
In this paper, we carry out stochastic comparisons of lifetimes of series and parallel systems with dependent heterogeneous Topp-Leone generated components. The usual stochastic order and the reversed hazard rate order are developed for the parallel systems with the help of vector majorization under Archimedean copula dependence, and the results for the usual stochastic order are also obtained for the series systems. Finally, some numerical examples are provided to illustrate the effectiveness of our theoretical findings.
Citation: Li Zhang, Rongfang Yan. Stochastic comparisons of series and parallel systems with dependent and heterogeneous Topp-Leone generated components[J]. AIMS Mathematics, 2021, 6(3): 2031-2047. doi: 10.3934/math.2021124
[1] | Xiao Zhang, Rongfang Yan . Stochastic comparisons of extreme order statistic from dependent and heterogeneous lower-truncated Weibull variables under Archimedean copula. AIMS Mathematics, 2022, 7(4): 6852-6875. doi: 10.3934/math.2022381 |
[2] | Bin Lu, Rongfang Yan . Ordering results of second order statistics from random and non-random number of random variables with Archimedean copulas. AIMS Mathematics, 2021, 6(6): 6390-6405. doi: 10.3934/math.2021375 |
[3] | Bin Lu . Stochastic comparisons of second-order statistics from dependent and heterogeneous modified proportional (reversed) hazard rates scale models. AIMS Mathematics, 2024, 9(4): 8904-8919. doi: 10.3934/math.2024434 |
[4] | Ahmed M. Gemeay, Najwan Alsadat, Christophe Chesneau, Mohammed Elgarhy . Power unit inverse Lindley distribution with different measures of uncertainty, estimation and applications. AIMS Mathematics, 2024, 9(8): 20976-21024. doi: 10.3934/math.20241021 |
[5] | Mingxia Yang . Orderings of the second-largest order statistic with modified proportional reversed hazard rate samples. AIMS Mathematics, 2025, 10(1): 311-337. doi: 10.3934/math.2025015 |
[6] | Miaomiao Zhang, Bin Lu, Rongfang Yan . Ordering results of extreme order statistics from dependent and heterogeneous modified proportional (reversed) hazard variables. AIMS Mathematics, 2021, 6(1): 584-606. doi: 10.3934/math.2021036 |
[7] | Mohamed Said Mohamed, Muqrin A. Almuqrin . Properties of fractional generalized entropy in ordered variables and symmetry testing. AIMS Mathematics, 2025, 10(1): 1116-1141. doi: 10.3934/math.2025053 |
[8] | Mohamed Said Mohamed, Najwan Alsadat, Oluwafemi Samson Balogun . Continuous Tsallis and Renyi extropy with pharmaceutical market application. AIMS Mathematics, 2023, 8(10): 24176-24195. doi: 10.3934/math.20231233 |
[9] | Mansour Shrahili, Mohamed Kayid, Mhamed Mesfioui . Stochastic inequalities involving past extropy of order statistics and past extropy of record values. AIMS Mathematics, 2024, 9(3): 5827-5849. doi: 10.3934/math.2024283 |
[10] | Yassine Sabbar, Aeshah A. Raezah . Threshold analysis of an algae-zooplankton model incorporating general interaction rates and nonlinear independent stochastic components. AIMS Mathematics, 2024, 9(7): 18211-18235. doi: 10.3934/math.2024889 |
In this paper, we carry out stochastic comparisons of lifetimes of series and parallel systems with dependent heterogeneous Topp-Leone generated components. The usual stochastic order and the reversed hazard rate order are developed for the parallel systems with the help of vector majorization under Archimedean copula dependence, and the results for the usual stochastic order are also obtained for the series systems. Finally, some numerical examples are provided to illustrate the effectiveness of our theoretical findings.
Order statistics play a significant role in many areas such as statistics, reliability, auction theory, risk management and many other branches of applied probability. For n random variables X1,…,Xn, denote Xk:n the kth smallest order statistic of X1,…,Xn. Then X1:n≤⋯≤Xn:n, and Xk:n represents the lifetime of (n−k+1)-out-of-n system. In particularly, X1:n and Xn:n respectively represent the lifetimes of series and parallel systems. Thus, to investigate the lifetime of k-out-of-n system is equivalent to study the stochastic properties of the order statistics. For more details, one may refer to Balakrishnan and Rao [1], Khaledi and Kochar [2], Zhao and Balakrishnan [3] and Fang and Zhang [4]. In the past few decades, stochastic comparisons of order statistics have been studied by many scholars, for example, Barlow and Proschan [5], Bartoszewicz [6], Boland et al. [7], Kundu et al. [8] and Balakrishnan et al. [9].
In the past decades, most of the work is developed on the independent and identically distributed random variables. For comprehensive references, interested readers may refer to Kochar [10] and Balakrishnan and Zhao [11]. Further, a considerable amount of work has also been carried out comparing systems with independent heterogeneous components under specific distributions. For more on this topic, please see Dykstra et al. [12], Balakrishnan et al. [13] and Torrado [14]. However, the components of the system share many complex factors when the system is functioning, such as environmental conditions and work stress. For these reasons, it would be practicable to consider dependent lifetimes of components. Recently, the dependence structure of the components is investigated by researchers with the help of copula theory. Archimedean copula has been considered by many scholars due to its flexibility, for instance, Clayton copula, Ali-Mikhail-Haq copula, and Gumbel-Hougaard copula. For example, Navarro and Spizzichino [15] studied the stochastic ordering of series and parallel systems with components sharing a common copula. Li and Li [16] considered ordering properties of the smallest order statistics of Weibull samples having a common Archimedean copula. Li et al. [17] discussed stochastic comparisons of extreme order statistics from scaled and interdependent random variables. Fang et al. [18] presented conditions to stochastically compare the extreme order statistics from dependent and heterogeneous random variables. Kundu and Chowdhury [19] discussed the lifetimes of two series and parallel systems with location-scale components assembled with some kind of Archimedean copulas under different stochastic orders. Fang et al. [20] obtained various ordering results for comparing the lifetimes of the series and the parallel systems, where each component follows scale proportional hazard or reversed hazard models with Archimedean copula. For proportional hazard rate and proportional reversed hazard rate models, Li and Li [21] developed ordering properties of extreme order statistics from heterogeneous dependent random variables in the sense of the hazard rate and the reversed hazard rate orders.
