Volume 2010, Article ID 769201,12pages doi:10.1155/2010/769201
Research Article
On Complete Convergence for Arrays of Rowwise
ρ
-Mixing Random Variables and Its Applications
Xing-cai Zhou
1, 2and Jin-guan Lin
11Department of Mathematics, Southeast University, Nanjing 210096, China
2Department of Mathematics and Computer Science, Tongling University, Tongling, Anhui 244000, China
Correspondence should be addressed to Jin-guan Lin,jglin@seu.edu.cn
Received 15 May 2010; Revised 23 August 2010; Accepted 21 October 2010
Academic Editor: Soo Hak Sung
Copyrightq2010 X.-c. Zhou and J.-g. Lin. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
We give out a general method to prove the complete convergence for arrays of rowwiseρ-mixing random variables and to present some results on complete convergence under some suitable conditions. Some results generalize previous known results for rowwise independent random variables.
1. Introduction
Let{Ω,F, P}be a probability space, and let{Xn;n≥ 1}be a sequence of random variables defined on this space.
Definition 1.1. The sequence{Xn;n≥1}is said to beρ-mixing if
ρn sup k≥1
sup X∈L2Fk1, Y∈L2F∞
nk
⎧ ⎪ ⎨ ⎪ ⎩
|EXY−EXEY|
EX−EX2EY−EY2
⎫ ⎪ ⎬ ⎪
⎭−→0 1.1
asn → ∞, whereFnmdenotes theσ-field generated by{Xi;m≤i≤n}.
Theρ-mixing random variables were first introduced by Kolmogorov and Rozanov
1. The limiting behavior ofρ-mixing random variables is very rich, for example, these in the study by Ibragimov2, Peligrad3, and Bradley4for central limit theorem; Peligrad
9for almost sure invariance principle; Peligrad10, Shao11and Liang and Yang12for convergence rate; Shao11, for the maximal inequality, and so forth.
For arrays of rowwise independent random variables, complete convergence has been extensively investigated see, e.g., Hu et al. 13, Sung et al. 14, and Kruglov et al. 15. Recently, complete convergence for arrays of rowwise dependent random variables has been considered. We refer to Kuczmaszewska16forρ-mixing andρ-mixing sequences, Kuczmaszewska17for negatively associated sequence, and Baek and Park18 for negatively dependent sequence. In the paper, we study the complete convergence for arrays of rowwiseρ-mixing sequence under some suitable conditions using the techniques of Kuczmaszewska 16,17. We consider the case of complete convergence of maximum weighted sums, which is different from Kuczmaszewska16. Some results also generalize some previous known results for rowwise independent random variables.
Now, we present a few definitions needed in the coming part of this paper.
Definition 1.2. An array{Xni;i ≥ 1, n ≥ 1}of random variables is said to be stochastically dominated by a random variableXif there exists a constantC, such that
P{|Xni|> x} ≤CP{C|X|> x} 1.2
for allx≥0,i≥1 andn≥1.
Definition 1.3. A real-valued functionlx, positive and measurable onA,∞for someA >0, is said to be slowly varying if
lim x→ ∞
lλx
lx 1 for eachλ >0. 1.3
Throughout the sequel,C will represent a positive constant although its value may change from one appearance to the next;xindicates the maximum integer not larger than
x;IBdenotes the indicator function of the setB. The following lemmas will be useful in our study.
Lemma 1.4Shao11. Let{Xn;n≥1}be a sequence ofρ-mixing random variables withEXi0
andE|Xi|q <∞for someq≥2. Then there exists a positive constantK Kq, ρ·depending only
onqandρ·such that for anyn≥1
Emax
1≤i≤n
i
j1 Xj
q
≤K
⎛
⎝n2/qexp
⎛ ⎝Klogn
i0 ρ2i
⎞ ⎠max
1≤i≤n
E|Xi|2
q/2
nexp
⎛ ⎝Klogn
i0
ρ2/q2i
⎞ ⎠max
1≤i≤nE|Xi| q
⎞ ⎠.
