Research article

Predicting future order statistics with random sample size

  • We suggest a new method for constructing an efficient point predictor for the future order statistics when the sample size is a random variable. The suggested point predictor is based on some characterization properties of the distributions of order statistics. For several distributions, including the mixture distribution, the performance of the suggested predictor is evaluated by means of a comprehensive simulation study. Three examples of real lifetime data-sets are analyzed by using this method and compared with an efficient recent method given by Barakat et al. [1], that deals with non-random sample sizes. One of these examples predicts the accumulative new cases per million for infection of the new Coronavirus (COVID-19).

    Citation: Haroon Barakat, Osama Khaled, Hadeer Ghonem. Predicting future order statistics with random sample size[J]. AIMS Mathematics, 2021, 6(5): 5133-5147. doi: 10.3934/math.2021304

    Related Papers:

    [1] Mohammed Shehu Shagari, Akbar Azam . Integral type contractions of soft set-valued maps with application to neutral differential equations. AIMS Mathematics, 2020, 5(1): 342-358. doi: 10.3934/math.2020023
    [2] 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
    [3] Sahibzada Waseem Ahmad, Muhammad Sarwar, Thabet Abdeljawad, Gul Rahmat . Multi-valued versions of Nadler, Banach, Branciari and Reich fixed point theorems in double controlled metric type spaces with applications. AIMS Mathematics, 2021, 6(1): 477-499. doi: 10.3934/math.2021029
    [4] Amjad Ali, Muhammad Arshad, Eskandar Emeer, Hassen Aydi, Aiman Mukheimer, Kamal Abodayeh . Certain dynamic iterative scheme families and multi-valued fixed point results. AIMS Mathematics, 2022, 7(7): 12177-12202. doi: 10.3934/math.2022677
    [5] Aftab Hussain . Fractional convex type contraction with solution of fractional differential equation. AIMS Mathematics, 2020, 5(5): 5364-5380. doi: 10.3934/math.2020344
    [6] Mesfer H. Alqahtani, Alaa M. Abd El-latif . Separation axioms via novel operators in the frame of topological spaces and applications. AIMS Mathematics, 2024, 9(6): 14213-14227. doi: 10.3934/math.2024690
    [7] Amer Hassan Albargi . Some new results in F-metric spaces with applications. AIMS Mathematics, 2023, 8(5): 10420-10434. doi: 10.3934/math.2023528
    [8] 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
    [9] Amer Hassan Albargi, Jamshaid Ahmad . Fixed point results of fuzzy mappings with applications. AIMS Mathematics, 2023, 8(5): 11572-11588. doi: 10.3934/math.2023586
    [10] Abdullah Shoaib, Tahair Rasham, Giuseppe Marino, Jung Rye Lee, Choonkil Park . Fixed point results for dominated mappings in rectangular b-metric spaces with applications. AIMS Mathematics, 2020, 5(5): 5221-5229. doi: 10.3934/math.2020335
  • We suggest a new method for constructing an efficient point predictor for the future order statistics when the sample size is a random variable. The suggested point predictor is based on some characterization properties of the distributions of order statistics. For several distributions, including the mixture distribution, the performance of the suggested predictor is evaluated by means of a comprehensive simulation study. Three examples of real lifetime data-sets are analyzed by using this method and compared with an efficient recent method given by Barakat et al. [1], that deals with non-random sample sizes. One of these examples predicts the accumulative new cases per million for infection of the new Coronavirus (COVID-19).



    In the past few decades, fixed point theory was developed by a large number of authors, especially in metric spaces, which can be observed in [1,2,3,4,5,6]. In 1993, Czerwik [7] initiated the concept of b-metric spaces. Later, many authors proved fixed point theorems in b-metric spaces [8,9,10]. However, the general metric notion was introduced by Branciari [11] in 2000, the so-called Branciari metric. The notion of generalization of Branciari b-metric spaces was introduced by George et al. [12] in 2015. Johnsonbaugh [13] explored certain fundamental mathematical principles, including foundational topics relevant to fixed-point theory and discrete structures, which underpin many concepts in fixed-point applications. Younis et al. [14] introduced graphical rectangular b-metric space and proved fixed point theorem. Younis et al. [15] presented graphical b-metric space and proved fixed point theorem. Younis et al. [16] presented graphical extended b-metric space and proved fixed point theorem. Younis et al. [17] proved fixed points results using graphical B c-Kannan-contractions by numerical iterations within the structure of graphical extended b-metric spaces. Younis et al. [18] presented a fixed point result for Kannan type mappings, in the framework of graphical b-metric spaces. Younis et al. [19] introduced the notion of controlled graphical metric type spaces and proved the fixed point theorem. Haroon Ahmad et al. [20] developed the graphical bipolar b-metric space and proved the fixed point theorem.

    The following preliminary is given for better understanding by the readers.

    Let (Υ,ϱ) be a metric space. Let Δ denote the diagonal of the Cartesian product Υ×Υ. Consider a directed graph Ω such that the set V(Ω) of its vertices coincides with Υ, and the set E(Ω) of its edges contains all loops, i.e., E(Ω)Δ. We assume Ω has no parallel edges, so we can identify Ω with the pair (V(Ω),E(Ω)). Moreover, we may treat Ω as a weighted graph (see [13], p.309) by assigning to each edge the distance between its vertices. By Ω1, we denote the conversion of a graph Ω, i.e., the graph obtained from Ω by reversing the direction of edges. Thus, we have

    E(Ω1)={(ϑ,σ)|(σ,ϑ)Ω}.

    The letter ˜Ω denotes the undirected graph obtained from Ω by ignoring the direction of edges. Actually, it will be more convenient for us to treat ˜Ω as a directed graph for which the set of its edges is symmetric. Under this convention,

    E(˜Ω)=E(Ω)E(Ω1). (1.1)

    We call (V,E) a subgraph of Ω if VV(Ω),EE(Ω) and, for any edge (ϑ,σ)E,ϑ,σV.