In this manuscript, we study the stochastic comparisons of series and parallel systems having Topp-Leone generated components with different scale and shape parameters. The Topp-Leone generated family of distribution was given by Sadegh Rezaei [22] as a generalization of Topp and Leone's distribution. It has the property to model bathtub-shaped hazard rates depending on the values of parameters and can be used for lifetime modeling. For more applications of the distribution, one may refer to Sadegh Rezaei [22]. A random variable X is said to be Topp-Leone generated (TL−G) family of distribution if its cumulative distribution function is
F(x;θ,α,ξ)=[G(x;ξ)θ(2−G(x;ξ)θ)]α,x≥0,θ>0,α>0, |
where θ is the scale parameter, α is the shape parameter, G(x;ξ) is the baseline distribution function, and ξ is the parameter specifying the baseline distribution, and denote X∼TL−G(α,θ,ξ). For convenience, G(x;ξ) is written by G(x). TL−G(α,θ,ξ) is reduced to Topp and Leone's distribution when G(x)∼U(0,1) and θ=1.
The organization of the paper is as follows. In Section 2, we present some fundamental definitions and lemmas. In Section 3, we develop the usual stochastic and the reversed hazard rate orders of series and parallel systems with dependent heterogeneous components under Archimedean copulas. Some numerical examples are provided to illustrate theoretical findings. Section 4 concludes the paper.
In this section, we first recall some basic definitions of some well-known notions of stochastic orders, majorization orders, and Archimedean copula, and introduce some lemmas may be used in the sequel. Denote R=(−∞,+∞) and Rn=(−∞,+∞)n.
Consider two absolutely continuous random variables X and Y have distribution functions F(x) and G(x), survival functions ˉF(x)=1−F(x) and ˉG(x)=1−G(x), density functions f(x) and g(x), hazard rate functions hX(x)=f(x)/ˉF(x) and hY(x)=g(x)/ˉG(x), and reversed hazard rate functions ˜γX(x)=f(x)/F(x) and ˜γY(x)=g(x)/G(x), respectively.
Definition 1. X is smaller than Y in the
(ⅰ) usual stochastic order (denoted by X≤stY) if ˉF(x)≤ˉG(x) for all x∈R;
(ⅱ) hazard rate order (denoted by X≤hrY) if ˉG(x)/ˉF(x) is increasing in x∈R, or hX(x)≥hY(x) for all x∈R;
(ⅲ) reversed hazard rate order (denoted by X≤rhY) if G(x)/F(x) is increasing in x∈R, or ˜γX(x)≤˜γY(x) for all x∈R.
X≤stY means X is less likely than Y to take on large values, where "large" means any value greater than x, and that this is the case for all x′s. The reversed hazard rate could be understood as the probability intensity of a component survival to the last moment t given that its lifetime does not exceed t (cf. Yan and Luo[23]). The hazard rate is well known and has been widely applied. As a dual concept of the hazard rate, the reversed hazard rate is far less popular and frequently used. Recently, the reversed hazard rate order has received great attention because it is more appropriate and effective than the hazard rate order for the study of some particular problems such as assessing waiting time, hidden failures, inactivity times, etc. (cf. Veres-Ferrer and Pavia[24]).
It is well known that the hazard rate order or the reversed hazard rate order implies the usual stochastic order, and not conversely. For more detailed discussions and applications of stochastic orders, please refer to Shaked and Shanthikumar [25]. In addition, some scholars recently have compared order statistics under some weaker orders such as the second-order stochastic dominance, the details can be referred to Lando et al.[26,27].
Let x1:n≤x2:n≤⋯≤xn:n and y1:n≤y2:n≤⋯≤yn:n be the increasing arrangements of the elements of vectors x=(x1,…,xn) and y=(y1,…,yn), respectively. In particular, x≤y means xi≤yi, for all i=1,…,n.
Definition 2. For two vectors x=(x1,…,xn) and y=(y1,…,yn), x is said to
(ⅰ) majorize y (written as xm⪰y) if ∑ji=1xi:n≤∑ji=1yi:n,j=1,…,n−1, and ∑ni=1xi:n=∑ni=1yi:n;
(ⅱ) supermajorize y (written as xw⪰y) if ∑ji=1xi:n≤∑ji=1yi:n,j=1,…,n;
(ⅲ) submajorize y (written as x⪰wy) if ∑nj=ixj:n≥∑nj=iyj:n,i=1,…,n.
The majorization, the weak supermajorization order, and the weak submajorization order are widely used to establish various stochastic inequalities. For more details on the notions and basic properties of majorization orders, one may refer to Marshall [28].
The following we recall the notions of copula.
Definition 3. For a random vector X=(X1,X2,…,Xn) with joint distribution function F, joint survival function ˉF, univariate distribution functions F1,F2,…,Fn and univariate survival functions ˉF1,…,ˉFn. If there exist some function C:[0,1]n→[0,1] and ˆC:[0,1]n→[0,1] such that, for all xi,i=1,…,n,
F(x1,…,xn)=C(F1(x1),…,Fn(xn)),ˉF(x1,…,xn)=ˆC(ˉF1(x1),…,ˉFn(xn)), |
then C and ˆC are called the copula and survival copula of random vector X, respectively.
Definition 4. For a decreasing and continuous function ψ:[0,∞)→[0,1] such that ψ(0)=1 and ψ(+∞)=0, let ϕ=ψ−1 be its pseudo-inverse. Then,
Cψ(u1,…,un)=ψ(ϕ(u1)+⋯+ϕ(un)),ui∈(0,1),i=1,2,…,n, |
is called an Archimedean copula with generator ψ, if (−1)kψ(k)(x)≥0 for k=1,2…,n−2 and (−1)n−2ψ(n−2)(x) is decreasing and convex.
For more on Archimedean copula, readers may refer to Nelsen [29]. For convenience, from now on, we denote
ϵ+={(x1,…,xn):0<x1≤x2≤⋯≤xn},D+={(x1,…,xn):x1≥x2≥⋯≥xn>0}. |
Definition 5. A function f is said to be super-additive if f(x+y)≥f(x)+f(y), for all x and y in the domain of f.
Definition 6. A real-valued function φ defined on a set A⊆Rn is said to be Schur-convex [Schur-concave] on A if xm⪰y implies φ(x)≥[≤]φ(y) on A.
In the following, we present some lemmas will be used in proving the main results.
