Dividend policies play a pivotal role in financial management by aiming to maximize shareholders' interest and effectively managing risk. In this paper, we explore the optimal dividend strategy in a discrete-time compound binomial dual risk framework. This model is suitable for a company whose income comes from occasional operating expenses and settlements only once per unit of time. We assume that expenses are subject to dynamic changes influenced by economic factors, following a Markov chain. With or without a ceiling constraint on dividend payments, we prove that the optimal value function serves as the exclusive solution to a discrete Hamilton-Jacobi-Bellman (HJB) equation through the utilization of the fixed-point theorem. Furthermore, we derive a straightforward computational approach for determining the optimal strategy. Finally, we provide numerical examples to illustrate the theoretical findings and calculation methods.
Citation: Li Deng, Zhichao Chen. Optimal dividends in a discrete-time dual risk model with stochastic expenses[J]. AIMS Mathematics, 2024, 9(11): 31696-31720. doi: 10.3934/math.20241524
[1] | Mahmoud S. Mehany, Faizah D. Alanazi . An η-Hermitian solution to a two-sided matrix equation and a system of matrix equations over the skew-field of quaternions. AIMS Mathematics, 2025, 10(4): 7684-7705. doi: 10.3934/math.2025352 |
[2] | Abdur Rehman, Muhammad Zia Ur Rahman, Asim Ghaffar, Carlos Martin-Barreiro, Cecilia Castro, Víctor Leiva, Xavier Cabezas . Systems of quaternionic linear matrix equations: solution, computation, algorithm, and applications. AIMS Mathematics, 2024, 9(10): 26371-26402. doi: 10.3934/math.20241284 |
[3] | Vladislav N. Kovalnogov, Ruslan V. Fedorov, Denis A. Demidov, Malyoshina A. Malyoshina, Theodore E. Simos, Spyridon D. Mourtas, Vasilios N. Katsikis . Computing quaternion matrix pseudoinverse with zeroing neural networks. AIMS Mathematics, 2023, 8(10): 22875-22895. doi: 10.3934/math.20231164 |
[4] | Wenxv Ding, Ying Li, Anli Wei, Zhihong Liu . Solving reduced biquaternion matrices equation k∑i=1AiXBi=C with special structure based on semi-tensor product of matrices. AIMS Mathematics, 2022, 7(3): 3258-3276. doi: 10.3934/math.2022181 |
[5] | Yang Chen, Kezheng Zuo, Zhimei Fu . New characterizations of the generalized Moore-Penrose inverse of matrices. AIMS Mathematics, 2022, 7(3): 4359-4375. doi: 10.3934/math.2022242 |
[6] | Abdur Rehman, Ivan Kyrchei, Muhammad Zia Ur Rahman, Víctor Leiva, Cecilia Castro . Solvability and algorithm for Sylvester-type quaternion matrix equations with potential applications. AIMS Mathematics, 2024, 9(8): 19967-19996. doi: 10.3934/math.2024974 |
[7] | Anli Wei, Ying Li, Wenxv Ding, Jianli Zhao . Three special kinds of least squares solutions for the quaternion generalized Sylvester matrix equation. AIMS Mathematics, 2022, 7(4): 5029-5048. doi: 10.3934/math.2022280 |
[8] | Dong Wang, Ying Li, Wenxv Ding . The least squares Bisymmetric solution of quaternion matrix equation AXB=C. AIMS Mathematics, 2021, 6(12): 13247-13257. doi: 10.3934/math.2021766 |
[9] | Qi Xiao, Jin Zhong . Characterizations and properties of hyper-dual Moore-Penrose generalized inverse. AIMS Mathematics, 2024, 9(12): 35125-35150. doi: 10.3934/math.20241670 |
[10] | Vladislav N. Kovalnogov, Ruslan V. Fedorov, Denis A. Demidov, Malyoshina A. Malyoshina, Theodore E. Simos, Vasilios N. Katsikis, Spyridon D. Mourtas, Romanos D. Sahas . Zeroing neural networks for computing quaternion linear matrix equation with application to color restoration of images. AIMS Mathematics, 2023, 8(6): 14321-14339. doi: 10.3934/math.2023733 |
Dividend policies play a pivotal role in financial management by aiming to maximize shareholders' interest and effectively managing risk. In this paper, we explore the optimal dividend strategy in a discrete-time compound binomial dual risk framework. This model is suitable for a company whose income comes from occasional operating expenses and settlements only once per unit of time. We assume that expenses are subject to dynamic changes influenced by economic factors, following a Markov chain. With or without a ceiling constraint on dividend payments, we prove that the optimal value function serves as the exclusive solution to a discrete Hamilton-Jacobi-Bellman (HJB) equation through the utilization of the fixed-point theorem. Furthermore, we derive a straightforward computational approach for determining the optimal strategy. Finally, we provide numerical examples to illustrate the theoretical findings and calculation methods.
The concept of fixed point theory has seen diverse extensions beyond the classical idea of a metric space. One notable variation is the b-metric space, also known as a metric type space according to certain authors. Bakhtin [1] introduced and extensively investigated this space, and it has been extensively employed by Czerwik in subsequent study [2]. Furthermore, Czerwik extended the classical Banach contraction principle in the context of a b-metric space that is complete. Following this, a collection of papers has been devoted to refining fixed point results for both single and multi-valued operators in b-metric space, taking into account diverse topological properties. Interested readers can explore some of these contributions in [3,4,5,6,7,8,9,10,11,12] and related references.
Bhaskar and Lakshmikantham [13] were the pioneers who introduced coupled fixed points to specific maps in partially ordered metric space. Their findings were employed to investigate solutions' existence and uniqueness for boundary value problems. Subsequently, Lakshmikantham and Ćirić [14] extended this work by introducing coupled coincidence, which is coupled common fixed point results for nonlinear contractive maps with mixed monotone properties in partially ordered metric space that is complete, thereby generalizing the findings from [13]. Since then, various authors have further generalized and improved the fixed point results in both partially ordered metric spaces and partially ordered b-metric spaces. Significant contributions have been made by the authors of papers [15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36]. Recent studies by Mituku et al. [37] and Rao et al. [38,39,40,41,42] have yielded results related to fixed points, coincidence points and coupled coincidence points of self-maps that adhere to generalized weak contractions in partially ordered b-metric spaces, taking into account necessary topological properties. The Banach fixed point theorem plays an essential role in the fields of optimization and inverse problems some of such works can be seen in [43,44,45].
This study was designed to demonstrate various types of results, including those on fixed points, coincidence points, coupled coincidence points and coupled common fixed points of the maps that fulfill the (ϕ,ψ)-weak contraction condition with auxiliary functions within a partially ordered b-metric space that is complete. These outcomes expand and generalize the findings from [13,14,37] and several analogous results found in the literature, as presented in [38,46]. To reinforce the validity of the results, the paper includes illustrative examples.
The forthcoming results often rely on the following definitions.
Definition 1. [38] A function ω: L×L→[0,+∞), with L≠∅ is termed as a b-metric, if it adheres to the specified properties for o1,o2,o3∈L and s≥1∈R:
(a) ω(o1,o2)=0 if and only if o1=o2.
(b) ω(o1,o2)=ω(o2,o1).
(c) ω(o1,o2)≤s(ω(o1,o3)+ω(o3,o2)).
Subsequently, the triple (L,ω,s) is referred to as a b-metric space. If the pair (L,⪯) continues as a partially ordered set (p.o.s), then (L,ω,s,⪯) is termed a partially ordered b-metric space (p.o.m.s).
Definition 2. [46] A sequence {on} in a b-metric space (L,ω,s) adheres to the following conditions:
(1) It converges to o if limn→+∞ω(on,o)=0, denoted as limn→+∞on=o,
(2) It is considered a Cauchy sequence in L when limn,m→+∞ω(on,om)=0.
Completeness of (L,ω,s) is affirmed when all Cauchy sequences within it converge.
Definition 3. In cases in which the metric ω achieves completeness, the term complete partially ordered b-metric space (c.m.s) is applied to characterize the structure (L,ω,s,⪯).
Definition 4. [46] Suppose that U and g are two self mappings defined on a partially ordered set (L,⪯). In this scenario, the following holds:
(1) A map U is considered monotone non-decreasing, if for every pair o and a in L such that o⪯a, the condition U(o)⪯U(a) is satisfied.
(2) When g(o)=U(o) or g(o)=U(o)=o for o∈L, then o is referred to as being a coincidence point or a common fixed point of g and U.
(3) g and U are considered commuting if g(U(o))=U(g(o)),∀o∈L holds.
(4) The mappings U and g are known to be compatible, if limn→+∞ω(Ugon,gUon)=0 whenever a sequence {on}∈L satisfies that limn→+∞Uon=limn→+∞gon=o, for some o∈L.