    If ϑ and σ are vertices in a graph Ω, then a path in Ω from ϑ to σ of length r(rN) is a sequence (ϑi)ri=0 of r+1 vertices such that ϑ0=ϑ,ϑr=σ and (ϑζ1,ϑζ)E(Ω) for i=1,.....,r. A graph Ω is connected if there is a path between any two vertices. Ω is weakly connected if, treating all of its edges as being undirected, there is a path from every vertex to every other vertex. More precisely, Ω is weakly connected if ˜Ω is connected.

    We define a relation P on Υ by: (ϑPσ)Ω if and only if there is a directed path from ϑ to σ in Ω. We write η,κ(ϑPσ)Ω if η,κ is contained in some directed path from ϑ to σ in Ω. For lN, we denote

    [ϑ]lΩ={σΥ:there is a directed path from ϑtoσof lengthl}.

    A sequence {ϑζ} in Υ is said to be Ω-term wise connected if (ϑζPσζ) for all ζN. Further details one can see [21,22,23,24,25].

    Definition 1.1. Let Υ be a nonempty set endowed with a graph Ω, 1 and ϱ:Υ×Υ[0,+) satisfy the assumptions below for every ϑ,σΥ:

    (T1) ϱ(ϑ,σ)=0 if and only if ϑ=σ;

    (T2) ϱ(ϑ,σ)=ϱ(σ,ϑ);

    (T3) (ϑPσ)Ω, η,φ(ϑPσ)Ω implies ϱ(ϑ,σ)[ϱ(ϑ,φ)+ϱ(φ,ω)+ϱ(ω,σ)] for all distinct points φ,ωΥ/{ϑ,σ}.

    In this case, the pair (Υ,ϱ) is called a graphical Branciari -metric space with constant 1.

    Example 1.1. Let Υ=BU, where B={0,12,13,14} and U=[1,2]. Define the graphical Branciari -metric space ϱ:Υ×Υ[0,+) as follows:

    {ϱ(ϑ,σ)=ϱ(σ,ϑ)for allϑ,σΥ,ϱ(ϑ,σ)=0ϑ=σ.

    and

    {ϱ(0,12)=ϱ(12,13)=0.2,ϱ(0,13)=ϱ(13,14)=0.02,ϱ(0,14)=ϱ(12,14)=0.5,ϱ(ϑ,σ)=|ϑσ|2,otherwise.

    equipped with the graph Ω=(V(Ω),E(Ω)) so that Υ=V(Ω) with E(Ω)).

    It can be seen that the above Figure 1 depicts the graph given by Ω=(V(Ω),E(Ω)).

    Figure 1.  Graphical Branciari -metric space.

    Definition 1.2. Let {ϑζ} be a sequence in a graphical Branciari -metric space (Υ,ϱ). Then,

    (S1) {ϑζ} converges to ϑΥ if, given ϵ>0, there is ζ0N so that ϱ(ϑζ,ϑ)<ϵ for each ζ>ζ0. That is, limζϱ(ϑζ,ϑ)=0.

    (S2) {ϑζ} is a Cauchy sequence if, for ϵ>0, there is ζ0N so that ϱ(ϑζ,ϑm)<ϵ for all ζ,m>ζ0. That is, limζ,mϱ(ϑζ,ϑm)=0.

    (S3) (Υ,ϱ) is complete if every Cauchy sequence in Υ is convergent in Υ.

    Definition 1.3. (see [8]) A function Q:(0,+)R belongs to F if it satisfies the following condition:

    (F1) Q is strictly increasing;

    (F2) There exists k(0,1) such that limϑ0+ϑkQ(ϑ)=0.

    In [8], the authors omitted Wardowski's (F2) condition from the above definition. Explicitly, (F2) is not required, if {αζ}ζN is a sequence of positive real numbers, then limζ+αζ=0 if and only if limζ+Q(αζ)=. The reason for this is the following lemma.

    Lemma 1.1. If Q:(0,+)R is an increasing function and {αζ}ζN(0,+) is a decreasing sequence such that limζ+Q(αζ)=, then limζ+αζ=0.

    We can also see some properties concerning Q, and Q,.

    Definition 1.4. (see [9]) Let 1 and >0. We say that QF belongs to F, if it also satisfies (Q) if infQ= and ϑ,σ(0,) are such that +Q(ϑ)Q(σ) and +Q(σ)Q(η), then

    +Q(2ϑ)Q(σ).

    In [10], the authors introduced the following condition (F4).

    (Q) if {αζ}ζN(0,+) is a sequence such that +Q(αζ)Q(αn1), for all ζN and for some 0, then +Q(ζαζ)Q(n1αn1), for all ζN.

    Proposition 1.1. (see [8]) If Q is increasing, then (F) is equivalent to (F).

    Definition 4.5. Let (Υ,ϱ) be a graphical Branciari -metric space. We say that a mapping Π:ΥΥ is a Ω-Q-contraction if

    (A1) Π preserves edges of Ω, that is, (ϑ,σ)E(Ω) implies (Πϑ,Πσ)E(Ω);

    (A2) There exists >0 and QF,, such that

    ϑ,σΥ,(ϑ,σ)E(Ω),ϱ(Πϑ,Πσ)>0+Q(ϱ(Πϑ,Πσ))Q(ϱ(ϑ,σ)).

    Chen, Huang, Li, and Zhao [24], proved fixed point theorems for Q-contractions in complete Branciari b-metric spaces. The aim of this paper is to study the existence of fixed point theorems for Q-contractions in complete Branciari b-metric spaces endowed with a graph Ω by introducing the concept of Ω-Q-contraction.

    Theorem 2.1. Let (Υ,ϱ) be a complete graphical Branciari -metric space and QF,. Let Π:ΥΥ be a self mapping such that

    (C1) there exists ϑ0Υ such that Πϑ0[ϑ0]lΩ, for some lN;

    (C2) Π is a Ω-Q-contraction.

    Then Π has a unique fixed point.