Lemma 1. (Li and Fang [30]. Lemma A.1) For two n-dimensional Archimedean copulas Cψ1 and Cψ2, if ϕ2∘ψ1 is super-additive, then Cψ1(u)≤Cψ2(u) for all u=(u1,…,un)∈[0,1]n.
Lemma 2. (Marshall et al. [28]. Theorem 3.A.4) Suppose J⊂R is an open interval and ϕ:Jn→R is continuously differentiable. Then, ϕ is Schur-convex [Schur-concave] on Jn if and only if
(ⅰ) ϕ is symmetric on Jn; and
(ⅱ) for all i≠j and all x∈Jn,
(xi−xj)(∂ϕ(x)∂xi−∂ϕ(x)∂xj)≥[≤]0, |
where ∂ϕ(x)∂xi represents the partial derivative of ϕ with respect to its i-th argument.
Lemma 3. (Marshall et al. [28]. Theorem 3.A.8) A real valued function ϕ on Rn, satisfies
xw≺y⇒ϕ(x)≤[≥]ϕ(y), |
if and only if ϕ is decreasing and Schur-convex [Schur-concave] on Rn. Similarly, ϕ satisfies
x≺wy⇒ϕ(x)≤[≥]ϕ(y), |
if and only if ϕ is increasing and Schur-convex [Schur-concave] on Rn.
Lemma 4. (Das and Kayal [31]. Lemma 2.6) Let the function h:(0,1)→(−∞,0) be defined as h(u)=ulnu1−u. Then, h(u) is decreasing in u for all u∈(0,1).
Lemma 5. (Li and Li [21]. Lemma 2.3) If λw⪯μ,ψ is log-concave and (ψlnψ)/ψ′ is increasing and concave, then
L(ϕ(Hλ1(x)),…,ϕ(Hλn(x)))≤L(ϕ(Hμ1(x)),…,ϕ(Hμn(x))), |
for any H(x)∈[0,1], where L(u)=ψ′(∑ni=1ui)ψ(∑ni=1ui)∑ni=1ψ(ui)lnψ(ui)ψ′(ui),ui∈[0,1],i=1,…,n and ϕ=ψ−1 is the pseudo-inverse of ψ.
Throughout this paper, the terms increasing and decreasing are used for non-decreasing and non-increasing, respectively. It is also assumed that all the random variables are non-negative and absolutely continuous.
In this section, we compare two system lifetimes with dependent heterogeneous components following the Topp-Leone generated distribution. The usual stochastic order and the reversed hazard rate order are obtained for series and parallel systems. Let X=(X1,…,Xn) be the dependent heterogeneous TL-G distributed random vectors, we denote X∼TL-G(α,θ,F), where F is the baseline distribution function, ψ1 is generator of the associated Archimedean copula, and α=(α1,…,αn) and θ=(θ1,…,θn) are the tilt parameter vector. Similarly, denote Y∼TL-G(β,δ,G). For convenience, we denote the distribution(survival) function of random variable X by HX(x)(ˉHX(x)). Then, the distribution functions of the parallel systems Xn:n and Yn:n are given by
HXn:n(x)=φ1(α,θ,F,ψ1)=ψ1(n∑i=1ϕ1((Fθi(2−Fθi))αi)) |
and
HYn:n(x)=φ2(β,δ,G,ψ2)=ψ2(n∑i=1ϕ2((Gδi(2−Gδi))βi)), |
respectively. And the reliability functions of series systems X1:n and Y1:n can be written as
ˉHX1:n(x)=ϑ1(α,θ,F,ψ1)=ψ1(n∑i=1ϕ1(1−(Fθi(2−Fθi))αi)) |
and
ˉHY1:n(x)=ϑ2(β,δ,G,ψ2)=ψ2(n∑i=1ϕ2(1−(Gδi(2−Gδi))βi)), |
respectively, where ϕ1=ψ−11 and ϕ2=ψ−12.
First, we develop sufficient conditions for the usual stochastic order between parallel systems with dependent components. In the following theorem, we consider the case of the heterogeneous shape parameters and common scale parameters.
Theorem 1. Suppose that X and Y are two random variables with distribution functions F and G, respectively. Define X∼TL-G(α,θ,F) with generator ψ1 and Y∼TL-G(β,θ,G) with generator ψ2, such that α,β and θ∈ϵ+. If X≥stY, ϕ2∘ψ1 is super-additive, and either ψ1 or ψ2 is log-convex. Then, αw⪰β⇒Xn:n≥stYn:n.
Proof. Without loss of generality, assume that ψ2 is log-convex, it follows from Lemma 1 that
φ1(α,θ,F,ψ1)≤φ1(α,θ,F,ψ2). | (3.1) |
By the decreasing property of ψ2 and ϕ2, and note that X≥stY, we have
φ1(α,θ,F,ψ2)≤φ2(α,θ,G,ψ2). | (3.2) |
Combining (3.1) and (3.2), we conclude
φ1(α,θ,F,ψ1)≤φ2(α,θ,G,ψ2). |
Thus, to prove the required result, it is sufficient to show that
φ2(α,θ,G,ψ2)≤φ2(β,θ,G,ψ2). |
Denote
φ2(α,θ,G,ψ2)=ψ2(n∑i=1ϕ2((Gθi(2−Gθi))αi)). | (3.3) |
On differentiating (3.3) with respect to αi (i=1,2,…,n), we obtain
∂φ2(α,θ,G,ψ2)∂αi=ψ′2(n∑i=1ϕ2((Gθi(2−Gθi))αi))ψ2(ϕ2((Gθi(2−Gθi))αi))ψ′2(ϕ2((Gθi(2−Gθi))αi))ln(Gθi(2−Gθi))≤0. |
Hence, φ2(α,θ,G,ψ2) is decreasing in αi (i=1,2,…,n). After simplifications, we have
∂φ2(α,θ,G,ψ2)∂αi−∂φ2(α,θ,G,ψ2)∂αj=ψ′2(n∑i=1ϕ2((Gθi(2−Gθi))αi))ψ2(ϕ2((Gθi(2−Gθi))αi))ψ′2(ϕ2((Gθi(2−Gθi))αi))ln(Gθi(2−Gθi))−ψ′2(n∑j=1ϕ2((Gθj(2−Gθj))αj))ψ2(ϕ2((Gθj(2−Gθj))αj))ψ′2(ϕ2((Gθj(2−Gθj))αj))ln(Gθj(2−Gθj)). |
Under the assumption of α,θ∈ϵ+, we have αi≤αj,θi≤θj (1≤i≤j≤n), which implies that (Gθi(2−Gθi))αi≥(Gθj(2−Gθj))αj. From the log-convexity of ψ2 that holds
ψ2(ϕ2((Gθi(2−Gθi))αi))ψ′2(ϕ2((Gθi(2−Gθi))αi))≥ψ2(ϕ2((Gθj(2−Gθj))αj))ψ′2(ϕ2((Gθj(2−Gθj))αj)). |
It is obvious that
∂φ2(α,θ,G,ψ2)∂αi−∂φ2(α,θ,G,ψ2)∂αj≥0. |
It follows immediately from Lemma 2 that φ2(α,θ,G,ψ2) is Schur-concave in α. Note that αw⪰β, we obtain that φ2(α,θ,G,ψ2)≤φ2(β,θ,G,ψ2) by Lemma 3. The proof is completed.
