB(f1,Qf2)+dB(f2,Qf1)2}, |
so if we choose ζ(α,β)=ββ+1−α, then we have
ζ(dB(Qf1,Qf2),ρR(f1,f2))≥0. |
Hence, all the assertions of Theorem 2.3 are verified and therefore we get the conclusion.
Example 4.1. Let W1=R=W2, U1=[1,50], U2=R+. We define η:U1×U2→U1, v:U1×U2→R and G:U1×U2R→R by
η(μ,λ)=50μ+λ6μ2+λ6, |
v(μ,λ)=sin(μ+λ2)7λ2+λ+1 |
and
A(μ,λ,z)=z|sin(μ2+λ2). |
Now for s=0.5, φ(s)=12. Therefore for ρ=0.65, it is clear that (4.2)-(4.3) are satisfied. Hence from Theorem 2.3 that the functional equation (4.1) possesses a bounded solution.
In this paper we introduced a new type of set valued contraction in Suzuki sense together with simulation functions and proved some new fixed point and data dependence theorems for such class of mappings. We gave an example to prove the validity of our results. At last, we presented an application to nonlinear functional equation arising in dynamical system to show the usability of obtained results.
The author Thabet Abdeljawad would like to thank Prince Sultan University for funding this work through the research group Nonlinear Analysis Methods in Applied Mathematics (NAMAM)(Group Number RG-DES-2017-01-17).
The authors declare no conflict of interest.
[1] |
I. Altun, H. A. Hnacer, G. Minak, On a general class of weakly picard operators, Miskolc Math. Notes, 16 (2015), 25–32. doi: 10.18514/MMN.2015.1168
![]() |
[2] |
S. Banach, Sur les opérations dans les ensembles abstraits et leur application aux equations intégrales, Fundamenta Math., 3 (1922), 133–181. doi: 10.4064/fm-3-1-133-181
![]() |
[3] |
G. Durmaz, Some theorems for a new type of multivalued contractive maps on metric space, Turkish J. Math., 41 (2017), 1092–1100. doi: 10.3906/mat-1510-75
![]() |
[4] |
G. Durmaz, I. Altun, A new perspective for multivalued weakly picard operators, Publications De L'institut Math., 101 (2017), 197–204. doi: 10.2298/PIM1715197D
![]() |
[5] |
A. A. Eldred, J. Anuradha, P. Veeramani, On equivalence of generalized multivalued contractions and Nadler's fixed point theorem, J. Math. Anal. Appl., 336 (2007), 751–757. doi: 10.1016/j.jmaa.2007.01.087
![]() |
[6] |
H. A. Hancer, G. Mmak, I. Altun, On a broad category of multivalued weakly Picard operators, Fixed Point Theory, 18 (2017), 229–236. doi: 10.24193/fpt-ro.2017.1.19
![]() |
[7] |
M. Jleli, B. Samet, A new generalization of the Banach contraction principle, J. Inequal. Appl., 2014 (2014), 8. doi: 10.1186/1029-242X-2014-8
![]() |
[8] | Z. Kadelburg, S. Radenovi, A note on some recent best proximity point results for non-self mappings, Gulf J. Math., 1 (2013), 36–41. |
[9] | T. Kamran, S. Hussain, Weakly (s, r)-contractive multi-valued operators, Rend. Circ. Mat. Palermo, 64 (2015), 475–482. |
[10] |
E. Karapinar, Fixed points results via simulation functions, Filomat, 30 (2016), 2343–2350. doi: 10.2298/FIL1608343K
![]() |
[11] |
F. Khojasteh, S. Shukla, S. Radenović, A new approach to the study of fixed point theorems via simulation functions, Filomat, 29 (2015), 1189–1194. doi: 10.2298/FIL1506189K
![]() |
[12] | S. B. Nadler Jr, Multivalued contraction mappings, Pacific J. Math., 30 (1969), 475–488. |
[13] |
O. Popescu, A new type of contractive multivalued operators, Bull. Sci. Math., 137 (2013), 30–44. doi: 10.1016/j.bulsci.2012.07.001
![]() |
[14] |
S. Radenović, S. Chandok, Simulation type functions and coincidence points, Filomat, 32 (2018), 141–147. doi: 10.2298/FIL1801141R
![]() |
[15] |
A. Rold, E. Karapinar, C. Rold, J. Martinez, Coincidence point theorems on metric spaces via simulation functions, J. Comput. Appl. Math., 275 (2015), 345–355. doi: 10.1016/j.cam.2014.07.011
![]() |
[16] | I. A. Rus, Basic problems of the metric fixed point theory revisited (II), Stud. Univ. Babes-Bolyai, 36 (1991), 81–89. |
[17] | J. Von Neuman, Über ein ökonomisches Gleichungssystem und eine Verallgemeinerung des Brouwerschen Fixpunktsatzes, Ergebn. Math. Kolloq., 8 (1937), 73–83. |
1. | Dumitru Baleanu, Muhammad Qamar Iqbal, Azhar Hussain, Sina Etemad, Shahram Rezapour, On solutions of fractional multi-term sequential problems via some special categories of functions and (AEP)-property, 2021, 2021, 1687-1847, 10.1186/s13662-021-03356-2 | |
2. | Arshad Khan, Muhammad Sarwar, Farhan Khan, Habes Alsamir, Hasanen A. Hammad, Santosh Kumar, Fixed Point Results for Multivalued Mappings with Applications, 2021, 2021, 2314-8888, 1, 10.1155/2021/9921728 |