(5) If g(U(o))=U(g(o)) whenever U(o)=g(o) then the maps g, U are referred to be weakly compatible.
(6) U is referred to as monotonically g non-decreasing if for all o,a∈L, the condition g(o)⪯g(a) implies that U(o)⪯U(a).
(7) A set L≠∅ is well-ordered if either o⪯a or a⪯o for every pair o, a∈L.
Definition 5. [14,15] Take a partially ordered set (L,⪯), and let us assume that there are two mappings U: L×L→L and g: L→L. In this scenario, the following holds:
(1) A map U is considered to exhibit a mixed g-monotone property, if it is increasing and decreasing g-monotone in its first and second arguments. In formal terms, for all o,a∈L,
o1,o2∈L,go1⪯go2⇒U(o1,a)⪯U(o2,a) |
and
a1,a2∈L,ga1⪯ga2⇒U(o,a1)⪰U(o,a2). |
In the specific scenario in which g=I, I is an identity mapping, U is described as having the mixed monotone property.
(2) (o,a)∈L×L with U(o,a)=g(o) and U(a,o)=g(a) is referred to as a coupled coincidence point for U and g. It is important to emphasize that if g=I, then (o,a) is denoted as a coupled fixed point of U.
(3) If
U(o,o)=g(o)=o, |
then o∈L is designated as a common fixed point for U, g.
(4) If
U(g(o),g(a))=g(U(o,a)) |
for all o,a∈L, then U, g are commutative.
(5) U and g are termed compatible, if
limn→+∞ω(g(U(on,an)),U(gon,gan))=0 |
and
limn→+∞ω(g(U(an,on)),U(gan,gon))=0, |
whenever the sequences {on} and {an} in L satisfy
limn→+∞U(on,an)=limn→+∞g(on)=o |
and
limn→+∞U(an,on)=limn→+∞g(an)=a |
for some o,a∈L.
The upcoming lemma is regularly utilized in our results within a b-metric space to verify the convergence of the sequences.
Lemma 6. [15] Consider a b-metric space (L,ω,s,⪯) with s≥1. Assuming that the sequences {on} and {an} b-converge to o and a respectively, then
1s2ω(o,a)≤limn→+∞infω(on,an)≤limn→+∞supω(on,an)≤s2ω(o,a). |
In particular, when o=a, the limit as n approaches infinity of ω(on,an) is zero. Also, for every d∈L, the subsequent inequalities are applicable:
1sω(o,d)≤lim infn→+∞ω(on,d)≤lim supn→+∞ω(on,d)≤sω(o,d). |
The subsequent distance functions are utilized consistently in the present work.
(i) A continuous, non-decreasing self-map ϕ over [0,+∞) is referred to as an alternating distance function if it satisfies the condition ϕ(t)=0 if and only if t=0. Refer to all such functions as Φ.
(ii) Ψ represents the collection of self-maps ψ over [0,+∞) with the characteristics that ψ is lower semi-continuous and ψ(t)=0 if and only if t=0.
For a self-map U over a partially ordered b-metric space (L,ω,s,⪯) let
G(o,a)=max{ω(a,Ua)[1+ω(o,Uo)]1+ω(o,a),ω(o,Uo)ω(a,Ua)1+ω(o,a),ω(o,Ua)+ω(a,Uo)2s,ω(o,a)} | (3.1) |
and
H(o,a)=max{ω(a,Ua)[1+ω(o,Uo)]1+ω(o,a),ω(o,a)}. | (3.2) |
Then, for ϕ∈Φ and ψ∈Ψ, U is called a (ϕ,ψ)-generalized contraction, if it adheres to the following condition:
ϕ(sω(Uo,Ua))≤ϕ(G(o,a))−ψ(H(o,a)). | (3.3) |
From here, we begin with the below theorem regarding fixed points in a b-metric space that is ordered.
Theorem 7. A non-decreasing self-map U on a c.m.s (L,ω,s,⪯) is continuous and satisfies the conditions of a (ϕ,ψ)-generalized contraction with respect to ⪯. If there exists o0∈L with o0⪯Uo0, then U admits a fixed point in L.
Proof. If a particular o0∈L satisfies that Uo0=o0, the proof is concluded. Assuming that o0≺Uo0, we form a sequence {on}⊂L through recursive definitions, where on+1=Uon for n≥0. By leveraging the non-decreasing characteristic of U, we infer, through an inductive argument, that
o0≺Uo0=o1⪯⋯⪯on⪯Uon=on+1⪯⋯. | (3.4) |
If there exists n0∈N such that on0=on0+1 as indicated by (3.4), it implies that on0 serves as a fixed point for U and therefore, no further proof is necessary. Let us consider the scenario in which on≠on+1,∀n≥1 holds. Applying the condition (3.3), we can conclude that
ϕ(ω(on,on+1))=ϕ(ω(Uon−1,Uon))≤ϕ(sω(Uon−1,Uon))≤ϕ(G(on−1,on))−ψ(H(on−1,on)). | (3.5) |
From (3.5), we acquire
ω(on,on+1)=ω(Uon−1,Uon)≤1sG(on−1,on), | (3.6) |
where
G(on−1,on)=max{ω(on,Uon)[1+ω(on−1,Uon−1)]1+ω(on−1,on),ω(on−1,Uon−1)ω(on,Uon)1+ω(on−1,on)ω(on−1,Uon)+ω(on,Uon−1)2s,ω(on−1,on)}=max{ω(on,on+1),ω(on−1,on)ω(on,on+1)1+ω(on−1,on),ω(on−1,on+1)+ω(on,on)2s,ω(on−1,on)}≤max{ω(on,on+1),ω(on−1,on)+ω(on,on+1)2,ω(on−1,on)}≤max{ω(on,on+1),ω(on−1,on)}. | (3.7) |
If
max{ω(on,on+1),ω(on−1,on)}=ω(on,on+1) |
for some n≥1, then from (3.6), it follows that
ω(on,on+1)≤1sω(on,on+1), |
which leads to a contradiction. Consequently,
max{ω(on,on+1),ω(on−1,on)}=ω(on−1,on) |
for n≥1. So, we infer from (3.6) that
ω(on,on+1)≤1sω(on−1,on). |
Given that 1s∈(0,1), it follows, according to [1,3,4,7], that the sequence {on} is a Cauchy sequence. Therefore, on→v in L by its completeness.
Additionally, the continuity of U results in
Uv=U(limn→+∞on)=limn→+∞Uon=limn→+∞on+1=v. | (3.8) |
Consequently, v constitutes a fixed point of U within the set L.
By easing the conditions for continuity on a mapping U as per Theorem 7, the subsequent result is obtained.
Theorem 8. Assume that a non-decreasing self-maping U on a c.m.s (L,ω,s,⪯), where ω is continuous in each variable satisfying the conditions of a (ϕ,ψ)-generalized contraction with respect to ⪯ and there exists o0∈L such that o0⪯Uo0. If L satisfies the condition that, for every non-decreasing sequence {on} in L such thaton→v∈L, then on⪯v,∀n≥1, i.e., v=supn∈Non, then U possesses a fixed point in L.
Proof. We form a non-decreasing Cauchy sequence {on} within L that converges to v∈L by applying Theorem 7. Therefore, based on the provided hypotheses, it follows that on⪯v for every n∈N, leading to the conclusion that v=supn∈Non.
Subsequently, we demonstrate that v serves as a fixed point of U in L. Let us assume the contrary, i.e., Uv≠v; then,
G(on,v)=max{ω(v,Uv)[1+ω(on,Uon)]1+ω(on,v),ω(on,Uon)ω(v,Uv)1+ω(on,v),ω(on,Uv)+ω(v,Uon)2s,ω(on,v)} | (3.9) |
and
H(on,v)=max{ω(v,Uv)[1+ω(on,Uon)]1+ω(on,v),ω(on,v)}. | (3.10) |
Taking the limit as n→∞ in (3.9) and (3.10), and by using the continuity in each variable of ω, we obtain
limn→+∞G(on,v)=max{ω(v,Uv),0,ω(v,Uv)s,0}=ω(v,Uv) | (3.11) |
and
limn→+∞H(on,v)=max{ω(v,Uv),0}=ω(v,Uv). | (3.12) |
Considering on⪯v, for all n, we can derive from condition (3.3) that
ϕ(ω(on+1,Uv))=ϕ(ω(Uon,Uv)≤ϕ(sω(Uon,Uv)≤ϕ(G(on,v))−ψ(H(on,v)). | (3.13) |
Allowing n to tend to infinity and employing (3.11) and (3.12), we obtain
ϕ(ω(v,Uv))≤ϕ(ω(v,Uv))−ψ(ω(v,Uv))<ϕ(ω(v,Uv)), | (3.14) |
which is absurd, unless Uv=v, indicating that U possesses a fixed point v in L.