Remark 1. (i) Note that the super-additivity of ϕ2∘ψ1 can be roughly interpreted as follows: Kendall's τ of the copula with generator ψ2 is larger than that with generator ψ1 and hence is more positive dependent (refer to Li and Fang [30]). Theorem 1 shows that the less positive dependence, and the more heterogeneous shape parameters (in the weakly supermajorized order) lead to the stochastically large lifetime of the parallel system and higher reliability could be achieved.
(ii) It is vital to note that the super-additivity of ϕ2∘ψ1(x) in Theorem 1 is easy to check for many well-known Archimedean copulas. For example, for the Clayton copula with generator ψ(x)=(ax+1)−1a for a∈(0,∞), it is easy to verify that ψ(x)=−1aln(ax+1) is log-convex in x∈[0,1]. Let ψ1(x)=(a1x+1)−1a1 and ψ2(x)=(a2x+1)−2a2. It can be observed that ϕ2∘ψ1(x)=((a1x+1)a2a1−1)/a2. Taking twice derivative of ϕ2∘ψ1(x) with respect to x, it can be seen that [ϕ2∘ψ1(x)]″≥0 for any a2≥a1>0, which implies the super-additivity of ϕ2∘ψ1(x).
(iii) To display the whole curves of distribution functions of random variables defined on [0,+∞), we take the transformation (x+1)−1:[0,+∞)↦(0,1] for variable X. Then X≤stY is equivalent to (X+1)−1≥st(Y+1)−1. Through this transformation, we can display the graphs of the CDFs in the (0,1] interval and further realize the panoramic observation of random variables.
The following example 1 illustrates the theoretical result of Theorem 1.
Example 1. Consider F(x)=e−2x and G(x)=1−11+x,x>0. It is obvious that X≥stY. Let Xi∼TL-G(αi,θi,F) and Yi∼TL-G(βi,δi,G),i=1,2. Set (α1,α2)=(2,6)w⪰(4,5)=(β1,β2),(θ1,θ2)=(0.6,0.1)=(δ1,δ2). We further take ψ1(x)=e1−ex2 and ψ2(x)=(0.1x+1)−10.1,x>0. It is clear that ψ2 is log-convex but ψ1 is log-concave. We can easily show that ϕ2∘ψ1(x) is convex in x and ϕ2∘ψ1(0)=0, and thus ϕ2∘ψ1(x) is super-additive. As is seen in Figure 1, the survival curve of (X2:2+1)−1 beneath that of (Y2:2+1)−1 confirms X2:2(x)≥stY2:2(x). This validates the result in Theorem 1.
The following result provides some sufficient conditions for the usual stochastic order of series systems under the weakly submajorized order. We consider equal scale parameter vectors.
Theorem 2. Suppose that X and Y are two random variables with distribution functions F and G, respectively. Define X∼TL-G(α,θ,F) with generator ψ1 and Y∼TL-G(β,θ,G) with generator ψ2, such that α,β and θ∈ϵ+. If X≤stY, and ϕ2∘ψ1 is super-additive, and either ψ1 or ψ2 is log-convex. Then, α⪰wβ⇒X1:n≤stY1:n.
Proof. Consider the case of ψ2 is log-convex. The super-additivity of ϕ2∘ψ1 implies that
ϑ1(α,θ,F,ψ1)≤ϑ1(α,θ,F,ψ2). | (3.4) |
Note that ψ2, ϕ2 are decreasing, by X≤stY, we have
ϑ1(α,θ,F,ψ2)≤ϑ2(α,θ,G,ψ2). | (3.5) |
Combining (3.4) and (3.5), we get
ϑ1(α,θ,F,ψ1)≤ϑ2(α,θ,G,ψ2). |
Hence, to prove the required results, it is sufficient to show that
ϑ2(α,θ,G,ψ2)≤ϑ2(β,θ,G,ψ2). |
Denote
ϑ2(α,θ,G,ψ2)=ψ2(n∑i=1ϕ2(1−(Gθi(2−Gθi))αi)). | (3.6) |
The partial derivative of (3.6) with respect to αi(i=1,…,n) can be given by
∂ϑ2(α,θ,G,ψ2)∂αi=−1αiψ′2(n∑i=1ϕ2(1−(Gθi(2−Gθi))αi))ψ2(ϕ2(1−(Gθi(2−Gθi))αi))ψ′2(ϕ2(1−(Gθi(2−Gθi))αi))×(Gθi(2−Gθi))αiln(Gθi(2−Gθi))αi1−(Gθi(2−Gθi))αi≥0. |
Thus, ϑ2(α,θ,G,ψ2) is increasing in αi (i=1,2,…,n). We obtain
∂ϑ2(α,θ,G,ψ2)∂αi−∂ϑ2(α,θ,G,ψ2)∂αj=−1αiψ′2(n∑i=1ϕ2(1−(Gθi(2−Gθi))αi))ψ2(ϕ2(1−(Gθi(2−Gθi))αi))ψ′2(ϕ2(1−(Gθi(2−Gθi))αi))(Gθi(2−Gθi))αiln(Gθi(2−Gθi))αi1−(Gθi(2−Gθi))αi+1αjψ′2(n∑j=1ϕ2(1−(Gθj(2−Gθj))αj))ψ2(ϕ2(1−(Gθj(2−Gθj))αj))ψ′2(ϕ2(1−(Gθj(2−Gθj))αj))(Gθj(2−Gθj))αjln(Gθj(2−Gθj))αj1−(Gθj(2−Gθj))αj. | (3.7) |
Consider that αi≤αj,θi≤θj (1≤i≤j≤n), which implies that (Gθi(2−Gθi))αi≥(Gθj(2−Gθj))αj. It follows from log-convexity of ψ2 that
ψ2(ϕ2(1−(Gθi(2−Gθi))αi))ψ′2(ϕ2(1−(Gθi(2−Gθi))αi))≤ψ2(ϕ2(1−(Gθj(2−Gθj))αj))ψ′2(ϕ2(1−(Gθj(2−Gθj))αj)). |
By Lemma 4, it holds that
(Gθi(2−Gθi))αiln(Gθi(2−Gθi))αi1−(Gθi(2−Gθi))αi≤(Gθj(2−Gθj))αjln(Gθj(2−Gθj))αj1−(Gθj(2−Gθj))αj. |
