Limit properties for ratios of order statistics from exponentials
Zhang and Ding Journal of Inequalities and Applications
Limit properties for ratios of order statistics from exponentials
Yong Zhang
In this paper, we study the limit properties of the ratio for order statistics based on samples from an exponential distribution and obtain the expression of the density functions, the existence of the moments, the strong law of large numbers for Rnij with 1 ≤ i < j < mn = m. We also discuss other limit theorems such as the central limit theorem, the law of iterated logarithm, the moderate deviation principle, the almost sure central limit theorem for selfnormalized sums of Rnij with 2 ≤ i < j < mn = m.
exponential distribution; order statistics; strong law of large numbers; central limit theorem; law of iterated logarithm

As we know, the exponential distribution can describe the lifetimes of the equipment, and
the ratios Rnij can measure the stability of equipment, it shows whether or not our system
Rnj =
l
(j – )!(m – j)!(α + ) l= Cj– (m – l – )
© The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License
(http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium,
provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and
indicate if changes were made.
Later on, Miao et al. [] proved the central limit theorem and the almost sure central
limit for Rn with fixed sample size, we state their results as the following theorem.
(Rn – ERn) →D N (, ) as N → ∞,
(Rn – ERn) ≤ x = (x) a.s.
In this paper, we will make a further study on the limit properties of Rnij. In the next
section, firstly, we give the expression of the density functions of Rnij for all ≤ i < j < mn,
it is more interesting that the density function is free of the sample mean λn, this allows
us to change the equipment from sample to sample as long as the underlying distribution
remains an exponential. Also we discuss the existence of the moments for fixed sample size
mn = m. Secondly, we establish the strong law of large number for Rnij with = i < j < m
and ≤ i < j < m, respectively. At last we give some limit theorems such as the central limit
theorem, the law of iterated logarithm, the moderate deviation principle, the almost sure
central limit theorem for selfnormalized sums of Rnij with ≤ i < j < m.
In the following, C denotes a positive constant, which may take different values
whenever it appears in different expressions. an ∼ bn means that an/bn → as n → ∞.
2 Main results and proofs
2.1 Density functions and moments of Rnij
The first theorem gives the expression of the density functions.
Proof It is easy to check that the joint density function of Xn(i) and Xn(j) is
Therefore the density function of Rnij is
fnij(r) =
f (w, r) dw
mn!
(i – )!(j – i – )!(mn – j)! λn k= l=
∞ we–(i–k+l)w/λn e–(mn–i–l)rw/λn dw
∞ te–[(i–k+l)+r(mn–i–l)]tdt
(–)j–k–l–Cik–Cjl–i–
(–)j–k–l–Cik–Cjl–i–
and with ≤ i < j ≤ m,
fnj(r) =
m!
(j – )!(m – j)! l= (–)j–l–Cjl– [ + l + r(m – l – )]
as r → ∞,
Remark . Miao et al. [] obtained the density function for Rnj for fixed sample size
mn = m, they also proved that the expectation of Rnj is finite and the truncated second
moment is slowly varying at ∞. Adler [] also claimed that all the Rnj have infinite
expectations for fixed sample size, so our theorems extended their results.
2.2 Strong law of large numbers of Rnij
From our assumptions, we know that {Rnij, n ≥ } is an independent sequence with the
same distribution for fixed sample size mn = m. As Theorem . states that the Rnj do not
have the expectation, so the strong law of large numbers with them is not typical. Here
we give the weighted strong law of large number as follows. At first, we list the following
lemma, that is, Theorem . from De la Peña et al. [], which will be used in the proof.
Lemma . Let {Xn, n ≥ } be a sequence of independent random variables, denote Sn =
in= Xi, if bn ∞, and i∞= Var(Xi)/bi < ∞, then (Sn – ESn)/bn → a.s.
Theorem . Let {an, n ≥ } be a sequence of positive real numbers and {bn, n ≥ } be a
sequence of nondecreasing positive real numbers with limn→∞ bn = ∞ and
< ∞,
∞ an
n= bn
bN n=
anRnj =
an RnjI{ ≤ Rnj ≤ cn} – ERnjI{ ≤ Rnj ≤ cn}
anRnj =
λm! l (–)j–l–
(j – )!(m – j)! l= Cj– (m – l – )
For mn → ∞,
bN n=
bN n=
bN n=
= I + I + I.
anRnjI{Rnj > cn}
anERnjI{ ≤ Rnj ≤ cn}
By (.) and (.), it is easy to show
∞ j–
n= cn
∞ an
n= bn
< ∞.