We will now present a condition that is adequate to guarantee the existence of a unique fixed point in Theorems 7 and 8.
For anyo,a∈L,there exists somew∈Lthat is comparable to botho,a. | (3.15) |
Theorem 9. Consider a non-decreasing self-mapping U on a c.m.s (L,ω,s,⪯) that satisfies the conditions of a (ϕ,ψ)-generalized contraction with respect to ⪯ and there exists o0∈L such that o0⪯Uo0. If L satisfies the condition (3.15) in Theorem 7 (or Theorem 8), then U has a unique fixed point in L.
Proof. We infer the existence of a nonempty set of fixed points for U based on Theorem 7 (or Theorem 8). Assuming that o∗, a∗ are two fixed points for U, we assert that o∗=a∗.
If o∗≠a∗, according to the given assumptions, we obtain
ϕ(ω(Uo∗,Ua∗))≤ϕ(sω(Uo∗,Ua∗))≤ϕ(G(o∗,a∗))−ψ(H(o∗,a∗)). | (3.16) |
Consequently, we acquire
ω(o∗,a∗)=ω(Uo∗,Ua∗)≤1sG(o∗,a∗), | (3.17) |
where
G(o∗,a∗)=max{ω(a∗,Ua∗)[1+ω(o∗,Uo∗)]1+ω(o∗,a∗),ω(o∗,Uo∗)ω(a∗,Ua∗)1+ω(o∗,a∗),ω(o∗,Ua∗)+ω(a∗,Uo∗)2s,ω(o∗,a∗)}=max{ω(a∗,a∗)[1+ω(o∗,o∗)]1+ω(o∗,a∗),ω(o∗,o∗)ω(a∗,a∗)1+ω(o∗,a∗),ω(o∗,a∗)+ω(a∗,o∗)2s,ω(o∗,a∗)}=max{0,0,ω(o∗,a∗)s,ω(o∗,a∗)}=ω(o∗,a∗). | (3.18) |
Therefore from (3.17), we derive that
ω(o∗,a∗)≤1sω(o∗,a∗)<ω(o∗,a∗). | (3.19) |
This results in a contradiction. Thus, we conclude that o∗=a∗, completing the proof.
Consider a p.o.m.s (L,ω,s,⪯) with s>1 and let U,g: L→L be two mappings. Define:
Gg(o,a)=max{ω(ga,Ua)[1+ω(go,Uo)]1+ω(go,ga),ω(go,Uo)ω(ga,Ua)1+ω(go,ga),ω(go,Ua)+ω(ga,Uo)2s,ω(go,ga)} | (3.20) |
and
Hg(o,a)=max{ω(ga,Ua)[1+ω(go,Uo)]1+ω(go,ga),ω(go,ga)}. | (3.21) |
Definition 10. A function U: L→L, where (L,ω,s,⪯) is a p.o.m.s satisfying the following condition is termed as a (ϕ,ψ)-generalized contraction by a self-mapping g over L:
ϕ(sω(Uo,Ua))≤ϕ(Gg(o,a))−ψ(Hg(o,a)), | (3.22) |
where ϕ∈Φ, ψ∈Ψ, o,a∈L with go⪯ga and where Gg(o,a) and Hg(o,a) are expressed as per (3.20) and (3.21), respectively.
Theorem 11. The mappings U and g in Definition 10 possess a coincidence point over a c.m.s. (L,ω,s,⪯) if there is an element o0∈L such that go0⪯Uo0 when U, g are continuous, UL⊆gL and U is compatible with g and a monotone g-non-decreasing mapping.
Proof. Utilizing the proof technique outlined in [18, Theorem 2.2], we construct two sequences {on} and {an} in L such that
an=Uon=gon+1,∀n≥0, | (3.23) |
for which
go0⪯go1⪯⋯⪯gon⪯gon+1⪯⋯. | (3.24) |
Again from [18], we can assume that an≠an+1 for all n≠0. For this case we have to show that
ω(an,an+1)≤λω(an−1,an) | (3.25) |
for all n≥1, where λ∈[0,1s].
Now, employing (3.22) and utilizing (3.23) and (3.24), we obtain that
ϕ(sω(an,an+1))=ϕ(sω(Uon,Uon+1))≤ϕ(Gg(on,on+1))−ψ(Hg(on,on+1)), | (3.26) |
where
Gg(on,on+1)=max{ω(gon+1,Uon+1)[1+ω(gon,Uon)]1+ω(gon,gon+1),ω(gon,Uon)ω(gon+1,Uon+1)1+ω(gon,gon+1),ω(gon,Uon+1)+ω(gon+1,Uon)2s,ω(gon,gon+1)}=max{ω(an,an+1)[1+ω(an−1,an)]1+ω(an−1,an),ω(an−1,an)ω(an,an+1)1+ω(an−1,an),ω(an−1,an+1)+ω(an,an)2s,ω(an−1,an)}≤max{ω(an,an+1),ω(an−1,an)+ω(an,an+1)2,ω(an−1,an)}≤max{ω(an−1,an),ω(an,an+1)} | (3.27) |
and
Hg(on,on+1)=max{ω(gon+1,Uon+1)[1+ω(gon,Uon)]1+ω(gon,gon+1),ω(gon,gon+1)}=max{ω(an,an+1)[1+ω(an−1,an)]1+ω(an−1,an),ω(an−1,an)}=max{ω(an−1,an),ω(an,an+1)}. | (3.28) |
Consequently, based on (3.26)–(3.28), we obtain
ϕ(sω(an,an+1))≤ϕ(max{ω(an−1,an),ω(an,an+1)})−ψ(max{ω(an−1,an),ω(an,an+1)}). | (3.29) |
For n∈N, 0<ω(an−1,an)≤ω(an,an+1); then, (3.29) becomes
ϕ(sω(an,an+1))≤ϕ(ω(an,an+1))−ψ(ω(an,an+1))<ϕ(ω(an,an+1)), |
or equivalently
sω(an,an+1)≤ω(an,an+1). |
This is a contradiction. Hence from (3.29) we obtain that
sω(an,an+1)≤ω(an−1,an). | (3.30) |
Therefore, (3.25) is valid, where λ=1s∈[0,1s]. because s>1, it follows that 0≤λ=1s<1. Now combining (3.25) and Lemma 3.1 of [8], we deduce that
{an}={Uon}={gon+1} |
forms a convergent Cauchy sequence in L which converges to some v∈L due to the completeness of (L,ω). i.e.,
limn→+∞Uon=limn→+∞gon+1=v. |
Hence, considering the compatibility of U and g, we conclude that
limn→+∞ω(gUon,Ugon)=0. | (3.31) |
Furthermore, utilizing the continuity properties of both U and g, we acquire
limn→+∞gUon=gv,limn→+∞Ugon=Uv. | (3.32) |
Moreover, by employing the triangle inequality and utilizing Eqs (3.31) and (3.32), we derive
1sω(Uv,gv)≤ω(Uv,U(gon))+sω(U(gon),g(Uon))+sω(g(Uon),gv). | (3.33) |
In conclusion, as n→+∞ in (3.33), we obtain that ω(Uv,gv)=0. Therefore, v is a coincidence point of U and g. Hence, we have the result.
By relaxing the continuity requirements imposed on g and U as stated in Theorem 11, we arrive at the following result:
Theorem 12. Consider an increasing sequence {gon}⊂L of Theorem 11 such that
(i) limn→+∞gon=goingL;
(ii) gLis a closed subset ofL;
(iii) gon⪯go and go⪯g(go)forn∈N.
Hence, a coincidence point exists for weakly compatible mappings U, g if go0⪯Uo0 for o0∈L. Additionally, at coincidence points U, g are commutative; then a common fixed point in L exists for U and g.
Proof. The sequence
{an}={Uon}={gon+1} |
forms a Cauchy sequence, as established in the proof of Theorem 11. Because of the closed nature of gL, we have
limn→+∞Uon=limn→+∞gon+1=gvforv∈L. |
Therefore, based on the given assumptions, it follows that gon⪯gv, for all n≥1. The subsequent step establishes that v serves as a coincidence point for both U and g.