    Proof. Let ϑ0Υ be such that Πϑ0[ϑ0]lΩ, for some lN, and {ϑζ} be the Π-Picard sequence with initial value ϑ0. Then, there exists a path {σi}li=0 such that ϑ0=σ0, Πϑ0=σl and (σi1,σi)E(Ω) for i=1,2,3....l. Since Π is a Ω-Q-contraction, by (A1), (Πσi1,Πσi)E(Ω) for i=1,2,3...l. Therefore, {Πσi}li=0 is a path from Πσ0=Πϑ0=ϑ1 to Πσ1=ρ2ϑ0=ϑ2 of length l, and so ϑ2[ϑ1]lΩ. Continuing this process, we obtain that  Πζσi}li=0, is a path from Πζσ0=Πζϑ0=ϑζ to Πζσl=ΠζΠϑ0=ϑζ+1 of length l, and so, ϑζ+1[ϑζ]lΩ, for all ζN. Thus {ϑζ} is a Ω-term wise connected sequence. For any ϑ0Υ, set ϑζ=Πϑζ1,γζ=ϱ(ϑζ+1,ϑζ), and βζ=ϱ(ϑζ+2,ϑζ) with γ0=ϱ(ϑ1,ϑ0) and β0=ϱ(ϑ2,ϑ0). Now, we consider the following two cases:

    (E1) If there exists ζ0N{0} such that ϑζ0=ϑζ0+1, then we have Πϑζ0=ϑζ0. It is clear that ϑζ0 is a fixed point of Π. Therefore, the proof is finished.

    (E2) If ϑζϑζ+1, for any ζN{0}, then we have γζ>0, for each ζN.

    +Q(ϱ(Πϑζ,Πϑζ+1))Q(ϱ(ϑζ,ϑζ+1)),Q(γζ+1)Q(γζ), for every ζN.

    By proposition 1.1, we obtain

    Q(ζ+1γζ+1)Q(ζγζ), ζN. (2.1)

    Furthermore, for any ζN, we have

    Q(ζγζ)Q(ζ1γζ1)Q(ζ2γζ2)2Q(γ0)ζ. (2.2)

    Since limζ(Q(γ0)ζ)=, then

    limζQ(ζγζ)=.

    From (2.1) and ((F1)), we derive that the sequence {ζγζ}ζ=1 is decreasing. By Lemma1.1, we derive that

    limζ(ζγζ)=0.

    By (F2), there exists k(0,1) such that

    limζ(ζγζ)kQ(ζγζ)=0.

    Multiplying (2.2) by (ζγζ)k results

    0ζ(ζγζ)k+(ζγζ)kQ(ζγζ)(ζγζ)kQ(γ0), ζN,

    which implies limζζ(ζγζ)k=0. Then, there exists ζ1N such that ζ(ζγζ)k1, ζζ1. Thus,

    ζγζ1ζ1k, ζζ1. (2.3)

    Therefore, the series i=1iγi is convergent. For all ζ,ωN, we drive the proof into two cases.

    (a) If ω>2 is odd, we obtain

    ϱ(ϑζ+3,ϑζ)ϱ(ϑζ+3,ϑζ+2)+ϱ(ϑζ+2,ϑζ+1)+ϱ(ϑζ+1,ϑζ),
    ϱ(ϑζ+5,ϑζ)ϱ(ϑζ+5,ϑζ+2)+ϱ(ϑζ+2,ϑζ+1)+ϱ(ϑζ+1,ϑζ)2ϱ(ϑζ+5,ϑζ+4)+2ϱ(ϑζ+4,ϑζ+3)+2ϱ(ϑζ+3,ϑζ+2)+γζ+1+γζ.

    Consequently,

    ϱ(ϑζ+ω,ϑζ)ϱ(ϑζ+ω,ϑζ+2)+ϱ(ϑζ+2,ϑζ+1)+ϱ(ϑζ+1,ϑζ)2ϱ(ϑζ+ω,ϑζ+4)+2ϱ(ϑζ+4,ϑζ+3)+2ϱ(ϑζ+3,ϑζ+2)+γζ+1+γζ3ϱ(ϑζ+ω,ϑζ+6)+3γζ+5+3γζ+4+2γζ+3+2γζ+2+γζ+1+γζω12γn+p1+ω12γζ+ω2+ω12γζ+ω3+ω32γζ+ω4+ω32γζ+ω5++2γζ+2+γζ+1+γζω+12γζ+ω1+ω2γζ+ω2+ω12γζ+ω3+ω22γζ+ω4+ω32γζ+ω5++32γζ+1+22γζω+1γζ+ω1+ωγζ+ω2+ω1γζ+ω3+ω2γζ+ω4+ω3γζ+ω5++3γζ+1+2γζ1ζ2(ζ+ω1γζ+ω1+ζ+ω2γζ+ω2+ζ+ω3γζ+ω3++ζ+1γζ+1+ζγζ)=1ζ2ζ+ω1i=ζiγi1ζ2i=ζiγi.

    (b) If ω>2 is even, we can obtain

    ϱ(ϑζ+4,ϑζ)ϱ(ϑζ+4,ϑζ+2)+ϱ(ϑζ+2,ϑζ+1)+ϱ(ϑζ+1,ϑζ).

    Furthermore, we conclude that

    ϱ(ϑζ+ω,ϑζ)ϱ(ϑζ+ω,ϑζ+2)+ϱ(ϑζ+2,ϑζ+1)+ϱ(ϑζ+1,ϑζ)2ϱ(ϑζ+ω,ϑζ+4)+2ϱ(ϑζ+4,ϑζ+3)+2ϱ(ϑζ+3,ϑζ+2)3ϱ(ϑζ+ω,ϑζ+6)+3γζ+5+3γζ+4+2γζ+3+2γζ+2ω22ϱ(ϑζ+ω,ϑζ+ω2)+ω22γζ+ω3+ω22γζ+ω4+ω42γζ+ω5+ω42γζ+ω6++γζ+1+γζω22ϱ(ϑζ+ω,ϑζ+ω2)+ω12γζ+ω3+ω22γζ+ω4+ω32γζ+ω5+ω42γζ+ω6++32γζ+1+22γζω22ϱ(ϑζ+ω,ϑζ+ω2)+1ζ2(ζ+ω3γζ+ω3+ζ+ω4γζ+ω4++ζ+1γζ+1+ζγζ)ω22ϱ(ϑζ+ω,ϑζ+ω2)+1ζ2ζ+ω1i=ζiγiω22βζ+ω2+1ζ2i=ζiγi.