Then by the BorelCantelli lemma, we get
RnjI{Rnj > cn} → a.s. n → ∞.
By (.) and (.), we can obtain
Therefore combining (.) with (.), we can easily conclude
I → a.s. n → ∞.
ERnjI{ ≤ Rnj ≤ cn}
For I, by (.) and noting cn → ∞, we get
dr ≤ C
n= cn
∞ an
n= bn
< ∞,
≤ C
≤ C
then by Lemma ., we have
I → a.s. n → ∞.
P{Rnj > cn} =
then combining with (.), we show
I → (j – λ)!m(m! – j)! l= Cjl– (m(––)lj––l–) , n → ∞.
So the proof of (.) is completed by combining (.), (.), (.), and (.).
By the same argument as in the proof of (.), we can get (.), so we omit it here.
(Rnij – ERnij) =
2.3 Other limit properties for Rnij, 2 ≤ i < j ≤ m
By the above discussion, we know that, for fixed sample size mn = m and ≤ i < j ≤ m,
{Rnij, n ≥ } is a sequence of independent and identically distributed random variables with
finite mean, and L(r) = E(Rnij – ERnij)I{Rnij – ERnij ≤ r} is a slowly varying function at ∞.
Therefore the limit properties of Rnij for fixed sample size can easily be established by those
of the selfnormalized sums. We list some of them, such as the central limit theorem (CLT),
the law of iterated logarithm (LIL), the moderate deviation principle (MDP), the almost
sure central limit theorem (ASCLT). Denote SN = nN=(Rnij – ERnij), VN = nN=(Rnij –
ERnij).
Proof By Theorem . from Giné et al. [], we can obtain the CLT for Rnij.
N→∞
VN
√ log log N
=
Proof By Theorem from Griffin and Kuelbs [], the LIL for Rnij holds.
Theorem . (MDP) Let {xn, n ≥ } be a sequence of positive numbers with xn → ∞ and
√
xn = o( n), as n → ∞, then, for fixed sample size mn = m, we conclude
≥ xN
Proof By Theorem . from Shao [], we can prove the MDP for Rnij.
n
k= dk . Then, for fixed sample size mn = m and any x ∈ R,
k→∞ Dk N=
≤ x = (x)
(·) is the distribution function of the standard normal random variable.
Proof By Corollary from Zhang [], we know ASCLT for Rnij holds.
p
Remark . It is easy to check that ηN /VN → , then by the Slutsky lemma and
Theorem ., we can get Theorem . from Miao et al. [].
Competing interests
The authors declares that they have no competing interests.
Authors’ contributions
Both authors contributed equally and significantly in writing this article. Both authors read and approved the final
manuscript.
Acknowledgements
This work was supported by National Natural Science Foundation of China (Grant Nos. 11101180, 11201175); the Science
and Technology Development Program of Jilin Province (Grant Nos. 20130522096JH, 20140520056JH).
1. Adler , A: Strong laws for ratios of order statistics from exponentials . Bull. Inst. Math. Acad. Sin . (N.S.) 10 ( 1 ), 101  111 ( 2015 )
2. Miao , Y, Wang , RJ, Adler , A: Limit theorems for order statistics from exponentials . Stat. Probab. Lett . 110 , 51  57 ( 2016 )
3. De la Peña , VH, Tai, TL, Shao, QM: SelfNormalized Process: Limit Theory and Statistical Applications . Springer, New York ( 2009 )
4. Adler , A: Exact strong laws . Bull. Inst. Math. Acad. Sin . 28 ( 3 ), 141  166 ( 2000 )
5. Feller , W: An Introduction to Probability Theory and Its Applications , vol. 1 , 3rd edn. Wiley, New York ( 1968 )
6. Giné , E, Götze, F, Mason, DM: When is the Student tstatistic asymptotically standard normal? Ann . Probab. 25 , 1514  1531 ( 1997 )
7. Griffin , PS, Kuelbs, JD: Selfnormalized laws of the iterated logarithm . Ann. Probab . 17 , 1571  1601 ( 1989 )
8. Shao , QM: Selfnormalized large deviations . Ann. Probab . 25 , 285  328 ( 1997 )
9. Zhang , Y: A general result on almost sure central limit theorem for selfnormalized sums for mixing sequences . Lith. Math. J . 53 ( 4 ), 471  483 ( 2013 )