For convenience, denote
A1=−1αi,B1=ψ2(ϕ2(1−(Gθi(2−Gθi))αi))ψ′2(ϕ2(1−(Gθi(2−Gθi))αi)),C1=(Gθi(2−Gθi))αiln(Gθi(2−Gθi))αi1−(Gθi(2−Gθi))αi, |
A2=−1αj,B2=ψ2(ϕ2(1−(Gθj(2−Gθj))αj))ψ′2(ϕ2(1−(Gθj(2−Gθj))αj)),C2=(Gθj(2−Gθj))αjln(Gθj(2−Gθj))αj1−(Gθj(2−Gθj))αj, |
and
D=ψ′2(n∑i=1ϕ2(1−(Gθi(2−Gθi))αi))<0. |
Note that A1≤A2<0,B1≤B2<0,C1≤C2<0. Then, equation (3.7) is equivalent to
A1B1C1D−A2B2C2D=(A1B1C1−A2B1C1+A2B1C1−A2B2C2)D={(A1−A2)B1C1+A2[(B1−B2)C1+B2(C1−C2)]}D≥0. |
Thus
∂ϑ2(α,θ,G,ψ2)∂αi−∂ϑ2(α,θ,G,ψ2)∂αj≥0. |
It follows from Lemma 2 that ϑ2(α,θ,G,ψ2) is Schur-concave in α. Based on α⪰wβ and by Lemma 3, we conclude ϑ2(α,θ,G,ψ2)≤ϑ2(β,θ,G,ψ2). This completes the proof.
Remark 2. In accordance with Theorem 2, for a series system, the more positive dependence, and the less heterogeneous (in the weakly subermajorized order) the shape parameters are, the stochastically longer lifetime of the system and hence higher reliability will be achieved. It should be pointed out that Theorem 2 expands Theorem 3.5 of Chanchal et al. [32] to the case of the heterogeneous components. It might be of great interest to study the hazard rate order between X1:n and Y1:n, which is left as an open problem.
The following example 2 illustrates that the theoretical result of Theorem 2.
Example 2. Consider F(x)=1−11+x and G(x)=e−1x,x>0. It is obvious that X≤stY. Let ψ1(x)=e1−(1+x)5 and ψ2(x)=e−x0.3,x>0, and Xi∼TL-G(αi,θi,F) and Yi∼TL-G(βi,δi,G),i=1,2, and set (α1,α2)=(2,8)⪰w(3,4)=(β1,β2),(θ1,θ2)=(1.2,3)=(δ1,δ2). It is easy to verify that ψ2 is log-convex but ψ1 is log-concave, and ϕ2∘ψ1(x) is convex in x and ϕ2∘ψ1(0)=0, which implies ϕ2∘ψ1(x) is super-additive. Then, the conditions of Theorem 2 are satisfied. Figure 2 plots distribution functions of (X1:2+1)−1 and (Y1:2+1)−1, from which it can be observed that H(X1:2+1)−1(x) is always smaller than H(Y1:2+1)−1(x), and this verifies X1:2≤stY1:2.
The next theorem establishes some sufficient conditions for the usual stochastic order, here, we consider different scale parameter vectors.
Theorem 3. Suppose that X and Y are two random variables with distribution functions F and G, respectively. Define X∼TL-G(α,θ,F) with generator ψ1 and Y∼TL-G(α,δ,G) with generator ψ2, such that θ,δ and α∈ϵ+. If X≥stY, ϕ2∘ψ1 is super-additive, and either ψ1 or ψ2 is log-convex. Then, θw⪰δ⇒Xn:n≥stYn:n.
Proof. First assume that ψ2 is log-convex. Similar to the arguments as in the proof of Theorem 1, we notice that to prove the present theorem, it is enough to show that
φ2(α,θ,G,ψ2)≤φ2(α,δ,G,ψ2). |
Denote
φ2(α,θ,G,ψ2)=ψ2(n∑i=1ϕ2((Gθi(2−Gθi))αi)). | (3.8) |
On differentiating (3.9) with respect to θi (i=1,2,…,n), we have
∂φ2(α,θ,G,ψ2)∂θi=2αilnG(1−Gθi2−Gθi)ψ′2(n∑i=1ϕ2((Gθi(2−Gθi))αi))ψ2(ϕ2((Gθi(2−Gθi))αi))ψ′2(ϕ2((Gθi(2−Gθi))αi))≤0. |
It follows that φ2(α,θ,G,ψ2) is decreasing in θi (i=1,2,…,n). We further get
∂φ2(α,θ,G,ψ2)∂θi−∂φ2(α,θ,G,ψ2)∂θj=2αilnG(1−Gθi2−Gθi)ψ′2(n∑i=1ϕ2((Gθi(2−Gθi))αi))ψ2(ϕ2((Gθi(2−Gθi))αi))ψ′2(ϕ2((Gθi(2−Gθi))αi))−2αjlnG(1−Gθj2−Gθj)ψ′2(n∑j=1ϕ2((Gθj(2−Gθj))αj))ψ2(ϕ2((Gθj(2−Gθj))αj))ψ′2(ϕ2((Gθj(2−Gθj))αj)). | (3.9) |
Further, we have αi≤αj,θi≤θj (1≤i≤j≤n). Therefore (Gθi(2−Gθi))αi≥(Gθj(2−Gθj))αj. By the log-convexity of ψ2, we conclude
ψ2(ϕ2((Gθi(2−Gθi))αi))ψ′2(ϕ2((Gθi(2−Gθi))αi))≥ψ2(ϕ2((Gθj(2−Gθj))αj))ψ′2(ϕ2((Gθj(2−Gθj))αj)) |
and
1−Gθi2−Gθi−1−Gθj2−Gθj=Gθj−Gθj(2−Gθi)(2−Gθj)≤0. |
For convenience, denote
A′1=ψ2(ϕ2((Gθi(2−Gθi))αi))ψ′2(ϕ2((Gθi(2−Gθi))αi)),B′1=1−Gθi2−Gθi,C′1=αi, |
A′2=ψ2(ϕ2((Gθj(2−Gθj))αj))ψ′2(ϕ2((Gθj(2−Gθj))αj)),B′2=1−Gθj2−Gθj,C′2=αj, |
and
D′=2lnGψ′2(n∑i=1ϕ2((Gθi(2−Gθi))αi))>0. |
Note that A′2≤A′1<0,0<B′1≤B′2,0<C′1≤C′2. Then, Eq (3.9) is equivalent to
A′1B′1C′1D′−A′2B′2C′2D′={(A′1−A′2)B′1C′1+A′2[(B′1−B′2)C′1+B′2(C′1−C′2)]}D′≥0. |
Thus
∂φ2(α,θ,G,ψ2)∂θi−∂φ2(α,θ,G,ψ2)∂θj≥0. |
Thus φ2(α,θ,G,ψ2) is Schur-concave in θ. Based on αw⪰β, we obtain φ2(α,θ,G,ψ2)≤φ2(α,δ,G,ψ2) by Lemma 3. The desired result is proved.