1.4
Lemma 1.5Sung19. Let{Xn;n≥1}be a sequence of random variables which is stochastically dominated by a random variableX. For anyα >0andb >0, the following statement holds:
Lemma 1.6Zhou20. Iflx>0is a slowly varying function asx → ∞, then
imn1nsln≤Cms1lmfors >−1,
ii∞nmnsln≤Cms1lmfors <−1.
This paper is organized as follows. InSection 2, we give the main result and its proof. A few applications of the main result are provided inSection 3.
2. Main Result and Its Proof
This paper studies arrays of rowwiseρ-mixing sequence. Letρnibe the mixing coefficient defined inDefinition 1.1for thenth row of an array{Xni;i≥1, n≥1}, that is, for the sequence
Xn1, Xn2, . . . , n≥1.
Now, we state our main result.
Theorem 2.1. Let{Xni;i≥ 1, n ≥1}be an array of rowwiseρ-mixing random variables satisfying
supn∞i1ρn2/q2i<∞for someq≥2, and let{ani;i≥1, n ≥1}be an array of real numbers. Let
{bn;n≥1}be an increasing sequence of positive integers, and let{cn;n≥1}be a sequence of positive
real numbers. If for some0< t <2and anyε >0the following conditions are fulfilled:
a∞n1cnbn
i1P{|aniXni| ≥εb 1/t n }<∞,
b∞n1cnbn−q/t1max1≤i≤bn|ani|qE|Xni|qI|aniXni|< εb1n/t<∞,
c∞n1cnbn−q/tq/2max1≤i≤bn|ani|2E|Xni|2I|aniXni|< εb1n/t
q/2 <∞,
then
∞
n1 cnP
⎧ ⎨ ⎩1max≤i≤bn
i
j1
anjXnj−anjEXnjIanjXnj< εb1n/t
> εb1n/t
⎫ ⎬
⎭<∞. 2.1
Remark 2.2. Theorem 2.1extends some results of Kuczmaszewska17to the case of arrays of rowwiseρ-mixing sequence and generalizes the results of Kuczmaszewska16to the case of maximum weighted sums.
Remark 2.3. Theorem 2.1firstly gives the condition of the mixing coefficient, so the conditions
a–c do not contain the mixing coefficient. Thus, the conditions a–c are obviously simpler than the conditionsi–iiiin Theorem 2.1 of Kuczmaszewska16. Our conditions are also different from those of Theorem 2.1 in the study by Kuczmaszewska17:q ≥ 2 is only required inTheorem 2.1, notq >2 in Theorem 2.1 of Kuczmaszewska17; the powers of
bninbandcofTheorem 2.1are−q/t1 and−q/tq/2, respectively, not−q/tin Theorem 2.1 of Kuczmaszewska17.
Proof. The conclusion of the theorem is obvious if∞n1cnis convergent. Therefore, we will consider that only∞n1cnis divergent. Let
YnjanjXnjIanjXnj< εbn1/t
, Tni
i
j1
Ynj, Sni i
j1 anjXnj,
Abn
i1
{aniXniYni}, B bn
i1
{aniXni/Yni}.
2.2
Note that
P
max
1≤i≤bn|Sni−ETni|> εb 1/t n
P
max
1≤i≤bn|Sni−ETni|> εb 1/t
n A
P
max
1≤i≤bn|Sni−ETni|> εb 1/t
n B
≤Pmax
1≤i≤bn|Tni−ETni|> εb 1/t n
bn
i1
P|aniXni|> εb1n/t
.
2.3
Byait is enough to prove that for allε >0
∞
n1 cnP
max
1≤i≤bn|Tni−ETni|> εb 1/t n
<∞. 2.4
By Markov inequality and Lemma 1.4, and note that the assumption supn∞i1ρn2/q2i<∞for someq≥2, we get
P
max
1≤i≤bn|Tni−ETni|> εb 1/t n
≤Cb−nq/tEmax
1≤i≤bn|Tni−ETni|
q
≤Cb−nq/t
⎧ ⎨ ⎩bnexp
⎛
⎝Klogbn
i0 ρ2/q
n
2i
⎞ ⎠max
1≤i≤bnE|aniXni|
q
×I|aniXni|< εbn1/t
Kexp
⎛
⎝Klogbn
i1 ρn
2i
×
bnmax
1≤i≤bnE|aniXni| 2I|a
niXni|< εb1n/t
q/2⎫⎬
⎭
≤Cb−q/t1
n max
1≤i≤bn|ani|
qE|X ni|qI
|aniXni|< εb1n/t
Cb−q/tq/2
n
max
1≤i≤bn|ani| 2E|X
ni|2I|aniXni|< εb1n/t
q/2 .