According to (3.22), it follows that
ϕ(sω(Uon,Uo))≤ϕ(Gg(on,o))−ψ(Hg(on,o)), | (3.34) |
where
Gg(on,v)=max{ω(gv,Uv)[1+ω(gon,Uon)]1+ω(gon,gv),ω(gon,Uon)ω(gv,Uv)1+ω(gon,gv),ω(gon,Uv)+ω(gv,Uon)2s,ω(gon,gv)}→max{ω(gv,Uv),0,ω(gv,Uv)2s,0}=ω(gv,Uv)asn→+∞ |
and
Hg(on,v)=max{ω(gv,Uv)[1+ω(gon,Uon)]1+ω(gon,gv),ω(gon,gv)}→max{ω(gv,Uv),0}=ω(gv,Uv)asn→+∞. |
Consequently, (3.34) transforms into the following:
ϕ(slimn→+∞ω(Uon,Uo))≤ϕ(ω(gv,Uv))−ψ(ω(gv,Uv))<ϕ(ω(gv,Uv)). | (3.35) |
As a result, we obtain
limn→+∞ω(Uon,Uo)<1sω(gv,Uv). | (3.36) |
Moreover, through the triangular inequality, we can express the following:
1sω(gv,Uv)≤ω(gv,Uon)+ω(Uon,Uv). | (3.37) |
Subsequently, if gv≠Uv, a contradiction arises from (3.36) and (3.37). Therefore, gv=Uv. Assuming that gv=Uv=ρ, which indicates the commutative property of U and g at ρ, we have that Uρ=U(gv)=g(Uv)=gρ. Given that gv=g(gv)=gρ, applying (3.34) with gv=Uv and gρ=Uρ yields
ϕ(sω(Uv,Uρ))≤ϕ(Gg(v,ρ))−ψ(Hg(v,ρ))<ϕ(ω(Uv,Uρ)), |
or equivalently,
sω(Uv,Uρ)≤ω(Uv,Uρ), |
which results in a contradiction, if Uv≠Uρ. Therefore, it follows that Uv=Uρ=ρ. Consequently, Uv=gρ=ρ, showing that ρ is a common fixed point for both U and g.
Definition 13. A mapping U: L×L→L, where (L,ω,s,⪯) is a c.m.s with respect to the function g: L→L is denoted as a (ϕ,ψ)-generalized contractive mapping if
ϕ(skω(U(o,a),U(ρ,d)))≤ϕ(Gg(o,a,ρ,d))−ψ(Hg(o,a,ρ,d)), | (3.38) |
where k>2, go⪯gρ, for o,a,ρ,d∈L and
Gg(o,a,ρ,d)=max{ω(gρ,U(ρ,d))[1+ω(go,U(o,a))]1+ω(go,gρ),ω(go,U(o,a))ω(gρ,U(ρ,d))1+ω(go,gρ),ω(go,U(ρ,d))+ω(gρ,U(o,a))2s,ω(go,gρ)} |
and
Hg(o,a,ρ,d)=max{ω(gρ,U(ρ,d))[1+ω(go,U(o,a))]1+ω(go,gρ),ω(go,gρ)}. |
Theorem 14. A coupled coincidence point for the mappings U, g exists in Definition 13 over a c.m.s (L,ω,s,⪯) when there exists (o0,a0)∈L×L with fo0⪯U(o0,a0), ga0⪰U(a0,o0), U(L×L)⊆g(L), U and commutes with g and has a mixed g-monotone property.
Proof. Based on the given assumptions and by following the argument presented in the proof of [18, Theorem 2.2], we form two sequences {on}, {an} within the space L such that
gon+1=U(on,an),gan+1=U(an,on)for alln≥0. |
Specifically, the sequences {gon} and {gan} are, respectively, non-decreasing and non-increasing in L. Substituting o=on,a=an,ρ=on+1 and d=an+1 into (3.38), we obtain
ϕ(skω(gon+1,gon+2))=ϕ(skω(U(on,an),U(on+1,an+1)))≤ϕ(Gg(on,an,on+1,an+1))−ψ(Hg(on,an,on+1,an+1)), | (3.39) |
where
Gg(on,an,on+1,an+1)≤max{ω(gon,gon+1),ω(gon+1,gon+2)} | (3.40) |
and
Hg(on,an,on+1,an+1)=max{ω(gon,gon+1),ω(gon+1,gon+2)}. | (3.41) |
Therefore from (3.39), we have
ϕ(skω(gon+1,gon+2))≤ϕ(max{ω(gon,gon+1),ω(gon+1,gon+2)})−ψ(max{ω(gon,gon+1),ω(gon+1,gon+2)}). | (3.42) |
Similarly by taking o=an+1,a=on+1,ρ=on and d=on in (3.38), we get
ϕ(skω(gan+1,gan+2))≤ϕ(max{ω(gan,gan+1),ω(gan+1,gan+2)})−ψ(max{ω(gan,gan+1),ω(gan+1,gan+2)}). | (3.43) |
Given the relation,
max{ϕ(d1),ϕ(d2)}=ϕ{max{d1,d2}} |
for all d1,d2∈[0,+∞), the combination of (3.42) and (3.43) yields
ϕ(skΥn)≤ϕ(max{ω(gon,gon+1),ω(gon+1,gon+2),ω(gan,gan+1),ω(gan+1,gan+2)})−ψ(max{ω(gon,gon+1),ω(gon+1,gon+2),ω(gan,gan+1),ω(gan+1,gan+2)}), | (3.44) |
where
Υn=max{ω(gon+1,gon+2),ω(gan+1,gan+2)}. | (3.45) |
Let us denote the following
Kn=max{ω(gon,gon+1),ω(gon+1,gon+2),ω(gan,gan+1),ω(gan+1,gan+2)}. | (3.46) |
As a result, by considering (3.42) to (3.45), we get
skΥn≤Kn. | (3.47) |
Subsequently, we establish that
Υn≤λΥn−1 | (3.48) |
for all n≥1, with λ=1sk∈[0,1).
Assume that if Kn=Υn, then according to (3.47), we derive skΥn≤Υn, which gives that Υn=0 because s>1. Consequently, (3.48) is satisfied. Alternatively, if
Kn=max{ω(gon,gon+1),ω(gan,gan+1)} |
denoted as Kn=Υn−1, then (3.47) implies (3.48).
Now, based on (3.47), we can conclude that Υn≤λnΥ0, leading to
ω(gon+1,gon+2)≤λnΥ0andω(gan+1,gan+2)≤λnΥ0. | (3.49) |
Consequently, in accordance with Lemma 3.1 from [8], the sequences {gon} and {gan} are Cauchy sequences in L. Thus, by implementing the remaining procedures detailed in [16, Theorem 2.2], we can demonstrate coincidence point existence for U and g in L.
Corollary 15. Consider a c.m.s (L,ω,s,⪯) where s>1. A continuous mapping U: L×L→L for a mixed monotone property. Assume the existence of ϕ∈Φ and ψ∈Ψ such that
ϕ(skω(U(o,a),U(ρ,d)))≤ϕ(Gg(o,a,ρ,d))−ψ(Hg(o,a,ρ,d)) |
for all o,a,ρ,d∈L satisfying that o⪯ρ and a⪰d, with k>2 and
Gg(o,a,ρ,d)=max{ω(ρ,U(ρ,d))[1+ω(o,U(o,a))]1+ω(o,ρ),ω(o,U(o,a))ω(ρ,U(ρ,d))1+ω(o,ρ),ω(o,U(ρ,d))+ω(ρ,U(o,a))2s,ω(o,ρ)} |
and
Hg(o,a,ρ,d)=max{ω(ρ,U(ρ,d))[1+ω(o,U(o,a))]1+ω(o,ρ),ω(o,ρ)}. |
Under those conditions, U possesses a coupled fixed point in L if (o0,a0)∈L×L satisfies that o0⪯U(o0,a0) and a0⪰U(a0,o0).
Proof. Put g=IL in Theorem 14.
Corollary 16. Consider a c.m.s (L,ω,s,⪯) where s>1 and U: L×L→L be a continuous mapping possessing a mixed monotone property. Suppose the existence of ψ∈Ψ such that
ω(U(o,a),U(ρ,d))≤1skGg(o,a,ρ,d)−1skψ(Hg(o,a,ρ,d)) | (3.50) |
for all o,a,ρ,d∈L with o⪯ρ, a⪰d and k>2, where
Gg(o,a,ρ,d)=max{ω(ρ,U(ρ,d))[1+ω(o,U(o,a))]1+ω(o,ρ),ω(o,U(o,a))ω(ρ,U(ρ,d))1+ω(o,ρ),ω(o,U(ρ,d))+ω(ρ,U(o,a))2s,ω(o,ρ)} | (3.51) |
and
Hg(o,a,ρ,d)=max{ω(ρ,U(ρ,d))[1+ω(o,U(o,a))]1+ω(o,ρ),ω(o,ρ)}. | (3.52) |
If for some (o0,a0)∈L×L such that o0⪯U(o0,a0) and a0⪰U(a0,o0), then U in L possesses a coupled fixed point.
Theorem 17. A coupled common fixed point for U and g is unique in Theorem 14, if for all (y,b)∈L×L, (U(y,b),U(b,y)) is comparable to both (U(o,a),U(a,o)) and (U(r,s),U(s,r)) for some (y,b)∈L×L.