Remark 3. Theorem 3 manifests that the less positive dependence, and the more heterogeneous scale parameters (in the weakly supermajorized order) lead to the stochastically large lifetime of the parallel system and higher reliability could be achieved. It should be noted that Theorem 3 generalizes Theorem 2 in Chanchal et al. [32] to the case of the dependent components. It might be of great interest to establish the (reversed) hazard rate order between Xn:n and Yn:n of Theorem 3, which is left as an open problem.
The following example 3 illustrates the theoretical result of Theorem 3.
Example 3. For F(x)=e−1x and G(x)=1−e−x,x>0. It is obvious that X≥stY. Let ψ1(x)=e1−(1+x)5 and ψ2(x)=0.2ex−0.8,x>0, and Xi∼TL-G(αi,θi,F) and Yi∼TL-G(βi,δi,G),i=1,2. Further set (θ1,θ2)=(0.2,0.6)w⪰(0.4,0.5)=(δ1,δ2), (α1,α2)=(2,3)=(β1,β2). It is easy to show that ψ2 is log-convex but ψ1 is log-concave, and ϕ2∘ψ1(x) is convex in x and ϕ2∘ψ1(0)=0, which implies ϕ2∘ψ1(x) is super-additive. Then, the conditions of Theorem 3 are satisfied. Figure 3 plots these survival functions of (X2:2+1)−1 and (Y2:2+1)−1, from which it can be observed that ˉH(X2:2+1)−1(x) is always smaller than ˉH(Y2:2+1)−1(x), and this confirms that X2:2≥stY2:2.
The theorem given below provides sufficient conditions for the usual stochastic order, here, we assume that the shape parameter vectors are the same.
Theorem 4. Suppose that X and Y are two random variables with distribution functions F and G, respectively. Define X∼TL-G(α,θ,F) with generator ψ1 and Y∼TL-G(α,δ,G) with generator ψ2, such that θ,δ and α∈ϵ+. If X≤stY and ϕ2∘ψ1 is super-additive, Then, θ≤δ⇒X1:n≤stY1:n.
Proof. Similar to the proof of Theorem 2, we only need to prove that
ϑ2(α,θ,G,ψ2)≤ϑ2(α,δ,G,ψ2). |
Denote
ϑ2(α,θ,G,ψ2)=ψ2(n∑i=1ϕ2(1−(Gθi(2−Gθi))αi)). | (3.10) |
On differentiating (3.10) with respect to θi (i=1,…,n) give rises to
∂ϑ2(α,θ,G,ψ2)∂θi=−ψ′2(n∑i=1ϕ2(1−(Gθi(2−Gθi))αi))ψ2(ϕ2(1−(Gθi(2−Gθi))αi))ψ′2(ϕ2(1−(Gθi(2−Gθi))αi))×2αilnG(1−Gθi2−Gθi)(Gθi(2−Gθi))αi1−(Gθi(2−Gθi))αi≥0. |
Thus, ϑ2(α,θ,G,ψ2) is increasing in θi (i=1,…,n). When θ≤δ, we obtain that ϑ2(α,θ,G,ψ2)≤ϑ2(α,δ,G,ψ2), hence the proof is finished.
Remark 4. It should be mentioned that the condition ϕ2∘ψ1(x) is super-additive in Theorem 4 is quite general and easy to be constructed for many well-known Archimedean copulas. For example, consider Ali-Mikhail-Haq(AMH) copula with generator ψ(x)=1−aex−a for a∈[0,1), it is easy to see that lnψ(x)=ln(1−a)−ln(ex−a) is convex in x∈[0,1]. Let ψ1(x)=1−a1ex−a1 and ψ2(x)=1−a2ex−a2. It can be observed that ϕ2∘ψ1(x)=ln[1−a21−a1(ex−a1)+a2]. Taking derivative of ϕ2∘ψ1(x) twice with respect to x, it can be seen that [ϕ2∘ψ1(x)]″≥0 for 1>a2>a1≥0, which implies the super-additivity of ϕ2∘ψ1(x).
It is natural to ask whether the condition in Theorem 4 can be weakened? The following numerical example provides a negative answer. We show that if we take θm⪯δ, then the result in Theorem 4 does not hold.
Counterexample 1. Let Xi∼TL-G(αi,θi,F) and Yi∼TL-G(βi,δi,G),i=1,2. For F(x)=1−11+x and G(x)=e−1x,x>0. It is obvious that X≤stY. Further set (θ1,θ2)=(0.4,0.6)m⪯(0.1,0.9)=(δ1,δ2), (α1,α2)=(2,3)=(β1,β2). Suppose we choose the Ali-Mikhail-Haq(AMH) copula with parameters a1=0.2,a2=0.8. We plot the graphs of the distribution functions of (X1:2+1)−1 and (Y1:2+1)−1 in Figure 4. The graphs cross each other. This implies that the usual stochastic ordering as mentioned in Theorem 4 does not hold.