2.5
Fromb,c, and2.5, we see that2.4holds.
3. Applications
Theorem 3.1. Let{Xni;i≥ 1, n ≥1}be an array of rowwiseρ-mixing random variables satisfying
supn∞i1ρn2/q2i < ∞for some q ≥ 2,EXni 0, andE|Xni|p < ∞ for alln ≥ 1, i ≥ 1, and
1≤p≤2. Let{ani;i≥1, n≥1}be an array of real numbers satisfying the condition
max
1≤i≤n|ani| pE|X
ni|pO
nν−1, as n−→ ∞,
3.1
for some0< ν <2/q. Then for anyε >0andαp≥1
∞
n1 nαp−2P
⎧ ⎨ ⎩max1≤i≤n
i
j1 anjXnj
> εnα
⎫ ⎬
⎭<∞. 3.2
Proof. Putcnnαp−2,bnn, and 1/tαinTheorem 2.1. By3.1, we get
∞
n1 cn
bn
i1
P|aniXni| ≥εb1n/t
≤C∞
n1
nαp−2n
i1 n−αp|a
ni|pE|Xni|p≤C
∞
n1 n−1max
1≤i≤n|ani| pE|X
ni|p≤C
∞
n1
n−2ν<∞,
∞
n1
cnb−nq/t1max
1≤i≤bnE|aniXni|
qI|a
niXni|< εbn1/t
≤C∞
n1
nαp−2n−αq1nαq−pmax 1≤i≤n|ani|
pE|X ni|p≤C
∞
n1
n−2ν<∞,
∞
n1
cnb−nq/tq/2
max
1≤i≤bnE|aniXni| 2I|a
niXni|< εb1n/t
≤C∞
n1
nαp−2n−αqq/2nα2−pq/2
max
1≤i≤n|ani|
pE|X ni|p
q/2
≤C∞
n1
nαp1−q/2νq/2−1−1 <∞
3.3
following fromνq/2−1<0. By the assumptionEXni0 forn≥1,i≥1 and by3.1, we have
n−αmax
1≤i≤n
i
j1
anjEXnjI anjXnj< εnα!
≤Cn−αn j1
anjEXnjI anjXnj< εnα!
≤Cn−αn j1
anjEXnjI anjXnj≥εnα!
≤Cn−αpn j1
anjpEX njp
≤Cn−αp1max 1≤j≤n|anj|
pE|X
nj|p≤Cn−αpν−→0, asn−→ ∞,
3.4
becauseν <1 andαp≥1. Thus, we complete the proof of the theorem.
Theorem 3.2. Let{Xni;i≥ 1, n ≥1}be an array of rowwiseρ-mixing random variables satisfying supn∞i1ρn2/q2i < ∞ for someq ≥ 2, EXni 0, and E|Xni|p < ∞for all n ≥ 1, i ≥ 1, and 1≤p≤2. Let the random variables in each row be stochastically dominated by a random variableX, such thatE|X|p<∞, and let{ani;i≥1, n≥1}be an array of real numbers satisfying the condition
max
1≤i≤n|ani|
pOnν−1, asn−→ ∞,
3.5
for some0< ν <2/q. Then for anyε >0andαp≥13.2holds.