    Since βζ0, we can assume that βζ>0, ζN. By a similar method, replacing γζ with βζ in (2.3), there exists ζ2N such that

    ζβζ1ζ1k, ζζ2,

    which implies limζζβζ=0 and limζβζ=0. Together (a) with (b), for every ωN, letting ζ,

    ϱ(ϑn+p,ϑζ)0.

    Thus, {ϑζ}ζN is a Cauchy sequence. Since (Υ,ϱ) is complete, there exists ϑΥ such that limζϑζ=ϑ. Now

    Q(ϱ(Πϑ,Πσ))+Q(ϱ(Πϑ,Πσ))+Q(ϱ(Πϑ,Πσ))Q(ϱ(ϑ,σ))

    holds for all ϑ,σΥ with ϱ(Πϑ,Πσ)>0. Since Q is increasing, then

    ϱ(Πϑ,Πσ)ϱ(ϑ,σ). (2.4)

    It follows that

    0ϱ(ϑζ+1,Πϑ)ϱ(ϑζ,ϑ)0 as ζ.

    Hence, ϑ=Πϑ. Suppose that ϑ and σ are two different fixed points of Π. Suppose that, Πϑ=ϑσ=Πσ and (ϑ,σ)E(Ω). Then

    +Q(ϱ(Πϑ,Πσ))Q(ϱ(ϑ,σ))Q(ϱ(ϑ,σ))=Q(ϱ(Πϑ,Πσ)),

    As ζ, which implies 0, a contradiction. Therefore ϑ=σ. Hence, Π has a unique fixed point in Υ.

    Next, we prove common fixed point theorems on complete graphical Branciari -metric space.

    Theorem 2.2. Let (Υ,ϱ) be a complete graphical Branciari -metric space with constant >1. If there exist >0 and QF,, such that Λ,Π:ΥΥ are two self mappings on Υ and satisfy

    (H1) for every ϑΥ, (ϑ,Λϑ)E(G) and (ϑ,Πϑ)E(G);

    (H2) Π and Λ are generalized G-Q contraction

    +Q(ϱ(Λϑ,Πσ))Q(ξ1ϱ(ϑ,σ)+ϱ(ϑ,Λϑ)+cϱ(σ,Πσ)), (2.5)

    for any ξ1,,c[0,1) with ξ1++c<1,<1, and min{ϱ(Λϑ,Πσ),ϱ(ϑ,σ),ϱ(ϑ,Λϑ),ϱ(σ,Πσ)}>0 for any (ϑ,σ)E(G). Then Λ and Π have a unique common fixed point.

    Proof. Let ϑ0Υ. Suppose that Λϑ0=ϑ0, then the proof is finished, so we assume that Λϑ0ϑ0. As (ϑ0,Λϑ0)E(G), so (ϑ0,ϑ1)E(G). Also, (ϑ1,Πϑ1)E(G) gives (ϑ1,ϑ2)E(G). Continuing this way, we define a sequence {ϑj} in Υ such that (ϑj,ϑj+1)E(G) with

    Λϑ2j=ϑ2j+1,Πϑ2j+1=ϑ2j+2,j=0,1,2,. (2.6)

    Combining with (2.5) and (2.6), we have

    +Q(ϱ(ϑ2j+1,ϑ2j+2))=+Q(ϱ(Λϑ2j,Πϑ2j+1))Q(ξ1ϱ(ϑ2j,ϑ2j+1)+ϱ(ϑ2j,Λϑ2j)+cϱ((ϑ2j+1,Πϑ2j+2))=Q(ξ1ϱ(ϑ2j,ϑ2j+1)+ϱ(ϑ2j,ϑ2j+1)+cϱ((ϑ2j+1,ϑ2j+2)).

    Let λ=(ξ1+)/(1c), 0<λ<1 since ξ1++c<1. Using the strictly monotone increasing property of Q,

    ϱ((ϑ2j+1,ϑ2j+2))<ξ1+1cϱ(ϑ2j,ϑ2j+1)=λϱ(ϑ2j,ϑ2j+1).

    Similarly,

    ϱ((ϑ2j+2,ϑ2j+3))<ξ1+1cϱ(ϑ2j+1,ϑ2j+2)=λϱ(ϑ2j+1,ϑ2j+2).

    Hence,

    ϱ((ϑζ,ϑζ+1))<λϱ(ϑζ1,ϑζ),ζN.

    For any ζN, we obtain

    ϱ((ϑζ,ϑζ+1))<λϱ(ϑζ1,ϑζ)<λ2ϱ(ϑζ2,ϑζ1)<<λζϱ(ϑ0,ϑ1).

    Notice that

    +Q(ϱ(ϑ1,ϑ3))=+Q(ϱ(Λϑ0,Λϑ2))Q(ϱ(ϑ0,ϑ2)),+Q(ϱ(ϑ2,ϑ4))=+Q(ϱ(Πϑ1,Πϑ3))Q(ϱ(ϑ1,ϑ3)).

    Since Q is strictly monotone increasing, we have

    ϱ(ϑ1,ϑ3)1ϱ(ϑ0,ϑ2),ϱ(ϑ2,ϑ4)1ϱ(ϑ1,ϑ3)12ϱ(ϑ0,ϑ2).

    By induction, we obtain

    ϱ(ϑζ,ϑζ+2)<1ζϱ(ϑ0,ϑ2),ζN.