Next, we switch our focus to the reversed hazard rate order between two parallel systems having dependent Topp-Leone generated distributed components. The following theorem presents sufficient conditions for the comparison of two parallel systems with the same scale parameters.
Theorem 5. Let X∼TL-G(α,θ,F) and Y∼TL-G(β,θ,F) have the common generator ψ, such that α,β and θ∈ϵ+. If ψ is log-concave, and (ψlnψ)/ψ′ is increasing and concave. Then αw⪯β⇒Xn:n≥rhYn:n.
Proof. The reversed hazard rate function of parallel system Xn:n is
˜γXn:n(x)=ψ′(∑ni=1ϕ((Fθ(2−Fθ))αi))ψ(∑ni=1ϕ((Fθ(2−Fθ))αi))(n∑i=11ψ′(ϕ((Fθ(2−Fθ))αi))αi(Fθ(2−Fθ))αi−1(2θFθ−1f(1−Fθ)))=ψ′(∑ni=1ϕ((Fθ(2−Fθ))αi))ψ(∑ni=1ϕ((Fθ(2−Fθ))αi))(n∑i=1ψ(ϕ((Fθ(2−Fθ))αi))ψ′(ϕ((Fθ(2−Fθ))αi))2θαi˜γX(1−Fθ2−Fθ))=2θ(1−Fθ2−Fθ)(˜γXln(Fθ(2−Fθ)))ψ′(∑ni=1ϕ((Fθ(2−Fθ))αi))ψ(∑ni=1ϕ((Fθ(2−Fθ))αi))×(n∑i=1ψ(ϕ((Fθ(2−Fθ))αi))ln(ψ(ϕ((Fθ(2−Fθ))αi)))ψ′(ϕ((Fθ(2−Fθ))αi))). |
Denote
L(u)=ψ′(∑ni=1ui)ψ(∑ni=1ui)(n∑i=1ψ(ui)lnψ(ui)ψ′(ui)), |
then ˜γXn:n(x) can be represented as
˜γXn:n(x)=2θ1−Fθ2−Fθ(˜γXln(Fθ(2−Fθ)))L(ϕ((Fθ(2−Fθ))α1),…,ϕ((Fθ(2−Fθ))αn)). |
Similarly, we have
˜γYn:n(x)=2θ1−Fθ2−Fθ(˜γXln(Fθ(2−Fθ)))L(ϕ((Fθ(2−Fθ))β1),…,ϕ((Fθ(2−Fθ))βn)). |
Notice that for θ>0,Fθ(2−Fθ)∈[0,1]. Since ψ is log-concave, αw⪯β, and note that (ψlnψ)/ψ′ is increasing concave. It follows from Lemma 5 that
L(ϕ((Fθ(2−Fθ))α1),…,ϕ((Fθ(2−Fθ))αn))≤L(ϕ((Fθ(2−Fθ))β1),…,ϕ((Fθ(2−Fθ))βn)). |
˜γXn:n(x)≥˜γYn:n(x) holds immediately from (1−Fθ)/(2−Fθ)≥0 and ln(Fθ(2−Fθ))≤0. The proof is completed.
Remark 5. Theorem 5 states that less heterogeneous shape parameters in the weakly supermajorized order lead to a large lifetime of the parallel system in the sense of the reversed hazard rate order. It might be of great interest to study the likelihood ratio order between Xn:n and Yn:n, which is left as an open problem.
The following example 4 illustrates that the theoretical results of Theorem 5.
Example 4. Let ψ(x)=e4−4ex,x>0. It is obvious that ψ(x) is log-concave in x>0, and (ψ(x)lnψ(x))/ψ′(x)=(ex−1)/ex is increasing concave. Assume Xi∼TL-G(αi,θi,F) and Yi∼TL-G(βi,θi,F),i=1,2, F(x)=[1−e−x]0.5,x>0. Further set (α1,α2)=(4,5)w⪯(2,6)=(β1,β2), (θ1,θ2)=(2,3). The difference between the hazard rate functions of (X2:2+1)−1 and (Y2:2+1)−1 is plotted in Figure 5, from which one can observe that h(X2:2+1)−1(x) is always larger than h(Y2:2+1)−1(x), and this confirms that X2:2≥rhY2:2.
In this paper, we study stochastic comparisons of series and parallel systems with heterogeneous Topp-Leone generated components with Archimedean (survival) copulas. We established the usual stochastic order of the series and parallel systems, and the reversed hazard rate order of the parallel system. These results generalize some existing results in the literature (see Chanchal et al. [32]) to the case of dependent components in the TL-G family.
The authors thank the editor and two anonymous reviewers for their insightful and valuable comments, which significantly helped in improving the presentation of this paper. This research was supported by the National Natural Science Foundation of China (11861058, 71471148).
The authors declare no conflict of interest.