Proof. By virtue of Theorem 14, a coupled coincidence point in L for U and g is established. To prove the uniqueness, let us take two coupled coincidence points (o,a) and (r,s), so that
U(o,a)=go,U(a,o)=ga,U(r,s)=grandU(s,r)=gs. |
Our objective is to demonstrate that go=gr and ga=gs.
Given the provided conditions, (U(y,b),U(b,y)) is comparable to (U(o,a),U(a,o)) and to (U(r,s),U(s,r)) for (y,b)∈L×L. Assume that
(U(o,a),U(a,o))≤(U(y,b),U(b,y))and(U(r,s),U(s,r))≤(U(y,b),U(b,y)). |
Define y0=y and b0=b and select (y1,b1)∈L×L as follows:
gy1=U(y0,b0),gb1=U(b0,y0),(n≥1). |
Construct the sequences {gyn} and {gbn} in L inductively by implementing the above process using the following equations;
gyn+1=U(yn,bn),gbn+1=U(bn,yn),(n≥0). |
In a similar manner, introduce the sequences {gon}, {gan}, {grn} and {gsn} in L according to the previous definitions, with the initial values o0=o, a0=a, r0=r, and s0=s. Additionally, it is apparent that
gon→U(o,a),gan→U(a,o),grn→U(r,s),gsn→U(s,r),(n≥1). |
Given that
(U(o,a),U(a,o))=(go,ga)=(go1,ga1) |
is comparable to
(U(y,b),U(b,y))=(gy,gb)=(gy1,gb1), |
it follows that (go1,ga1)≤(gy1,gb1). Therefore, through induction, we can deduce that
(gon,gan)≤(gyn,gbn),(n≥0). |
Hence, according to (3.38), we obtain
ϕ(ω(go,gyn+1))≤ϕ(skω(go,gyn+1))=ϕ(skω(U(o,a),U(yn,bn)))≤ϕ(Gg(o,a,yn,bn))−ψ(Hg(o,a,yn,bn)), | (3.53) |
where
Gg(o,a,yn,bn)=max{ω(gyn,U(yn,bn))[1+ω(go,U(o,a))]1+ω(go,gyn),ω(go,U(o,a))ω(gyn,U(yn,bn))1+ω(go,gyn),ω(go,U(yn,bn))+ω(gyn,U(o,a))2s,ω(go,gyn)}=max{0,0,ω(go,gyn)s,ω(go,gyn)}=ω(go,gyn) |
and
Hg(o,a,yn,bn)=max{ω(gyn,U(yn,bn))[1+ω(go,U(o,a))]1+ω(go,gyn),ω(go,gyn)}=ω(go,gyn). |
Hence, according to (3.38), we obtain
ϕ(ω(go,gyn+1))≤ϕ(ω(go,gyn))−ψ(ω(go,gyn)). | (3.54) |
By an analogous procedure, we find that
ϕ(ω(ga,gbn+1))≤ϕ(ω(ga,gbn))−ψ(ω(ga,gbn)). | (3.55) |
Combining (3.54) and (3.55), we obtain
ϕ(max{ω(go,gyn+1),ω(ga,gbn+1)})≤ϕ(max{ω(go,gyn),ω(ga,gbn)})−ψ(max{ω(go,gyn),ω(ga,gbn)})<ϕ(max{ω(go,gyn),ω(ga,gbn)}). | (3.56) |
Therefore, utilizing the property of ϕ, we deduce the following:
max{ω(go,gyn+1),ω(ga,gbn+1)}<max{ω(go,gyn),ω(ga,gbn)}, |
which demonstrates that max{ω(go,gyn),ω(ga,gbn)} is a decreasing sequence; then,
limn→+∞max{ω(go,gyn),ω(ga,gbn)}=γ,γ≥0. |
By the limiting case in (3.56), we get
ϕ(γ)≤ϕ(γ)−ψ(γ), |
this suggests that if ψ(γ)=0, then it follows that γ=0. As a result,
limn→+∞max{ω(go,gyn),ω(ga,gbn)}=0. |
Consequently, we get
limn→+∞ω(go,gyn)=0andlimn→+∞d(ga,gbn)=0. | (3.57) |
By a similar argument, we acquire
limn→+∞ω(gr,gyn)=0andlimn→+∞ω(gs,gbn)=0. | (3.58) |
Thus, utilizing (3.57) and (3.58), we conclude that go=gr and ga=gs. Given that go=U(o,a) and ga=U(a,o), then the fact that U and g are commutative results in
g(go)=g(U(o,a))=U(go,ga)andg(ga)=g(U(a,o))=U(ga,go). | (3.59) |
Let go=c∗ and ga=e∗; then, (3.59) becomes
g(c∗)=U(c∗,e∗)andg(e∗)=U(e∗,c∗). | (3.60) |
This establishes that U, g have a coupled coincidence point of (c∗,e∗). Consequently, g(c∗)=gr and g(e∗)=gs, indicating that g(c∗)=c∗ and g(e∗)=e∗. Therefore, we can deduce from (3.60) that
c∗=g(c∗)=U(c∗,e∗)ande∗=g(e∗)=U(e∗,c∗). |
Thus, U, g have (c∗,e∗) which is a coupled common fixed point.
Assuming uniqueness, consider another pair (u∗,m∗) as a coupled common fixed point for both U and g. This implies that
u∗=gu∗=U(u∗,m∗)andm∗=gm∗=U(m∗,u∗). |
Given that (u∗,m∗) is a coupled common fixed point of U and g, it follows that gu∗=go=c∗ and, gm∗=ga=e∗. Consequently,
u∗=gu∗=gc∗=c∗andm∗=gm∗=ge∗=e∗. |
Hence, we have the proof.
Theorem 18. Under the additional conditions of Theorem 17, when go0 and ga0 are comparable, a common fixed point which is unique exists in L for U and g.
Proof. Following Theorem 17, the existence of a coupled common fixed point (o,a) which is a unique for U and g in L is established. To prove that o=a, consider the hypothesis that go0 and ga0 are comparable, assuming, without loss of generality, that go0⪯ga0. Through induction, it is inferred that gon⪯gan holds for all n≥0, where {gon} and {gan} are sequences derived from Theorem 14.
Now, employing Lemma 6, we obtain
ϕ(sk−2ω(o,a))=ϕ(sk1s2ω(o,a))≤limn→+∞supϕ(skω(on+1,an+1))=limn→+∞supϕ(skω(U(on,an),U(an,on)))≤limn→+∞supϕ(Gg(on,an,an,on))−limn→+∞infψ(Hg(on,an,an,on))≤ϕ(ω(o,a))−limn→+∞infψ(Hg(on,an,an,on))<ϕ(ω(o,a)). |
This contradiction suggests that o=a. Hence, we have the proof.
Remark 19. The condition outlined is based on the known fact that a b-metric space transforms into a metric space for s=1. Consequently, drawing upon Jachymski's[36] findings, the specified condition unfolds;
ϕ(ω(U(o,a),U(ρ,d)))≤ϕ(max{ω(go,gρ),ω(ga,gd)})−ψ(max{ω(go,gρ),ω(ga,gd)}) |
is equivalent to
ω(U(o,a),U(ρ,d))≤φ(max{ω(go,gρ),ω(ga,gd)}). |
In this scenario, considering that ϕ∈Φ, ψ∈Ψ and a self-map φ over [0,+∞) exists as a continuous function with φ(t)<t,∀t>0 and φ(t)=0 under the circumstance that t=0, it can be concluded that our findings provide a broader and more extensive perspective than the results presented in [13,14,37], along with other similar outcomes.
Corollary 20. A non-decreasing self-map U over a c.m.s. (L,ω,s,⪯) is continuous and o0⪯Uo0 for o0∈L. Then U possesses a fixed point in L if for (o,a)∈L×L,
ϕ(sω(Uo,Ua))≤ϕ(G(o,a))−ψ(G(o,a)), |
where G(o,a), ϕ and ψ are the same as in Theorem 7.
Proof. Choose H(o,a)=G(o,a) in the contraction condition (3.3) of Theorem 7.
Note 1. Additionally, given that the requirements of Theorem 8 are fulfilled by L, it establishes the presence of a fixed point for the mapping U. Also, under the condition (3.15) satisfied by L, the uniqueness of the fixed point is ensured.
Note 2. By adapting the proofs provided in Theorems 11 and 12, one can establish a coincidence point for U and g within L. Similarly, leveraging Theorems 14, 17 and 18, one can derive a coupled coincidence point which is unique, along with a unique common fixed point for U and g within both L×L and L. These results are contingent upon the fulfillment of a generalized contraction condition in Corollary 20, where Gg(o,a), Gg(o,a,ρ,d), and the conditions imposed on ϕ and ψ remain consistent with the previously defined formulations.