Theorem 3.3. Let{Xni, n ≥1, i ≥1}be an array of rowwiseρ-mixing random variables satisfying supn∞i1ρn2/q2i<∞for someq≥2andEXni0for alln≥1, i≥1. Let the random variables in
each row be stochastically dominated by a random variableX, and let{ani;i≥1, n≥1}be an array of real numbers. If for some0< t <2,ν >1/2
sup i≥1
|ani|O
n1/t−ν, E|X|12/ν<∞,
then for anyε >0
∞
n1 P
⎧ ⎨ ⎩max1≤i≤n
i
j1 anjXnj
> εn1/t
⎫ ⎬
⎭<∞. 3.7
Proof. Takecn 1 andbn nforn ≥ 1. Then we see thataandbare satisfied. Indeed, takingq≥max2,12/ν, byLemma 1.5and3.6, we get
∞
n1 cn
bn
i1
P|aniXni| ≥εbn1/t
∞
n1
n
i1
P|aniXni| ≥εn1/t
≤C∞
n1
n
i1
P{|X| ≥Cnν}
C∞
n1 n∞
kn
PCkν≤ |X|< Ck1ν
≤C∞
k1
k2PCkν≤ |X|< Ck1ν≤CE|X|2/ν<∞,
∞
n1
cnb−nq/t1max
1≤i≤bn|ani|
qE|X ni|qI
|aniXni|< εb1n/t
≤C∞
n1
n−q/t1max 1≤i≤n|ani|
q
"
E|X|qI|aniX|< εn1/t
nq/t
|ani|qP
|aniX| ≥εn1/t
#
≤C∞
n1
n−12/ν/t1max 1≤i≤n|ani|
12/νE|X|12/νC∞ n1
nmax
1≤i≤nP
|aniX| ≥εn1/t
≤C∞
n1
n−12/ν/t1
$
sup i≥1
|ani|
%12/ν
E|X|12/νC∞ n1
nP{|X| ≥Cnν}
≤C∞
n1
n−ν−1E|X|12/νC∞ n1
n−ν−1E|X|12/ν≤C∞ n1
n−ν−1<∞ 3.8
In order to prove thatcholds, we consider the following two cases. Ifν >2, byLemma 1.5,Cr inequality, and3.6, we have
∞
n1
cnb−nq/tq/2
max
1≤i≤bn|ani| 2E|X
ni|2I
|aniXni|< εb1n/t
q/2
≤C∞
n1
n−q/tq/2
max
1≤i≤n|ani|
2E|X|2I|a
niXni|< εn1/t
C∞
n1 nq/2
max
1≤i≤nP
|aniX| ≥εn1/t
q/2
≤C∞
n1
n−q/tq/2n1/t1−2/νq/2max 1≤i≤n|ani|
12/νE|X|12/νq/2
C∞
n1
nq/2P{|X| ≥Cnν}q/2
≤C∞
n1
n−q/tq/21/t1−2/νq/2 $
sup i≥1
|ani|
%12/νq/2
E|X|12/νq/2
C∞
n1
nq/2n−12/ννq/2E|X|12/νq/2
≤C∞
n1
n−ν1q/2E|X|12/νq/2<∞.
3.9
If 1/2 < ν ≤ 2, take q > 2/2ν−1. We have that 2ν−1q/2 > 1. Note that in this case
E|X|2<∞. We have
∞
n1
cnb−nq/tq/2
max
1≤i≤bn|ani| 2E|X
ni|2I
|aniXni|< εb1n/t
q/2
≤C∞
n1
n−q/tq/2
max
1≤i≤n|ani|
2E|X|2I|a
niXni|< εn1/t
q/2
C∞
n1
nq/2P{|X| ≥Cnν}q/2
≤C∞
n1
n−q/tq/2
$
sup i≥1
|ani|
%q
E|X|2q/2C∞
n1
nq/2−νqE|X|2q/2
≤C∞
n1
n−2ν−1q/2E|X|2q/2<∞. 3.10
The proof will be completed if we show that
n−1/tmax
1≤i≤n
i
j1
anjEXnjIanjXni< εn1/t
Indeed, byLemma 1.5, we have
n−1/tmax
1≤i≤n
i
j1
anjEXnjIanjXni< εn1/t
≤Cn−1/t
n
j1
anjE|X|Cn j1
PanjX≥εn1/t
≤Cn−νE|X|CnP{|X| ≥εnν}
≤Cn−νE|X|Cn−ν1E|X|12/ν−→0, asn−→ ∞.