    We consider the following two cases:

    (ⅰ) Let m=ζ+ω, if ω is odd and ω>2, we have

    ϱ(ϑζ,ϑm)(ϱ(ϑζ,ϑζ+1)+ϱ(ϑζ+1,ϑζ+2)+ϱ(ϑζ+2,ϑm))ϱ(ϑζ,ϑζ+1)+ϱ(ϑζ+1,ϑζ+2)+2ϱ(ϑn+2,ϑζ+3)+2ϱ(ϑζ+3,ϑζ+4)+2ϱ(ϑζ+4,ϑm)ϱ(ϑζ,ϑζ+1)+ϱ(ϑζ+1,ϑζ+2)+2ϱ(ϑn+2,ϑζ+3)+2ϱ(ϑζ+3,ϑζ+4)+3ϱ(ϑζ+4,ϑζ+5)+3ϱ(ϑζ+5,ϑζ+6)++mζ2ϱ(ϑm2,ϑm1)+mζ2ϱ(ϑm1,ϑm)λζϱ(ϑ0,ϑ1)+λζ+1ϱ(ϑ0,ϑ1)+2λζ+2ϱ(ϑ0,ϑ1)+2λζ+3ϱ(ϑ0,ϑ1)+3λζ+3ϱ(ϑ0,ϑ1)+3λζ+4ϱ(ϑ0,ϑ1)++mζ2λm2ϱ(ϑ0,ϑ1)+mζ2λm1ϱ(ϑ0,ϑ1)(λζ+2λζ+2+3λζ+4++mζ2λm2)ϱ(ϑ0,ϑ1)+(λζ+1+2λζ+3+3λζ+5++mζ2λm1)ϱ(ϑ0,ϑ1)(λζ+2λζ+2+3λζ+4++mζ2λm2)(1+λ)ϱ(ϑ0,ϑ1)λζ(1+λζ+2+2λζ+4++mζ22λmζ2)(1+λ)ϱ(ϑ0,ϑ1)λζ.1mζ2λmζ1λ2(1+λ)ϱ(ϑ0,ϑ1)λζ.1ω2λω1λ2(1+λ)ϱ(ϑ0,ϑ1).

    (ⅱ) Let m=ζ+ω, if ω is even and ω>2, we have

    ϱ(ϑζ,ϑm)(ϱ(ϑζ,ϑζ+1)+ϱ(ϑζ+1,ϑζ+2)+ϱ(ϑζ+2,ϑm))ϱ(ϑζ,ϑζ+1)+ϱ(ϑζ+1,ϑζ+2)+2ϱ(ϑn+2,ϑζ+3)+2ϱ(ϑζ+3,ϑζ+4)+2ϱ(ϑζ+4,ϑm)ϱ(ϑζ,ϑζ+1)+ϱ(ϑζ+1,ϑζ+2)+2ϱ(ϑn+2,ϑζ+3)+2ϱ(ϑζ+3,ϑζ+4)+3ϱ(ϑζ+4,ϑζ+5)+3ϱ(ϑζ+5,ϑζ+6)++mζ22ϱ(ϑm4,ϑm3)+mζ22ϱ(ϑm3,ϑm2)+mζ22ϱ(ϑm2,ϑm)λζϱ(ϑ0,ϑ1)+λζ+1ϱ(ϑ0,ϑ1)+2λζ+2ϱ(ϑ0,ϑ1)+2λζ+3ϱ(ϑ0,ϑ1)+3λζ+3ϱ(ϑ0,ϑ1)+3λζ+4ϱ(ϑ0,ϑ1)++mζ22λm4ϱ(ϑ0,ϑ1)+mζ22λm3ϱ(ϑ0,ϑ1)+mζ22ϱ(ϑm2,ϑm)(λζ+2λζ+2+3λζ+4++mζ22λm4)ϱ(ϑ0,ϑ1)+(λζ+1+2λζ+3+3λζ+5++mζ22λm3)ϱ(ϑ0,ϑ1)+mζ22ϱ(ϑm2,ϑm)(λζ+2λζ+2+3λζ+4++mζ22λm4)(1+λ)ϱ(ϑ0,ϑ1)+mζ22ϱ(ϑm2,ϑm)λζ(1+λζ+2+2λζ+4++mζ42λmζ4)(1+λ)ϱ(ϑ0,ϑ1)+mζ22ϱ(ϑm2,ϑm)λζ.1mζ22λmζ21λ2(1+λ)ϱ(ϑ0,ϑ1)+ζϱ(ϑ0,ϑ2)λζ.1ω22λω21λ2(1+λ)ϱ(ϑ0,ϑ1)+ζϱ(ϑ0,ϑ2).

    As m,ζ, ϱ(ϑζ,ϑm)0 for all ω>2. Hence, {ϑζ} is a Cauchy sequence in Υ. Since (Υ,ϱ) is complete, there exists zΥ such that

    limζϱ(ϑζ,z)=0.

    Suppose that ϱ(Λz,z)>0, then

    +Q(ϱ(Λz,ϑ2j+2))Q(ξ1ϱ(z,ϑ2j+1)+ϱ(z,Λz)+cϱ(ϑ2j+1,ϑ2j+2)).

    Using the strictly monotone increasing property of Q, we get

    ϱ(Λz,ϑ2j+2)<ξ1ϱ(z,ϑ2j+1)+ϱ(z,Λz)+cϱ(ϑ2j+1,ϑ2j+2)).

    We can also see that

    ϱ(Λz,z)<[ϱ(Λz,ϑ2j+2)+ϱ(ϑ2j+2,ϑ2j+1)+ϱ(ϑ2j+1,z)].

    It follows that

    1ϱ(Λz,z)limj infϱ(Λz,ϑ2j+2)limj supϱ(Λz,ϑ2j+2)ϱ(z,Λz).

    Hence, 1 which is an absurdity. Therefore, ϱ(Λz,z)=0. Similarly, we can obtain Πz=z. Therefore, we have

    Πz=Λz=z.

    Suppose that ϑ and σ are two different common fixed points of Λ and Π. Suppose that, Λϑ=ϑσ=Πσ and (ϑ,σ)E(G). Then,

    +Q(ϱ(ϑ,σ))=+Q(ϱ(Λϑ,Πσ))Q(ξ1ϱ(ϑ,σ)+ϱ(ϑ,Λϑ)+cϱ(σ,Πσ))=Q(ξ1ϱ(ϑ,σ)+ϱ(ϑ,ϑ)+cϱ(σ,σ)).

    Using the strictly monotone increasing property of Q, (1ξ1)ϱ(ϑ,σ)<0, which is an absurdity. Hence ϑ=σ.