Corollary 21. A non-decreasing self-map U over a c.m.s. (L,ω,s,⪯) is continuous. Assume that there is a constant k∈[0,1), and for each pair of elements o,a∈L with o⪯a,
ω(Uo,Ua)≤ksmax{ω(a,Ua)[1+ω(o,Uo)]1+ω(o,a),ω(o,Uo)ω(a,Ua)1+ω(o,a),ω(o,Ua)+ω(a,Uo)2s,ω(o,a)}. |
If there exists o0∈L with o0⪯Uo0, then U has a fixed point in L.
Proof. Define ϕ(t)=t and ψ(t)=(1−k)t, for all t∈(0,+∞) in accordance with Corollary 20.
Note 3. By relaxing the continuity constraints for the mapping U as described in Corollary 21, it is possible to acquire a fixed point of U by considering a decreasing sequence {on}⊂L in Theorem 8.
We demonstrate the practical relevance of the obtained results for diverse scenarios, including situations in which a metric ω exhibits both continuity and discontinuity within a space L.
Example 22. A metric ω: L×L→L is defined as below and ≤ is a usual order on L, where L={1,2,3,4,5,6} given the following:
ω(o,a)=ω(a,o)=0,ifo,a=1,2,3,4,5,6ando=a,ω(o,a)=ω(a,o)=3,ifo,a=1,2,3,4,5ando≠a,ω(o,a)=ω(a,o)=12,ifo=1,2,3,4anda=6,ω(o,a)=ω(a,o)=20,ifo=5anda=6. |
Define a map U: L→L by U1=U2=U3=U4=U5=1,U6=2 and set ϕ(t)=t2, ψ(t)=t4, ∀t∈[0,+∞). Then U possesses a fixed point in L.
Proof. It is clear that when s=2, (L,ω,s,⪯) forms a c.m.s. Let us examine the various scenarios for elements o and a within L.
Case 1. Assume that o and a are elements of the set {1,2,3,4,5}, with o<a. Consequently,
ω(Uo,Ua)=ω(1,1)=0. |
Thus,
ϕ(2ω(Uo,Ua))=0≤ϕ(G(o,a))−ψ(G(o,a)). |
Case 2. Assume that o is any one of {1,2,3,4,5} and a=6. In this scenario, ω(Uo,Ua)=ω(1,2)=3, G(6,5)=20 and G(o,6)=12, for o∈{1,2,3,4}. As a result, we have
ϕ(2ω(Uo,Ua))≤G(o,a)4=ϕ(G(o,a))−ψ(G(o,a)). |
Therefore, the condition established in Corollary 20 is satisfied. Additionally, the other prerequisites outlined in Corollary 20 are met. Consequently, U possesses a fixed point in L since Corollary 20 is applicable for U, ϕ, ψ and the space (L,ω,s,⪯).
Example 23. A metric ω: L×L→L is specified on the set
L={0,1,12,13,14,…,1n,…} |
with the standard order ≤ in the following manner:
ω(o,a)={0,foro=a,1,foro≠a∈{0,1},|o−a|,foro,a∈{0,12n,12m:n≠m≥1},2,otherwise. |
Define U: L→L under the conditions that U0=0,U1n=112n,∀n≥1 and ϕ(t)=t, ψ(t)=4t5,0≤t<∞. Then U possesses a fixed point in L.
Proof. Clearly, when s=125, the space (L,ω,s,⪯) forms a c.m.s. Additionally, as defined, the metric ω is a b-metric space which is discontinuous. When examining o,a∈L where o<a, the following situations arise:
Case 1. If o=0 and a=1n,∀n≥1, then,
ω(Uo,Ua)=ω(0,112n)=112nandG(o,a)=1norG(o,a)={1,2}. |
Therefore, we have
ϕ(125ω(Uo,Ua))≤G(o,a)5=ϕ(G(o,a))−ψ(G(o,a)). |
Case 2. If o=1m and a=1n with m>n≥1, then,
ω(Uo,Ua)=ω(112m,112n)andG(o,a)≥1n−1morG(o,a)=2. |
Therefore,
ϕ(125ω(Uo,Ua))≤G(o,a)5=ϕ(G(o,a))−ψ(G(o,a)). |
Thus, the conditions specified in Corollary 20, including condition the defined condition, along with the fulfillment of the remaining assumptions, are met. Consequently, there exists a fixed point for U within the space L.
Example 24. Consider L such that it consists of all continuous functions on [a,b], denoted by L=C[a,b]. A b-metric ω on L defined as
ω(θ1,θ2)=supt∈[a,b]{|θ1(t)−θ2(t)|2} |
for any θ1,θ2∈L with the partial order ⪯ defined with θ1⪯θ2, if a≤θ1(t)≤θ2(t)≤b, where 0≤a<b and t∈[a,b]. Define U: L→L under the conditions that Uθ=θ5,θ∈L and the distance functions ϕ(t)=t, ψ(t)=t3, 0≤t<∞. Thus, U possesses a unique fixed point in L.
Proof. Given the specified assumptions, it is evident that (L,ω,s,⪯) forms a c.m.s. with s=2, satisfying the conditions stated in Corollary 20 and Note 1. Moreover, for all θ1,θ2∈L, the function
min(θ1,θ2)(t)=min{θ1(t),θ2(t)} |
is continuous, satisfying the assumptions of Corollary 20 and Note 1. Thus in L, U possesses a unique fixed point θ=0.
We examine the presence of a unique solution to a nonlinear quadratic integral [15], considering its applicability in light of the coupled fixed point defined in Corollary 16 outlined in this paper.
Let us examine the following integral equation of a quadratic and nonlinear nature:
x(t)=δ(t)+α∫10k1(t,q)g1(q,x(q))dq∫10k2(t,q)g2(q,x(q))dq,t∈I=[0,1],α≥0. | (4.1) |
Denote Γ that consists of self-maps γ over [0,+∞) to satisfy the following conditions:
(i) (γ(t))q≤γ(tq),∀q≥1 and γ is non-decreasing.
(ii) ∃ ϕ∈Φ with γ(t)=t−ϕ(t),∀t∈[0,∞).
For example, γ1(t)=rt, 0≤r<1 and γ2(t)=tt+1 are in Γ [15].
We will examine the equation given by (4.1) as subject to the following:
(c1) The functions gi: I×R→R, with gi(t,o)≥0 are continuous and ξi∈L1(I) are two functions such that gi(t,o)≤ξi(t) for i=1,2.
(c2) g1(t,o) exhibits monotonic non-decreasing behavior in o, while g2(t,a) demonstrates monotonic descending characteristics in a for every o,a∈R and t∈I.
(c3) The function δ: I→R is continuous.
(c4) The functions denoted by ki: I×I→R, i=1,2 exhibit continuity in I at every t for each q∈I and measurability in q∈I for each t∈I under the given conditions
∫10ki(t,q)ξi(q)dq≤K,i=1,2andki(t,o)≥0. |
(c5) For any o,a∈R with o≥a a function γ∈Γ and the constants 0≤Li<1,(i=1,2) exist such that
|gi(t,o)−gi(t,a)|≤Liγ(o−a),(i=1,2). |
(c6) There are functions l1,l2∈C(I) with
l1(t)≤δ(t)+α∫10k1(t,q)g1(q,l1(q))dq∫10k2(t,q)g2(q,l2(q))dq≤δ(t)+λ∫10k1(t,q)g1(q,l2(q))dq∫10k2(t,q)g2(q,l1(q))dq≤l2(t). |
(c7) max{L1q,L2q}αqK2q≤124q−3.
Let us consider the space L=C(I) with I=[0,1], which denotes the set of continuous maps equipped with a metric.
d(o,a)=supt∈I|o(t)−a(t)|foro,a∈C(I). |
It is evident that the space can be endowed with a partial order defined by
o,a∈C(I),o≤a⟺o(t)≤a(t),∀t∈I. |
For q≥1, a metric ω defined as below,
ω(o,a)=(d(o,a)q)=(supt∈I|o(t)−a(t)|)q=supt∈I|o(t)−a(t)|qforo,a∈C(I), |
forms a complete b-metric space for s=2q−1 according to [16].
Additionally, L×L=C(I)×C(I) forms a p.o.s with the specified order relation, as follows
(o,a),(ρ,d)∈L×L,(o,a)≤(ρ,d)⟺o≤ρanda≥d. |
Moreover, for all o,a∈L and any t∈I, max{o(t),a(t)} serves as both an upper and lower bound for o,a in L. Consequently, for any
(o,a),(ρ,d)∈L×L,(max{o,ρ},min{a,d})∈L×L |
is comparable to both (o,a) and (ρ,d).