3.12
Theorem 3.4. Let{Xni;i≥1, n≥ 1}be an array of rowwiseρ-mixing random variables satisfying supn∞i1ρn2/q2i<∞for someq≥2, and let{ani;i≥1, n ≥1}be an array of real numbers. Let lx>0be a slowly varying function asx → ∞. If for some0 < t <2and real numberλ, and any ε >0the following conditions are fulfilled:
A∞n1nλlnin1P{|aniXni| ≥εn1/t}<∞,
B∞n1nλ−q/t1lnmax
1≤i≤nE|aniXni|qI|aniXni|< εn1/t<∞,
C∞n1nλ−q/tq/2lnmax
1≤i≤n|ani|2E|Xni|2I|aniXni|< εn1/tq/2 <∞,
then
∞
n1
nλlnP
⎧ ⎨ ⎩max1≤i≤n
i
j1
anjXnj−anjEXnjIanjXnj< εn1/t
> εn1/t
⎫ ⎬
⎭<∞. 3.13
Proof. Letcnnλlnandbnn. UsingTheorem 2.1, we obtain3.13easily.
Theorem 3.5. Let{Xni;i≥1, n≥1}be an array of rowwiseρ-mixing identically distributed random
variables satisfying∞i1ρ2n/q2i<∞for someq≥2andEX110. Letlx>0be a slowly varying
function asx → ∞. If forα >1/2,αp >1, and0< t <2
E|X11|αptl
|X11|t
<∞, 3.14
then
∞
n1
nαp−2lnP
⎧ ⎨ ⎩max1≤i≤n
i
j1 Xnj
> εn1/t
⎫ ⎬
⎭<∞. 3.15
ByLemma 1.6, we obtain
∞
n1
nαp−1lnP|X
11|> εn1/t
∞
n1
nαp−1ln∞
mnP
εm1/t<|X
11| ≤εm11/t
≤C∞
m1
Pεm1/t<|X
11| ≤εm11/t
m
n1
nαp−1ln
≤C∞
m1
mαplmPεm1/t<|X
11| ≤εm11/t
≤CE|X11|αptl
|X11|t
<∞,
3.16
which proves that conditionAis satisfied.
Takingq >max2, αpt, we haveαp−q/t <0. ByLemma 1.6, we have
∞
n1
nαp−1−q/tlnE|X 11|qI
|X11| ≤εn1/t
∞
n1
nαp−1−q/tlnn
m1
E|X11|qI
εm−11/t≤ |X11|< εm1/t
≤C∞
m1
E|X11|qI
εm−11/t≤ |X11|< εm1/t
∞
nmn
αp−1−q/tln
≤C∞
m1
mαp−q/tlmE|X 11|qI
εm−11/t≤ |X11|< εm1/t
≤CE|X11|αptl
|X11|t
<∞,
3.17
which proves thatBholds.
In order to prove thatCholds, we consider the following two cases. Ifαpt <2, takeq >2. We have
∞
n1
nαp−2−q/tq/2lnE|X 11|2I
|X11|< εn1/t
q/2
≤C∞
n1
nαp−2−q/tq/2lnnq/t−αpq/2E|X 11|αptI
|X11|< εn1/t
q/2
≤C∞
n1
nαp−11−q/2−1ln<∞.
Ifαpt≥2, takeq >max2,2tαp−1/2−t. We haveαp−q/tq/2<1. Note that in this caseE|X11|2 <∞. We obtain
∞
n1
nαp−2−q/tq/2lnE|X 11|2I
|X11|< εn1/t
q/2
≤C∞
n1
nαp−2−q/tq/2ln<∞. 3.19
The proof will be completed if we show that
n−1/tmax
1≤i≤n
i
j1 EXnjI
|X11|< εn1/t
−→0, asn−→ ∞. 3.20
Ifαpt <1, then
n−1/tmax
1≤i≤n
i
j1 EXnjI
|X11|< εn1/t
≤Cn1−αpE|X11|αpt−→0, asn−→ ∞. 3.21
Ifαpt≥1, note thatEX110, then
n−1/tmax
1≤i≤n
i
j1 EXnjI
|X11|< εn1/t
≤n−1/t1EX 11I
|X11| ≥εn1/t≤Cn1−αpE|X11|αpt−→0, asn−→ ∞.