    Example 2.1. Let Υ=ΓΨ, where Γ={1ζ:ζ{2,3,4,5}} and Ψ=[1,2]. For any ϑ,σΥ, we define ϱ:Υ×Υ[0,+) by

    {ϱ(ϑ,σ)=ϱ(σ,ϑ)for allϑ,σΥ,ϱ(ϑ,σ)=0ϑ=σ.

    and

    {ϱ(12,13)=ϱ(13,14)=ϱ(14,15)=16,ϱ(12,14)=ϱ(13,15)=17,ϱ(12,15)=ϱ(12,14)=12,ϱ(ϑ,σ)=|ϑσ|2,otherwise.

    Clearly, (Υ,ϱ) is a complete graphical Branciari -metric space with constant =3>1. Define the graph Ω by E(Ω)=Δ+{(13,14),(13,15),(12,2),(12,14),(15,12),(15,2),(2,1),(2,13),(1,14),(1,12)}.

    Figure 2 represents the directed graph Ω. Let Π:ΥΥ be a mapping satisfying

    Πϑ={12,ϑΓ,13,ϑΨ.
    Figure 2.  Graph Ω described in Example 2.3.

    Now, we verify that Π is a Ω-Q-contraction. We take ϑ=14Γ,σ=2Ψ, and =0.1. Then, ϱ(Πϑ,Πσ)=ϱ(12,13)=16>0 and

    0.1+3ϱ(Πϑ,Πσ)=0.6<3.0625=ϱ(ϑ,σ).

    Let Q:(0,+)R be a mapping defined by Q(ϑ)=ϑ, then it is easy to see that QF,. Therefore

    +Q(.ϱ(Πϑ,Πσ))Q(ϱ(ϑ,σ)).

    Hence, Π fulfills the conditions of Theorem 2.1 and ϑ=12 is the unique fixed point of Π.

    Consider the integral equation:

    ϑ(ρ)=μ(ρ)+ξ10m(ρ,φ)θ(φ,ϑ(φ))dφ,ρ[0,ξ1],ξ1>0. (3.1)

    Let Υ=C([0,ξ1],R) be the set of real continuous functions defined on [0,ξ1] and the mapping Π:ΥΥ defined by

    Π(ϑ(ρ))=μ(ρ)+ξ10m(ρ,φ)θ(φ,ϑ(φ))dφ,ρ[0,ξ1]. (3.2)

    Obviously, ϑ(ρ) is a solution of integral Eq (3.1) iff ϑ(ρ) is a fixed point of Π.

    Theorem 3.1. Suppose that

    (R1) The mappings m:[0,ξ1]×R[0,+), θ:[0,ξ1]×RR, and μ:[0,ξ1]R are continuous functions.

    (R2) >0 and >1 such that

    |θ(φ,ϑ(φ))θ(φ,σ(φ))|e|ϑ(φ)σ(φ)| (3.3)

    for each φ[0,ξ1] and ϑσ (i.e., ϑ(φ)σ(φ))

    (R3) ξ10m(ρ,φ)dφ1.

    (R4) ϑ0C([0,ξ1],R) such that ϑ0(ρ)μ(ρ)+ξ10m(ρ,φ)θ(φ,ϑ0(φ))dφ for all ρ[0,ξ1]

    Then, the integral Eq (3.1) has a unique solution in the set {ϑC([0,ξ1],R):ϑ(ρ)ϑ0(ρ)orϑ(ρ)ϑ0(ρ),for allφ[0,ξ1]}.

    Proof. Define ϱ:Υ×Υ[0,+) given by

    ϱ(ϑ,σ)=supρ[0,ξ1]|ϑ(ρ)σ(ρ)|2

    for all ϑ,σΥ. It is easy to see that, (Υ,ϱ) is a complete graphical Branciari -metric space with 1. Define Π:ΥΥ by

    Π(ϑ(ρ))=μ(ρ)+ξ10m(ρ,φ)θ(φ,ϑ(φ))dφ,ρ[0,ξ1]. (3.4)

    Consider a graph Ω consisting of V(Ω):=Υ and E(Ω)={(ϑ,σ)Υ×Υ:ϑ(ρ)σ(ρ)}. For each ϑ,σΥ with (ϑ,σ)E(Ω), we have

    |Πϑ(ρ)Πσ(ρ)|2=|a0m(ρ,φ)[θ(φ,ϑ(φ))θ(φ,σ(φ))]du|2(a0m(ρ,φ)e|ϑ(φ)σ(φ)|du)2e(a0m(ρ,φ)du)2supφ[0,a]|ϑ(φ)σ(φ)|2eϱ(ϑ,σ).

    Thus,

    ϱ(Πϑ,Πσ)eϱ(ϑ,σ),

    which implies that

    +ln(ϱ(Πϑ,Πσ))ln(ϱ(ϑ,σ)),

    for each ϑ,σΥ. By (R4), we have (ϑ0,Πϑ0)E(Ω), so that [ϑ0]lΩ={ϑC([0,ξ1],R):ϑ(ρ)ϑ0(ρ)orϑ(ρ)ϑ0(ρ),for allφ[0,ξ1]}. Therefore, all the hypotheses of Theorem 2.1 are fulfilled. Hence, the integral equation has a unique solution.

    We recall many important definitions from fractional calculus theory. For a function ϑC[0,1], the Reiman-Liouville fractional derivative of order δ>0 is given by

    1Γ(ξδ)dξdtξt0ϑ(e)de(te)δξ+1=Dδϑ(t),

    provided that the right hand side is pointwise defined on [0,1], where [δ] is the integer part of the number δ,Γ is the Euler gamma function. For more details, one can see [26,27,28,29].

    Consider the following fractional differential equation

    eDηϑ(t)+f(t,ϑ(t))=0,0t1,1<η2;ϑ(0)=ϑ(1)=0, (4.1)

    where f is a continuous function from [0,1]×R to R and eDη represents the Caputo fractional derivative of order η and it is defined by

    eDη=1Γ(ξη)t0ϑξ(e)de(te)ηξ+1.

    Let Υ=(C[0,1],R) be the set of all continuous functions defined on [0,1]. Consider ϱ:Υ×ΥR+ to be defined by

    ϱ(ϑ,ϑ)=supt[0,1]|ϑ(t)ϑ(t)|2

    for all ϑ,ϑΥ. Then (Υ,ϱ) is a complete graphical Branciari -metric space with 1. The given fractional differential equation (4.1) is equivalent to the succeeding integral equation

    ϑ(t)=10G(t,e)f(q,ϑ(e))de,

    where

    G(t,e)={[t(1e)]η1(te)η1Γ(η),0et1,[t(1e)]η1Γ(η),0te1.