Theorem 25. Given the conditions (c1)–(c7), the integral Eq (4.1) admits a unique solution within the set C(I).
Proof. Define a mapping U: L×L→L as follows:
U(o,a)(t)=δ(t)+α∫10k1(t,q)g1(q,o(q))dq∫10k2(t,q)g2(q,a(q))dqfort∈I. |
Therefore, the well-definiteness of U is ensured by the given hypotheses. To demonstrate the mixed monotone property of U, consider the case in which o1≤o2 and t∈I
U(o1,a)(t)−U(o2,a)(t)=δ(t)+α∫10k1(t,q)g1(q,o1(q))dq∫10k2(t,q)g2(q,a(q))dq−δ(t)−α∫10k1(t,q)g1(q,o2(q))dq∫10k2(t,q)g2(q,a(q))dq=α∫10k1(t,q)[g1(q,o1(q))−g1(q,o2(q))]dq∫10k2(t,q)g2(q,a(q))ds≤0. |
Similarly, we establish that
U(o,a1)(t)≤U(o,a2)(t) |
holds true if a1≤a2 and t∈I. Thus, U possesses the mixed monotone property. Furthermore, for (o,a)≤(ρ,d), which implies that o≤ρ and a≥d, we obtain
d|U(o,a)(t)−U(ρ,d)(t)|≤|α∫10k1(t,q)g1(q,o(q))dq∫10k2(t,q)[g2(q,a(q))−g2(q,d(q))]dq+α∫10k2(t,q)g2(q,d(q))dq∫10k1(t,q)[g1(q,o(q))−g1(q,ρ(q))]dq|≤α∫10k1(t,q)g1(q,o(q))dq∫10k2(t,q)|g2(q,a(q))−g2(q,d(q))|dq+α∫10k2(t,q)g2(q,d(q))dq∫10k1(t,q)|g1(q,o(q))−g1(q,ρ(q))|dq≤α∫10k1(t,q)l1(q)dq∫10Λ2(t,q)L2γ[a(q)−d(q)]dq+α∫10k2(t,q)l2(q)dq∫10k1(t,q)L1γ[ρ(q)−o(q)]dq. |
Given that γ is an increasing function and under the conditions o≤ρ, a≥d, we can conclude that
γ(ρ(q)−o(q))≤γ(supt∈I|o(q)−ρ(q)|)=γ(d(o,ρ)) |
and
γ(a(q)−d(q))≤γ(supt∈I|a(q)−d(q)|)=γ(d(a,d)). |
Thus,
|U(o,a(t))−U(ρ,d)(t)|≤αK∫10k2(t,q)L2γ(d(a,d))dq+αK∫10k1(t,q)L1γ(d(ρ,o))dq≤αK2max{L1,L2}[γ(d(ρ,o))+γ(d(a,d))]. |
Therefore,
ω(U(o,a),U(ρ,d))=supt∈I|U(o,a)(t)−U(ρ,d)(t)|q≤{αK2max{L1,L2}[γ(d(ρ,o))+γ(d(a,d))]}q=αqK2qmax{L1q,L2q}[γ(d(ρ,o))+γ(d(a,d))]q, |
and for q>1, we have that (m+n)q≤2q−1(mq+nq) for m,n∈(0,∞); then,
ω(U(o,a),U(ρ,d))≤2q−1αqK2qmax{L1q,L2q}[(γ(d(ρ,o)))q+(γ(d(a,d)))q]≤2q−1αqK2qmax{L1q,L2q}[γ(d(ρ,o))+γ(d(a,d))]≤2qαqK2qmax{L1q,L2q}[γGg(o,a,ρ,d)]≤2qαqK2qmax{L1q,L2q}[Gg(o,a,ρ,d)−ψ(Hg(o,a,ρ,d))]≤123q−kGg(o,a,ρ,d)−123q−kψ(Hg(o,a,ρ,d)), |
this implies that the mapping U fulfills the contraction condition (3.50) specified in Corollary 16.
Lastly, consider the functions l1,l2 as indicated in assumption (c6). By (c6), we obtain
l1≤U(l1,l2)≤U(l2,l1)≤l2. |
Therefore, based on Theorem 17, the mapping U possesses a unique coupled fixed point (c∗,e∗)∈L×L. Given l1≤l2, according to Theorem 18, we conclude that c∗=e∗, suggesting that c∗=U(c∗,c∗). Consequently, c∗∈C(I) stands as the unique solution to (4.1).
This study has established results related to the fixed point of a mapping that adheres to the conditions of a rational contraction with auxiliary functions, demonstrating its uniqueness. Furthermore, we extend these findings to obtain coupled fixed points for two mappings within an ordered metric space. The results presented here generalize and extend existing comparable findings in the literature. Numerical examples have been provided to illustrate and support the outcomes. Moreover, these results have the potential for further generalization and extension in various ordered metric spaces, given the necessary topological conditions.
Additionally, the aforementioned findings can be extended and generalized by incorporating the idea of wt-distance in a metric type space [47], cone b-metric spaces over Banach algebras [48] and (ϕ,ψ)-weak contractions as presented in [49].
The authors declare they have not used artificial intelligence (AI) tools in the creation of this article. The authors declare that they have only used the Grammarly free AI tools to improve the manuscript's grammar.
The authors A. Aloqaily and N. Mlaiki would like to thank Prince Sultan University for paying the APC and for the support through the TAS research lab.
The authors declare that they have no conflicts of interest.
[1] |
D. Konstantinides, Q. Tang, G. Tsitsiashvili, Estimates for the ruin probability in the classical risk model with constant interest force in the presence of heavy tails, Insur. Math. Econ., 31 (2002), 447–460. https://doi.org/10.1016/S0167-6687(02)00189-0 doi: 10.1016/S0167-6687(02)00189-0
![]() |
[2] |
H. U. Gerber, E. S. W. Shiu, On optimal dividend strategies in the compound Poisson model, N. Am. Actuar. J., 10 (2006), 76–93. http://doi.org/10.1080/10920277.2006.10596249 doi: 10.1080/10920277.2006.10596249
![]() |
[3] |
B. Avanzi, H. U. Gerber, E. S. W. Shiu, Optimal dividends in the dual model, Insur. Math. Econ., 41 (2007), 111–123. http://doi.org/10.1016/j.insmatheco.2006.10.002 doi: 10.1016/j.insmatheco.2006.10.002
![]() |
[4] |
H. U. Gerber, N. Smith, Optimal dividends with incomplete information in the dual model, Insur. Math. Econ., 43 (2008), 227–233. https://doi.org/10.1016/j.insmatheco.2008.06.002 doi: 10.1016/j.insmatheco.2008.06.002
![]() |
[5] |
A. C. Y. Ng, On a dual model with a dividend threshold, Insur. Math. Econ., 44 (2009), 315–324. https://doi.org/10.1016/j.insmatheco.2008.11.011 doi: 10.1016/j.insmatheco.2008.11.011
![]() |
[6] |
D. J. Yao, H. L. Yang, R. M. Wang, Optimal dividend and capital injection problem in the dual model with proportional and fixed transaction costs, Eur. J. Oper. Res., 211 (2011), 568–576. https://doi.org/10.1016/j.ejor.2011.01.015 doi: 10.1016/j.ejor.2011.01.015
![]() |
[7] |
Y. X. Zhao, R. M. Wang, D. J. Yao, P. Chen, Optimal dividends and capital injections in the dual model with a random time horizon, J. Optimiz. Theory Appl., 167 (2015), 272–295. https://doi.org/10.1007/s10957-014-0653-0 doi: 10.1007/s10957-014-0653-0
![]() |
[8] |
C. Yang, K. P. Sendova, Z. Li, Parisian ruin with a threshold dividend strategy under the dual Lévy risk model, Insur. Math. Econ., 90 (2020), 135–150. https://doi.org/10.1016/j.insmatheco.2019.11.002 doi: 10.1016/j.insmatheco.2019.11.002
![]() |
[9] |
A. Fahim, L. J. Zhu, Asymptotic analysis for optimal dividends in a dual risk model, Stoch. Models, 38 (2022), 605–637. https://doi.org/10.1080/15326349.2022.2080709 doi: 10.1080/15326349.2022.2080709
![]() |
[10] |
Z. J. Song, F. Y. Sun, The dual risk model under a mixed ratcheting and periodic dividend strategy, Commun. Stat. Theor. M., 52 (2023), 3526–3540. https://doi.org/10.1080/03610926.2021.1974483 doi: 10.1080/03610926.2021.1974483
![]() |
[11] |
Z. Liu, P. Chen, Y. J. Hu, On the dual risk model with diffusion under a mixed dividend strategy, Appl. Math. Comput., 376 (2020), 125115. https://doi.org/10.1016/j.amc.2020.125115 doi: 10.1016/j.amc.2020.125115
![]() |
[12] |
J. L. Pérez, K. Yamazaki, Optimality of hybrid continuous and periodic barrier strategies in the dual model, Appl. Math. Optim., 82 (2020), 105–133. https://doi.org/10.1007/s00245-018-9494-9 doi: 10.1007/s00245-018-9494-9
![]() |
[13] |