3.22
We complete the proof of the theorem.
Noting that for typical slowly varying functions,lx 1 andlx logx, we can get the simpler formulas in the above theorems.
Acknowledgments
The authors thank the academic editor and the reviewers for comments that greatly improved the paper. This work is partially supported by the Anhui Province College Excellent Young Talents Fund Project of Chinano. 2009SQRZ176ZD and National Natural Science Foundation of Chinanos. 11001052, 10871001, 10971097.
References
1 A. N. Kolmogorov and G. Rozanov, “On the strong mixing conditions of a stationary Gaussian process,”Theory of Probability and Its Applications, vol. 2, pp. 222–227, 1960.
2 I. A. Ibragimov, “A note on the central limit theorem for dependent random variables,”Theory of Probability and Its Applications, vol. 20, pp. 134–139, 1975.
4 R. C. Bradley, “A central limit theorem for stationaryρ-mixing sequences with infinite variance,”The Annals of Probability, vol. 16, no. 1, pp. 313–332, 1988.
5 M. Peligrad, “Invariance principles for mixing sequences of random variables,” The Annals of Probability, vol. 10, no. 4, pp. 968–981, 1982.
6 Q. M. Shao, “A remark on the invariance principle forρ-mixing sequences of random variables,”
Chinese Annals of Mathematics Series A, vol. 9, no. 4, pp. 409–412, 1988.
7 Q. M. Shao, “On the invariance principle forρ-mixing sequences of random variables,”Chinese Annals of Mathematics Series B, vol. 10, no. 4, pp. 427–433, 1989.
8 Q. M. Shao, “Complete convergence ofρ-mixing sequences,”Acta Mathematica Sinica, vol. 32, no. 3, pp. 377–393, 1989.
9 Q. M. Shao, “Almost sure invariance principles for mixing sequences of random variables,”Stochastic Processes and Their Applications, vol. 48, no. 2, pp. 319–334, 1993.
10 M. Peligrad, “Convergence rates of the strong law for stationary mixing sequences,”Zeitschrift f ¨ur Wahrscheinlichkeitstheorie und Verwandte Gebiete, vol. 70, no. 2, pp. 307–314, 1985.
11 Q. M. Shao, “Maximal inequalities for partial sums ofρ-mixing sequences,”The Annals of Probability, vol. 23, no. 2, pp. 948–965, 1995.
12 H. Liang and C. Yang, “A note of convergence rates for sums of ρ-mixing sequences,” Acta Mathematicae Applicatae Sinica, vol. 15, no. 2, pp. 172–177, 1999.
13 T.-C. Hu, M. Ord ´o ˜nez Cabrera, S. H. Sung, and A. Volodin, “Complete convergence for arrays of rowwise independent random variables,”Communications of the Korean Mathematical Society, vol. 18, no. 2, pp. 375–383, 2003.
14 S. H. Sung, A. I. Volodin, and T.-C. Hu, “More on complete convergence for arrays,”Statistics & Probability Letters, vol. 71, no. 4, pp. 303–311, 2005.
15 V. M. Kruglov, A. I. Volodin, and T.-C. Hu, “On complete convergence for arrays,” Statistics & Probability Letters, vol. 76, no. 15, pp. 1631–1640, 2006.
16 A. Kuczmaszewska, “On complete convergence for arrays of rowwise dependent random variables,”
Statistics & Probability Letters, vol. 77, no. 11, pp. 1050–1060, 2007.
17 A. Kuczmaszewska, “On complete convergence for arrays of rowwise negatively associated random variables,”Statistics & Probability Letters, vol. 79, no. 1, pp. 116–124, 2009.
18 J.-I. Baek and S.-T. Park, “Convergence of weighted sums for arrays of negatively dependent random variables and its applications,”Journal of Theoretical Probability, vol. 23, no. 2, pp. 362–377, 2010.
19 S. H. Sung, “Complete convergence for weighted sums of random variables,”Statistics & Probability Letters, vol. 77, no. 3, pp. 303–311, 2007.