    Define Π:ΥΥ defined by

    Πϑ(t)=10G(t,e)f(q,ϑ(e))de.

    It is easy to note that if ϑΠ is a fixed point of Π then ϑ is a solution of the problem (4.1).

    Theorem 4.1. Assume the fractional differential Eq (4.1). Suppose that the following conditions are satisfied:

    (S1) there exists t[0,1], (0,1) and ϑ,ϑΥ such that

    |f(t,ϑ)f(t,ϑ)|e|ϑ(t)ϑ(t)|

    for all ϑϑ (i.e., ϑ(t)ϑ(t)).

    (S2)

    supt[0,1]10|G(t,e)|de1.

    (S3) ϑ0C([0,1],R) such that ϑ0(t)10G(t,e)f(q,ϑ(e))de for all t[0,1].

    Then the fractional differential Eq (4.1) has a unique solution in the set {ϑC([0,1],R):ϑ(t)ϑ0(t)orϑ(t)ϑ0(t),for allt[0,1]}.

    Proof. Consider a graph Ω consisting of V(Ω):=Υ and E(Ω)={(ϑ,ϑ)Υ×Υ:ϑ(ρ)σ(ρ)}. For each ϑ,ϑΥ with (ϑ,ϑ)E(Ω), we have

    |Πϑ(t)Πϑ(t)|2=|10G(t,e)f(q,ϑ(e))de1oG(t,e)f(q,ϑ(e))de|2(10|G(t,e)|de)2(10|f(q,ϑ(e))f(q,ϑ(e))|de)2e|ϑ(t)ϑ(t)|2.

    Taking the supremum on both sides, we get

    +ln(ϱ(Πϑ,Πϑ))ln(ϱ(ϑ,ϑ)),

    for each ϑ,ϑΥ. By (S3), we have (ϑ0,Πϑ0)E(Ω), so that [ϑ0]lΩ={ϑC([0,1],R):ϑ(t)ϑ0(t)orϑ(t)ϑ0(t),for allt[0,1]}. Therefore, all the hypotheses of Theorem 2.1 are fulfilled. Hence, the fractional differential Eq (4.1) has a unique solution.

    In this paper, we have established fixed point results for Ω-Q-contraction in the setting of complete graphical Branciari -metric spaces. The directed graphs have been supported by Figures 1 and 2. The proven results have been supplemented with a non-trivial example and also applications to solve Fredholm integral equation and fractional differential equation have also been provided.

    All authors equally contributed to this work. All authors read and approved the final manuscript.

    The authors extend their appreciation to Prince Sattam bin Abdulaziz University for funding this research work through the Project Number PSAU/2023/01/24098.

    The authors convey their sincere appreciation to the Chief/Guest Editor and the anonymous reviewers for their suggestion and valuable comments which helped to bring the manuscript to its present form.

    The authors declare no conflict of interest.