E. Bayraktar, A. E. Kyprianou, K. Yamazaki, Optimal dividends in the dual model under transaction costs, Insur. Math. Econ., 54 (2014), 133–143. https://doi.org/10.1016/j.insmatheco.2013.11.007 doi: 10.1016/j.insmatheco.2013.11.007
![]() |
[14] |
Y. X. Zhao, P. Chen, H. L. Yang, Optimal periodic dividend and capital injection problem for spectrally positive Lévy processes, Insur. Math. Econ., 74 (2017), 135–146. https://doi.org/10.1016/j.insmatheco.2017.03.006 doi: 10.1016/j.insmatheco.2017.03.006
![]() |
[15] |
D. C. M. Dickson, A. D. E. Reis, H. R. Waters, Some stable algorithms in ruin theory and their applications, Astin Bull., 25 (1995), 153–175. https://doi.org/10.2143/AST.25.2.563245 doi: 10.2143/AST.25.2.563245
![]() |
[16] |
H. Cossette, D. Landriault, É. Marceau, Compound binomial risk model in a Markovian environment, Insur. Math. Econ., 35 (2004), 425–443. https://doi.org/10.1016/j.insmatheco.2004.07.009 doi: 10.1016/j.insmatheco.2004.07.009
![]() |
[17] |
H. U. Gerber, Mathematical fun with the compound binomial process, Astin Bull., 18 (1988), 161–168. https://doi.org/10.2143/AST.18.2.2014949 doi: 10.2143/AST.18.2.2014949
![]() |
[18] |
Z. H. Bao, A note on the compound binomial model with randomized dividend strategy, Appl. Math. Comput., 194 (2007), 276–286. https://doi.org/10.1016/j.amc.2007.04.023 doi: 10.1016/j.amc.2007.04.023
![]() |
[19] |
D. Landriault, Randomized dividends in the compound binomial model with a general premium rate, Scand. Actuar. J., 2008 (2008), 1–15. https://doi.org/10.1080/03461230701642489 doi: 10.1080/03461230701642489
![]() |
[20] |
X. X. Yang, J. Y. Tan, H. J. Zhang, Z. Q. Li, An optimal control problem in a risk model with stochastic premiums and periodic dividend payments, Asia Pac. J. Oper. Res., 34 (2017), 1740013. https://doi.org/10.1142/S0217595917400139 doi: 10.1142/S0217595917400139
![]() |
[21] |
S. Drekic, J. K. Woo, R. Xu, A threshold-based risk process with a waiting period to pay dividends, J. Ind. Manag. Optim., 14 (2018), 1179–1201. https://doi.org/10.3934/jimo.2018005 doi: 10.3934/jimo.2018005
![]() |
[22] |
J. Y. Tan, S. L. Yuan, A dividend optimization problem with constraint of survival probability in a Markovian environment model, Commun. Stat. Theor. M., 50 (2021), 3522–3546. https://doi.org/10.1080/03610926.2019.1705981 doi: 10.1080/03610926.2019.1705981
![]() |
[23] |
X. Lin, Compound binomial risk model in a Markovian environment with capital cost and the calculation algorithm, Appl. Math. Comput., 424 (2022), 126969. https://doi.org/10.1016/j.amc.2022.126969 doi: 10.1016/j.amc.2022.126969
![]() |
[24] |
A. S. Dibu, M. J. Jacob, On a double barrier hybrid dividend strategy in a compound Poisson risk model with stochastic income, Ann. Oper. Res., 315 (2022), 969–984. https://doi.org/10.1007/s10479-021-03937-0 doi: 10.1007/s10479-021-03937-0
![]() |
[25] |
R. E. Greenblatt, A dual theory of price and value in a meso-scale economic model with stochastic profit rate, Physica A, 416 (2014), 518–531. https://doi.org/10.1016/j.physa.2014.08.061 doi: 10.1016/j.physa.2014.08.061
![]() |
[26] |
R. J. Boucherie, O. J. Boxma, K. Sigman, A note on negative customers, GI/G/1 workload, and risk processes, Probab. Eng. Inform. Sc., 11 (1997), 305–311. https://doi.org/10.1017/S0269964800004848 doi: 10.1017/S0269964800004848
![]() |
[27] |
K. P. Sendova, C. Yang, R. X. Zhang, Dividend barrier strategy: Proceed with caution, Stat. Probabil. Lett., 137 (2018), 157–164. https://doi.org/10.1016/j.spl.2018.01.016 doi: 10.1016/j.spl.2018.01.016
![]() |
[28] |
E. C. K. Cheung, J. T. Y. Wong, On the dual risk model with Parisian implementation delays in dividend payments, Eur. J. Oper. Res., 257 (2017), 159–173. https://doi.org/10.1016/j.ejor.2016.09.018 doi: 10.1016/j.ejor.2016.09.018
![]() |
[29] |
K. Hu, J. C. Li, J. M. Zhou, On the dual risk model with Parisian implementation delays under a mixed dividend strategy, Probab. Eng. Inform. Sc., 37 (2023), 442–461. https://doi.org/10.1017/S0269964822000481 doi: 10.1017/S0269964822000481
![]() |
[30] |
D. J. Yao, H. L. Yang, R. M. Wang, Optimal financing and dividend strategies in a dual model with proportional costs, J. Ind. Manag. Optim., 6 (2010), 761–777. https://doi.org/10.3934/jimo.2010.6.761 doi: 10.3934/jimo.2010.6.761
![]() |
[31] |
J. Y. Tan, X. Q. Yang, Optimal dividend strategy in compound binomial model with bounded dividend rates, Acta Math. Appl. Sin. Engl. Ser., 30 (2014), 859–870. https://doi.org/10.1007/s10255-014-0428-2 doi: 10.1007/s10255-014-0428-2
![]() |
[32] |
A. Bazyari, On the evaluation of ruin probabilities in a generalized dual binomial risk model using Markov property, Commun. Stat. Theor. M., 53 (2024), 1162–1187. https://doi.org/10.1080/03610926.2022.2093910 doi: 10.1080/03610926.2022.2093910
![]() |
[33] |
S. Asmussen, Risk theory in a Markovian environment, Scand. Actuar. J., 1989 (1989), 69–100. https://doi.org/10.1080/03461238.1989.10413858 doi: 10.1080/03461238.1989.10413858
![]() |
[34] |
E. Marciniak, Z. Palmowski, On the optimal dividend problem in the dual model with surplus-dependent premiums, J. Optim. Theory Appl., 179 (2018), 533–552. https://doi.org/10.1007/s10957-016-1050-7 doi: 10.1007/s10957-016-1050-7
![]() |
[35] |
R. Xu, J. K. Woo, Optimal dividend and capital injection strategy with a penalty payment at ruin: Restricted dividend payments, Insu. Math. Econ., 92 (2020), 1–16. https://doi.org/10.1016/j.insmatheco.2020.02.008 doi: 10.1016/j.insmatheco.2020.02.008
![]() |
[36] |
J. Y. Tan, Y. Yang, S. R. Liu, K. N. Xiang A consistent estimation of optimal dividend strategy in a risk model with delayed claims, Commun. Stat.-Simul. C., 51 (2022), 6840–6853. https://doi.org/10.1080/03610918.2020.1818096 doi: 10.1080/03610918.2020.1818096
![]() |
[37] |
R. G. Alcoforado, A. I. Bergel, R. M. R. Cardoso, A. D. E. Reis, E. V. Rodríguez-Martínez, Ruin and dividend measures in the renewal dual risk model, Methodol. Comput. Appl. Probab., 24 (2022), 537–569. https://doi.org/10.1007/s11009-021-09876-4 doi: 10.1007/s11009-021-09876-4
![]() |
[38] |
R. Wang, J. Q. Huang, L. Z. Zhang, Y. Xia, X. Xu, T. L. Nong, Assessments of air pollution control effectiveness based on a sharp regression discontinuity design—evidence from China's Environmental Big Data, Front. Environ. Sci., 9 (2021), 724716. https://doi.org/10.3389/fenvs.2021.724716 doi: 10.3389/fenvs.2021.724716
![]() |
1. | Mahmoud S. Mehany, Faizah D. Alanazi, An η-Hermitian solution to a two-sided matrix equation and a system of matrix equations over the skew-field of quaternions, 2025, 10, 2473-6988, 7684, 10.3934/math.2025352 |