[1] | N. Balakrishnan, C. R. Rao, Order statistics: applications, Elsevier Science, 1998. |
[2] |
B. Khaledi, S. C. Kochar, Some new results on stochastic comparisons of parallel systems, J. Appl. Probab., 37 (2000), 1123–1128. doi: 10.1239/jap/1014843091
![]() |
[3] |
P. Zhao, N. Balakrishnan, New results on comparisons of parallel systems with heterogeneous gamma components, Stat. Probab. Lett., 81 (2011), 36–44. doi: 10.1016/j.spl.2010.09.016
![]() |
[4] |
L. X. Fang, X. S. Zhang, New results on stochastic comparison of order statistics from heterogeneous Weibull populations, J. Korean Stat. Soc., 41 (2012), 13–16. doi: 10.1016/j.jkss.2011.05.004
![]() |
[5] | E. Barlow, F. Proschan, Statistical theory of reliability and life testing: probability models, Holt, Rinehart & Winston of Canada Ltd, 1975. |
[6] |
J. Bartoszewicz, Dispersive ordering and the total time on test transformation, Stat. Probab. Lett., 4 (1986), 285–288. doi: 10.1016/0167-7152(86)90045-3
![]() |
[7] | P. J. Boland, T. Hu, M. Shaked, J. G. Shanthikumar, Stochastic ordering of order statistics Ⅱ, Springer, 2002. |
[8] |
A. Kundu, S. Chowdhury, A. K. Nanda, N. K. Hazra, Some results on majorization and their applications, J. Comput. Appl. Math., 301 (2016), 161–177. doi: 10.1016/j.cam.2016.01.015
![]() |
[9] |
N. Balakrishnan, P. Nanda, S. Kayal, Ordering of series and parallel systems comprising heterogeneous generalized modified Weibull components, Appl. Stoch. Models. Bus. Ind., 34 (2018), 816–834. doi: 10.1002/asmb.2353
![]() |
[10] | S. Kochar, Stochastic comparisons of order statistics and spacings: a review, ISRN Probability and Statistics, 2012 (2012), 222–229. |
[11] |
N. Balakrishnan, P. Zhao, Ordering properties of order statistics from heterogeneous populations: a review with an emphasis on some recent developments, Probab. Eng. Inform. Sci., 27 (2013), 403–443. doi: 10.1017/S0269964813000156
![]() |
[12] |
R. L. Dykstra, S. C. Kochar, J. Rojo, Stochastic comparisons of parallel systems of heterogeneous exponential components, J. Stat. Plan. Infer., 65 (1997), 203–211. doi: 10.1016/S0378-3758(97)00058-X
![]() |
[13] |
N. Balakrishnan, G. Barmalzan, A. Haidari, On usual multivariate stochastic ordering of order statistics from heterogeneous beta variables, J. Multivar. Anal., 127 (2014), 147–150. doi: 10.1016/j.jmva.2014.02.008
![]() |
[14] |
N. Torrado, On magnitude orderings between smallest order statistics from heterogeneous beta distributions, J. Math. Anal. Appl., 426 (2015), 824–838. doi: 10.1016/j.jmaa.2015.02.003
![]() |
[15] |
J. Navarro, F. Spizzichino, Comparisons of series and parallel systems with components sharing the same copula, Appl. Stoch. Models. Bus. Ind., 26 (2010), 775–791. doi: 10.1002/asmb.819
![]() |
[16] |
C. Li, X. H. Li, Likelihood ratio order of sample minimum from heterogeneous Weibull random variables, Stat. Probab. Lett., 97 (2015), 46–53. doi: 10.1016/j.spl.2014.10.019
![]() |
[17] | C. Li, R. Fang, X. H. Li, Stochastic comparisons of order statistics from scaled and interdependent random variables, Metrika, 79 (2015), 553–578. |
[18] |
R. Fang, C. Li, X. H. Li, Stochastic comparisons on sample extremes of dependent and heterogenous observations, Statistics, 50 (2016), 930–955. doi: 10.1080/02331888.2015.1119151
![]() |
[19] | A. Kundu, S. Chowdhury, Stochastic comparisons of lifetimes of two series and parallel systems with location-scale family distributed components having archimedean copulas, arXiv: 1710.00769, 2017. |
[20] |
R. Fang, C. Li, X. H. Li, Ordering results on extremes of scaled random variables with dependence and proportional hazards, Statistics, 52 (2018), 458–478. doi: 10.1080/02331888.2018.1425998
![]() |
[21] |
C. Li, X. H. Li, Hazard rate and reversed hazard rate orders on extremes of heterogeneous and dependent random variables, Stat. Probab. Lett., 146 (2019), 104–111. doi: 10.1016/j.spl.2018.11.005
![]() |
[22] |
S. Rezaei, B. B. Sadr, M. Alizadeh, S. Nadarajah, Topp-Leone generated family of distributions: Properties and applications, Commun. Stat. Theor. M., 46 (2017), 2893–2909. doi: 10.1080/03610926.2015.1053935
![]() |
[23] |
R. F. Yan, T. Q. Luo, On the optimal allocation of active redundancies in series system, Commun. Stat. Theor. M., 47 (2018), 2379–2388. doi: 10.1080/03610926.2015.1054942
![]() |
[24] |
E. J. Veres-Ferrer, J. M. Pavia, On the relationship between the reversed hazard rate and elasticity, Stat. Papers, 55 (2014), 275–284. doi: 10.1007/s00362-012-0470-1
![]() |
[25] | M. Shaked, G. Shanthikumar, Stochastic orders, New York: Springer, 2007. |
[26] | T. Lando, L. Bertoli-Barsotti, Stochastic dominance relations for generalised parametric distributions obtained through compositio, METRON, 78 (2020), 297–311. |
[27] | T. Lando, L. Bertoli-Barsotti, Second-order stochastic dominance for decomposable multiparametric families with applications to order statistics, Stat. Probab. Lett., 159 (2020), 108691. |
[28] | A. W. Marshall, I. Olkin, B. C. Arnold, Inequalities: Theory of majorization and its applications, 2 Eds., New York: Springer, 2011. |
[29] | R. B. Nelsen, An introduction to copulas, New York: Springer, 2006. |
[30] |
X. H. Li, R. Fang, Ordering properties of order statistics from random variables of Archimedean copulas with applications, J. Multivar. Anal., 133 (2015), 304–320. doi: 10.1016/j.jmva.2014.09.016
![]() |
[31] |
S. Das, S. Kayal, Ordering extremes of exponentiated location-scale models with dependent and heterogeneous random samples, Metrika, 83 (2020), 869–893. doi: 10.1007/s00184-019-00753-2
![]() |
[32] | R. Chanchal, V. Gupta, A. K. Misra, Stochastic comparisons of series and parallel systems with Topp-Leone generated family of distributions, arXiv: 1908.05896, 2019. |
1. | Jing Zhang, Aoshuang Li, Rui Fang, Stochastic comparisons on extremes of burr type XII samples associated with Archimedean copula and heterogeneous shape parameters, 2022, 0361-0926, 1, 10.1080/03610926.2022.2092146 | |
2. | Xiao Zhang, Rongfang Yan, Stochastic comparisons of extreme order statistic from dependent and heterogeneous lower-truncated Weibull variables under Archimedean copula, 2022, 7, 2473-6988, 6852, 10.3934/math.2022381 | |
3. | Ernesto-Jesús Veres-Ferrer, Jose M. Pavía, The Elasticity of a Random Variable as a Tool for Measuring and Assessing Risks, 2022, 10, 2227-9091, 68, 10.3390/risks10030068 | |
4. | Liang Jiao, Rongfang Yan, Stochastic Comparisons of Lifetimes of Series and Parallel Systems with Dependent Heterogeneous MOTL-G Components under Random Shocks, 2021, 13, 2073-8994, 2248, 10.3390/sym13122248 |