    [1] H. M. Barakat, O. M. Khaled, H. A. Ghonem, New method for prediction of future order statistics, QTQM, 18 (2021), 101-116.
    [2] N. Balakrishnan, E. Beutner, E. Cramer, Exact two-sample non-parametric confindence, prediction, and tolerance intervals based on ordinary and progressively type-II right censored data, Test, 19 (2010), 68-91. doi: 10.1007/s11749-008-0133-7
    [3] H. M. Barakat, E. M. Nigm, M. E. El-Adll, M. Yusuf, Prediction for future exponential lifetime based on random number of generalized order statistics under a general set-up, Stat. Pap., 59 (2018), 605-631. doi: 10.1007/s00362-016-0779-2
    [4] P. Dellaportas, D. Wright, Numerical prediction for the two-parameter Weibull distribution, Statistician, 40 (1991), 365-372. doi: 10.2307/2348725
    [5] K. S. Kaminsky, L. S. Rhodin, Maximum likelihood prediction, Ann. Inst. Stat. Math., 37 (1985), 507-517. doi: 10.1007/BF02481119
    [6] D. Kundu, M. Z. Raqab, Bayesian inference and prediction of order statistics for a Type-II censored Weibull distribution, J. Stat. Plan. Infer., 142 (2012), 41-47. doi: 10.1016/j.jspi.2011.06.019
    [7] M. Z. Raqab, H. N. Nagaraja, On some predictors of future order statistic, Metron, 53 (1995), 85-204.
    [8] A. Saadati Nik, A. Asgharzadeh, M. Z. Raqab, Prediction methods for future failure times based on type-II right-censored samples from new Pareto-type distribution, J. Stat. Theory Pract., 2020, 1-20.
    [9] G. Volovskiy, U. Kamps, Maximum observed likelihood prediction of future record values, TEST, 29 (2020), 1072-1097. doi: 10.1007/s11749-020-00701-7
    [10] A. E. Aly, H. M. Barakat, M. E. El-Adll, Prediction intervals of the record-values process, Revstat-Stat. J., 17 (2019), 401-427.
    [11] H. M. Barakat, M. E. El-Adll, A. E. Aly, Exact prediction intervals for future exponential lifetime based on random generalized order statistics, Comput. Math. Appl., 61 (2011), 1366-1378. doi: 10.1016/j.camwa.2011.01.002
    [12] H. M. Barakat, M. E. El-Adll, A. E. Aly, Prediction intervals of future observations for a sample of random size for any continuous distribution, Math. Comput. Simulat., 97 (2014), 1-13. doi: 10.1016/j.matcom.2013.06.007
    [13] H. M. Barakat, E. M. Nigm, R. A. Aldallal, Exact prediction intervals for future current records and record range from any continuous distribution, SORT, 38 (2014), 251-270.
    [14] H. M. Barakat, O. M. Khaled, H. A. Ghonem, PredictionR: Prediction for future data from any continuous distribution, 2020. Available from: https://CRAN.R-project.org/package=PredictionR.
    [15] H. M. Barakat, O. M. Khaled, H. A. Ghonem, Predicting future lifetime for mixture exponential distribution, Commun. Stat.-Simul. Comput., 2020. DOI: 10.1080/03610918.2020.1715434.
    [16] M. E. El-Adll, A. E. Aly, Prediction intervals of future generalized order statistics from pareto distribution, J. Appl. Stat. Sci., 22 (2016), 111-125.
    [17] W. Fan, Y. Jiang, S. Huang, W. Liu, Research and prediction of opioid crisis based on BP neural network and Markov chain, AIMS Math., 4 (2019), 1357-1368. doi: 10.3934/math.2019.5.1357
    [18] H. K. Hsieh, Prediction interval for Weibull observation, based on early-failure data, IEEE Trans. Reliab., 45 (1996), 666-670. doi: 10.1109/24.556591
    [19] K. S. Kaminsky, P. I. Nelson, Prediction of order statistics, Handb. Stat., 17 (1998), 431-450. doi: 10.1016/S0169-7161(98)17017-7
    [20] J. F. Lawless, Statistical models and methods for lifetime data, Wiley, New York, 2003.
    [21] J. K. Patel, Prediction intervals review, Commun. Stat.-Theory Methods, 18 (1989), 2393-2465. doi: 10.1080/03610928908830043
    [22] W. Peng, B. Liang, Y. Xia, X. Tong, Predicting disease risks by matching quantiles estimation for censored data, Math. Biosci. Eng. (AIMS), 17 (2020), 4544-4562. doi: 10.3934/mbe.2020251
    [23] M. Z. Raqab, H. M. Barakat, Prediction intervals for future observations based on samples of random sizes, J. Math. Stat., 14 (2018), 16-28. doi: 10.3844/jmssp.2018.16.28
    [24] I. A. Shah, H. M. Barakat, A. H. Khan, Characterizations through generalized and dual generalized order statistics, with an application to statistical prediction problem, Stat. Probabil. Lett., 163 (2020), 108782. doi: 10.1016/j.spl.2020.108782
    [25] R. Valiollahi, A. Asgharzadeh, D. Kundu, Prediction of future failures for generalized exponential distribution under type-I or type-II hybrid censoring, Braz. J. Probab. Stat., 31 (2017), 41-61. doi: 10.1214/15-BJPS302
    [26] E. K. Al-Hussaini, F. Al-Awadhi, Bayes two-sample prediction of generalized order statistics with fixed and random sample size, J. Stat. Comput. Sim., 80 (2010), 13-28. doi: 10.1080/00949650802440871
    [27] F. Louzada, E. Bereta, M. Franco, On the distribution of the minimum or maximum of a random number of i.i.d. lifetime random variables, Appl. Math., 3 (2012), 350-353. doi: 10.4236/am.2012.34054
    [28] A. H. Khan, I. A. Shah, M. Ahsanullah, Characterization through distributional properties of order statistics, J. Egypt. Math. Soc., 20 (2012), 211-214. doi: 10.1016/j.joems.2012.10.002
    [29] S. Y. Oncel, M. Ahsanullah, F. A. Aliev, F. Aygun, Switching record and order statistics via random contraction, Stat. Probabil. Lett., 73 (2005), 207-217. doi: 10.1016/j.spl.2005.03.004
    [30] I. A. Shah, A. H. Khan, H. M. Barakat, Random translation, dilation and contraction of order statistics, Stat. Probabil. Lett., 92 (2014), 209-214. doi: 10.1016/j.spl.2014.05.025
    [31] I. A. Shah, A. H. Khan, H. M. Barakat, Translation, contraction and dilation of dual generalized order statistics, Stat. Probabil. Lett., 107 (2015), 131-135. doi: 10.1016/j.spl.2015.08.015
    [32] I. A. Shah, H. M. Barakat, A. H. Khan, Characterization of Pareto and power function distributions by conditional variance of order statistics, C. R. Acad. Bulg. Sci., 71 (2018), 313-316.
    [33] J. Wesolowski, M. Ahsanullah, Switching order statistics through random power contractions, Aust. N. Z. J. Stat., 46 (2004), 297-303. doi: 10.1111/j.1467-842X.2004.00330.x
    [34] C. Bernard, S. Vanduffel, Quantile of a mixture, arXiv: 1411.4824v1 [stat.OT], 2014.
    [35] M. G. Badar, A. M. Priest, Statistical aspects of fiber and bundle strengthin hybrid composites, Prog. Sci. Eng. Compos., (1982), 1129-1136.
    [36] M. Z. Raqab, D. Kundu, Comparison of different estimators of P[Y<X] for a scaled Burr type X distribution, Commun. Stat.-Simul. Comput., 34 (2005), 465-483. doi: 10.1081/SAC-200055741
    [37] M. R. Ahmad, S. A. Ali, Combining two Weibull distributions using a mixing parameter, Eur. J. Sci. Res., 31 (2009), 296-305.
    [38] S. F. Ateya, Maximum likelihood estimation under a finite mixture of generalized exponential distributions based on censored data, Stat. Pap., 55 (2014), 311-325. doi: 10.1007/s00362-012-0480-z
    [39] L. Li, Z. Yang, Z. Dang, C. Meng, J. Huang, H. Meng, et al., Propagation analysis and prediction of the COVID-19, Infect. Dis. Modell., 5 (2020), 282-292. doi: 10.1016/j.idm.2020.03.002
    [40] J. M. Read, J. R. Bridgen, D. A. Cummings, A. Ho, C. P. Jewell, Novel coronavirus 2019-nCoV: early estimation of epidemiological parameters and epidemic predictions, MedRxiv, 2020.
    [41] M. R. Leadbetter, G. Lindgren, H. Rootzén, Extremes and related properties of random sequences and processes, Springer, Berlin, 1983.
    [42] W. Nelson, Weibull prediction of a future number of failures, Qual. Reliab. Eng. Int., 16 (2000), 23-26. doi: 10.1002/(SICI)1099-1638(200001/02)16:1<23::AID-QRE283>3.0.CO;2-Q
  • Reader Comments
  • © 2021 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(3073) PDF downloads(186) Cited by(8)

Figures and Tables

Tables(8)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog