Hindawi Publishing Corporation Mathematical Problems in Engineering Volume 2015, Article ID 756548, 9 pages http://dx.doi.org/10.1155/2015/756548
Research Article Normal Limiting Distribution of the Size of Binary Interval Trees
Jie Liu1 and Yang Yang2,3
1 Department of Statistics and Finance, University of Science and Technology of China, Hefei 230026, China 2Department of Statistics, Nanjing Audit University, Nanjing 211815, China 3The Key Lab of Financial Engineering of Jiangsu Province, Nanjing Audit University, Nanjing 211815, China
Correspondence should be addressed to Yang Yang; [email protected]
Received 1 August 2015; Revised 3 September 2015; Accepted 9 September 2015
Academic Editor: Xinguang Zhang
Copyright Β© 2015 J. Liu and Y. Yang. 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.
The limiting distribution of the size of binary interval tree is investigated. Our illustration is based on the contraction method, and it is quite different from the case in one-sided binary interval tree. First, we build a distributional recursive equation ofthe size. Then, we draw the expectation, the variance, and some high order moments. Finally, it is shown that the size (with suitable standardization) approaches the standard normal random variable in the Zolotarev metric space.
1. Introduction a one-sided version of a random tree with a digital flavor. Fill et al. [12] followed with a study of the nonexistence of Random trees are usually generated based on combinatorics limit distribution for the height of the incomplete trie. Itoh and occur also in the context of algorithms from com- and Mahmoud [13] considered five incomplete one-sided puter science. There are many kinds of random trees with variants of binary interval trees and proved that their sizes all different structures, such as recursive trees, search trees, approach some normal random variables. Janson [14] drew binary trees, and interval trees. The asymptotic probability the same result for a larger scale of one-sided interval trees by behavior of random variables in random trees has attracted the renewal theory, and one kind of fragmentation trees was more scholarsβ attention and has become a popular research area. Drmota [1] introduced some labelled and unlabelled discussed by Janson and Neininger [15]. Javanian et al. [16] random trees in his book. Devroye and Janson [2] studied investigated the paths in m-ary interval trees. Su et al. [17] the protected nodes in several random trees. Feng and Hu studied the complete binary interval trees and got the Law [3] researched the phase changes of scale-free trees. The of Large Numbers. In addition, Pan et al. [18] considered the limitinglawfortheheight,size,andsubtreeofbinarysearch construction algorithm about binary interval trees. trees was also considered (see [4β6]). There were also some The binary interval tree is a tree associated with repeated researchers investigating the Zagreb index and nodes of divisions of a line interval of length π₯. The process of divisions random recursive trees (see [7β9]). is as follows. If π₯<1, there is no division in effect; the Binary interval tree is a random structure that underlies associated interval tree consists only of one terminal node. the process of random division of a line interval and parking Supposing that π₯β₯1,webeginwiththeinterval(0, π₯).Divide problems. It has recently been a popular subject. Sibuya the interval (0, π₯) into two subintervals by choosing ππ₯,a andItoh[10]showedthatthenumberofinternaland point uniformly distributed over the interval (0, π₯).Then, external nodes in different levels of binary internal tree is we get two intervals, (0, ππ₯) and (ππ₯,π₯). Each of the two asymptotically normal, from which the asymptotic normality subintervals is further divided at a uniform point of its length, ofthesizeofthetreecouldnotbeachieveddirectly.Prodinger and two smaller subintervals are got as before. If the length of [11] looked into various parameters of the incomplete trie, the subinterval is less than 1, we stop the division. Repeat this 2 Mathematical Problems in Engineering
0 4
0 2.5 2.5 4
0 0.8 0.8 2.5 2.5 3.2 3.2 4
0.8 1.6 1.6 2.5 (a) (b)
Figure 1: (a) The division process. (b) The binary interval tree.
process until the length of every interval (or subinterval) is From the definition of binary interval tree, it is easy to less than 1. see that π1 =3and ππ₯ =1,forπ₯<1.Forourpurposeto We take π₯=4, for instance. Figures 1(a) and 1(b) show investigate the case of π₯β₯1,letππ₯ denote the point chosen (0, π₯) π βΌ π(0, π₯) how the above random division process of interval generates uniformly from interval ;hence, π₯ .Forany (1) abinaryintervaltree. fixed real number 0<π’<π₯,ifππ₯ =π’, we denote ππ’ If some different conditions are added and those intervals to be the size of the left subtree associated with the interval (0, π’) π(2) satisfying the conditions are not allowed to be divided (see . Correspondingly, π₯βπ’ denotes the size of the right subtree associated with the interval (π’, π₯). According to the [13, 14]), then we can get different incomplete interval trees. In (1) (2) particular, if we only divide one subinterval of every interval, rule of division, we can see that ππ’ and ππ₯βπ’ are mutually then the interval tree we get is the so-called one-sided interval independent. Thus, we have tree (see [13]). σ΅¨ π π σ΅¨ =1+π(1) +π(2) ,β0<π’<π₯. (1) It is obvious that interval tree could embody many π₯σ΅¨π=π’ π’ π₯βπ’ properties of random division, so it can elicit lots of valuable This formula implies that if ππ₯ =π’is given, ππ₯ has the same subjects related to probability. For example, for π₯>0, (1) (2) distribution as 1+π +π .Obviously,wecanrewritethe the height of the interval tree is the greatest level of all π’ π₯βπ’ above formula as subintervals after the divisions, denoted by π»π₯;thetotal number of nodes of an interval tree is the total number of π (1) (2) ππ₯ = 1+ππ +ππ₯βπ . (2) intervalsthatweregotfromtherandomdivisionprocess,and π₯ π₯ so on. Let ππ₯ bethesizeoftheintervaltrees,thatis,thetotal Define number of nodes of the binary interval trees. Our intention π1 (π₯) := Eππ₯; is to investigate the random variable ππ₯,thesizeofbinary (3) interval trees. 2 π2 (π₯) := E (ππ₯) . In this paper, the central limit theorem of the size of binary interval trees is investigated. In view of the difficulty It is easy to see that π to calculate the moment generating function of π₯,the π (1) = Eπ =3; method we used is completely different from that in the 1 1 case of one-sided interval trees. In Section 2, we build a π2 (1) =9; (4) distributional recursive equation of ππ₯ andgivetheexpec- tation, the variance, and some high order moments of ππ₯. π1 (π₯) =π2 (π₯) =1, 0<π₯<1. In Section 3, via the contraction method, the limit law of From the distributional recursive equation (2) and the ππ₯ isshowntoapproachtheuniquesolutionofafixed- point distributional equation in the Zolotarev metric space. above boundary conditions, Su et al. [17] calculated the expectation Eππ₯ and the variance Var ππ₯,foranyπ₯β₯0. Finally, we demonstrate that ππ₯,withsuitablestandardization, converges to a normal limiting random variable, as π₯ββ. Lemma 1. Let ππ₯ bethesizeofabinaryintervaltree.Then Eπ =π (π₯) =4π₯β1, π₯β₯1. 2. The Moments of ππ₯ π₯ 1 (5) Lemma 2. π Compared with the one-sided interval trees, the properties Let π₯ bethesizeofabinaryintervaltree.Then of binary interval trees are much more complex. There are 2 {32π₯ ln π₯β16π₯ +8π₯+8, 1β€π₯β€2, a lot of difficulties when it comes to obtaining the moment Var ππ₯ = { (6) π 32 2β20 π₯, π₯β₯2. generating function of π₯. Therefore, the method used in {( ln ) the case of one-sided interval trees (see [13]) is no longer applicable. Here, we build a distributional recursive equation In order to prove that the asymptotic distribution of ππ₯ is 4 of ππ₯. We can calculate the expectation and the variance of normal, we also need the order of E(ππ₯ β Eππ₯) as π₯ββ. ππ₯. Furthermore, we find that the order of the fourth central The following proposition shows the fourth central moment 2 moment of ππ₯ is π(π₯ ) as π₯ goes to infinity. of ππ₯. Mathematical Problems in Engineering 3
β Proposition 3. Let ππ₯ bethesizeofabinaryintervaltree.Then But, we can find that, in the set D , there is only one distri- bution, the standard normal N(0, 1),satisfying(10). 4 2 E (ππ₯ β E [ππ₯]) =π(π₯), π₯σ³¨ββ. (7) (π) Suppose that π is a nonnegative integer. Denote F by Proof. See the appendix. thesetofallrealfunctionsthatareπ times continuous and differentiable, defined on the real line. Let π 3. The CLT for π₯ (π) FπΌ π In this section, we will prove the asymptotic normality of π₯ σ΅¨ σ΅¨ (13) π₯ββ (π) σ΅¨ (π) (π) σ΅¨ σ΅¨ σ΅¨πΌ as . The main method is the contraction method and := {π : π β F , σ΅¨π (π₯) βπ (π¦)σ΅¨ β€ σ΅¨π₯βπ¦σ΅¨ }, some metrics are needed especially the Zolotarev metrics (see [19]). where 0<πΌβ€1is a fixed real number. Let π =π+πΌand First we introduce the Zolotarev metrics. Denote the π L(π) D distribution of the random variable by .Let be the π (π, π) := π (L (π) , L (π)) set of the distributions of all real random variables, and define π π (14) σ΅¨ σ΅¨ (π) = sup {σ΅¨Eπ (π) β Eπ (π)σ΅¨ :πβFπΌ }, Dβ ={πΉ:πΉβD, β« π₯ππΉ (π₯) =0, β« π₯2ππΉ (π₯) R R (8) and then ππ is the Zolotarev metrics with order π on the set D. =1, β« |π₯|3 ππΉ (π₯) <β}. According to the properties of the Zolotarev metric, we know R
ππ (π, π) <β It can be verified that random variable π with L(π) = 2 π π N(0, π ) satisfies the following formula. For any π’β[0,1], ββ E |π| + E |π| <β, (15) π π π Eπ = Eπ , π=1,...,π. π =πβπ’+πβ1βπ’, (9)
Therefore, we can choose π3 as the metric we need on the β and more generally, we have the following lemma. subset D (see [20, 21]); that is, π=2, πΌ=1. This is due β β to the fact that, for any L(π) β D and L(π) β D ,wehave Lemma 4. π π β If and are standard normal random variables, π3(π, π),butif =β L(π), L(π) β D ,thenπ3(π, π) <β. π is uniformly distributed over interval [0, 1],and(π, π, π) are The metric ππ (π, π) has several properties as follows (see mutually independent and then one has [20]):
π π>0 π =πβπ+πβ1βπ. (10) (1) For any constant , π ππ (ππ, ππ) =πππ (π, π) ; (16) Proof. In fact, for any π’β[0,1],wehave
E exp {iπ‘(βπ’π + β1βπ’π)} (2) if random variables π and (π1,π2) are mutually independent, then = E exp {iπ‘(βπ’π + β1βπ’π)} ππ (π1 +π,X2 +π)β€ππ (π1,π2); (17) = Eπi(π‘βπ’)πEπi(π‘β1βπ’)π (11)
π’π‘2 (1βπ’) π‘2 π‘2 π {π , π=1,2,3,...} = {β } {β }= {β }. (3) for random variables and π , exp 2 exp 2 exp 2 π ππ (ππ,π)σ³¨β0σ³¨βππ βπ.σ³¨ (18) Therefore,
E exp {iπ‘(βππ + β1βππ)} Now,webegintoprovethemainresultinthispaper.
1 Theorem 5. Let ππ₯ be the size of a binary interval tree. Then, = β« E exp {iπ‘(βπ’π + β1βπ’π)} ππ’ π₯ββ 0 (12) as ,
1 2 2 π‘ π‘ ππ₯ β Eππ₯ π = β« exp {β }ππ’=exp {β }. σ³¨β N (0, 1) . (19) 0 2 2 βVar ππ₯ 4 Mathematical Problems in Engineering
Proof. Denote Suppose that π₯β₯4;by(A.1)and(21),wehave
β ππ₯ β (4π₯ β 1) σ΅¨ π := ,π₯>0, βσ΅¨ π β (4π₯ β 1) σ΅¨ π₯ π σ΅¨ = π₯ σ΅¨ β π₯σ΅¨π =π‘ σ΅¨ (32 ln 2β20) π₯ π₯ β(32 2β20) π₯σ΅¨ ln ππ₯=π‘ (20) 2 32π₯ ln π₯β16π₯ +8π₯+8 π(1) β E [π(1)] π(2) β E [π(2) ] β (π₯) := β , π₯>0. { π‘ π‘ π₯βπ‘ π₯βπ‘ { + , 2β€π‘β€π₯β2, (32 ln 2β20) π₯ {β β { (32 ln 2β20) π₯ (32 ln 2β20) π₯ { { (1) (1) (2) (2) { π β E [π ] π β E [π ] { π‘ π‘ π₯βπ‘ π₯βπ‘ ThenfromLemmas1and2,weknowthat { + , 1β€π‘<2, {β(32 ln 2β20) π₯ β(32 ln 2β20) π₯ { { { (1) (1) (2) (2) π ππ‘ β E [ππ‘ ] ππ₯βπ‘ β E [ππ₯βπ‘] ππ₯ β Eππ₯ = { + , π₯β2<π‘β€π₯β1, { , π₯β₯2, {β(32 2β20) π₯ β(32 2β20) π₯ { { ln ln { βVar ππ₯ { { { (2) (2) { {π β E [π ]β(4π‘ β 2) {π β Eπ { π₯βπ‘ π₯βπ‘ ,0<π‘<1 β π₯ π₯ { π = β β(π₯) , 1β€π₯<2, (21) { β(32 ln 2β20) π₯ π₯ { βVar π { { π₯ { { {π(1) β E [π(1)]β[4 (π₯βπ‘) β2] { { π‘ π‘ (27) { 2β4π₯ { , π₯β1<π‘<π₯ { , 0<π₯<1. { β(32 ln 2β20) π₯ {β(32 ln 2β20) π₯ π‘ π₯βπ‘ { ββ Μβ β {ππ‘ + ππ₯βπ‘ ,2β€π‘β€π₯β2; β β β β { π₯ π₯ L(π )βD π₯β₯2 L(π )βD { So, we have π₯ for and π₯ for { 0<π₯<2 { π‘ π₯βπ‘ . { β β Μβ β {ππ‘ β (π‘) + ππ₯βπ‘ ,1β€π‘<2, According to the correlative inequality in [21], for any { π₯ π₯ β β { L(π) β D , L(π) β D { , { π‘ π₯βπ‘ π ββ Μβ β = {ππ‘ + ππ₯βπ‘β (π₯βπ‘) , π₯β2<π‘β€π₯β1, { π₯ π₯ { Ξ (2) 3 { π (π, π) β€ β« |π‘| π |π (π<π‘) βπ(π<π‘)| , { π₯βπ‘ 4π‘ β 2 3 (22) {πΜβ β β ,0<π‘<1; Ξ (4) R { π₯βπ‘ { π₯ β(32 ln 2β20) π₯ { { { π‘ 4 (π₯βπ‘) β2 {πββ β , π₯β1<π‘<π₯, where Ξ is the gamma function. Assume that the distribution π‘ { π₯ β(32 ln 2β20) π₯ of random variable π is N(0, 1).ItfollowsfromProposition3 that
where ππ₯ =π‘is the first point chosen from interval (0, π₯) and β 4 Μβ β sup E (ππ₯) <β. {ππ₯,π₯>0}is an independent copy of {ππ₯,π₯>0}. π₯β₯4 (23) If we denote π:=ππ₯/π₯,thenπβΌπ(0,1)and we can rewritetheaboveformulaas Therefore, there exists a constant πΆ>0such that σ΅¨ π β (4π₯ β 1) σ΅¨ βσ΅¨ π₯ σ΅¨ ππ₯σ΅¨ = σ΅¨ π=π’ β 32 2β20 π₯σ΅¨ (28) β σ΅¨ βσ΅¨3 3 ( ln ) σ΅¨π=π’ sup π3 (ππ₯,π)β€πΆ(sup E σ΅¨ππ₯σ΅¨ + E |π| )<β. (24) π₯β₯4 π₯β₯4 2 2 {πβ βπ’+πΜβ β1βπ’, β€π’β€1β ; { π’π₯ (1βπ’)π₯ { π₯ π₯ { Denote { { β β 1 2 {π β (π’π₯) βπ’+πΜ β1βπ’, β€π’< , { π’π₯ (1βπ’)π₯ π₯ π₯ { β β { ππ₯ := π3 (L (ππ₯) ,Ξ¦) =π3 (ππ₯,π) , (25) { 2 1 π πβ βπ’+πΜβ β ((1βπ’) π₯) β1βπ’, 1β <π’β€1β , = { π’π₯ (1βπ’)π₯ (29) { π₯ π₯ { Ξ¦ π { where is standard normal distribution and is standard { β 4π’π₯ β2 1 {πΜ β1βπ’β ,0<π’<; { (1βπ’)π₯ β(32 2β20) π₯ π₯ normal random variable; then we can see that { ln { { 4 (1βπ’) π₯β2 1 { β ππ’π₯βπ’β ,1β<π’<1. 0β€π:=lim sup ππ₯ <β. { β(32 ln 2β20) π₯ π₯ π₯ββ (26)
β Now, we just need to prove that π=0; then the theorem According to the definition of D and π3,itcouldbe β β β β follows. found that (ππ₯ |ππ₯ =π‘)βD for 2<π‘<π₯β2and ππ₯ β D . Mathematical Problems in Engineering 5
σΈ β If we define ππ₯ := (ππ₯ |ππ₯ <2or ππ₯ >π₯β2),thenwe Thus, σΈ β σΈ 4 can also see that π β D .Furthermore,E((π ) )β€πΆ1 for πΏ πΏ πΏ π₯ π₯ 3/2 3/2 3/2 some positive constant πΆ1 by conditioning on ππ₯ and using 2 β« π’ ππ₯π’ππ’ β€ 2π½ β« π’ ππ’ β€ 2π½ β« π’ ππ’ the similar calculation in the appendix. Hence, 2/π₯ 2/π₯ 0 4π½πΏ5/2 π = < ; σΈ 5 10 π3 (π ,π)β€π½, (30) π₯ (33) 1β2/π₯ 1β2/π₯ 3/2 3/2 2 β« π’ ππ₯π’ππ’ β€2 (π+π) β« π’ ππ’ forsomepositiveconstantπ½. πΏ πΏ As we had pointed out before, the standard normal 1 4 (π+π) distribution is the only distribution satisfying (10) in the set β€2(π+π) β« π’3/2ππ’ < , β 5 D . From (25), (14), and Lemma 4, for π₯>4,wehave 0 where π½ is the constant as before and π₯ is sufficiently large. It β σΈ 4 implies that ππ₯ =π3 (π ,π)β€π3 (π ,π)β π₯ π₯ π₯ 4π½ πΏ 1β2/π₯ π β€ +2β« π’3/2π ππ’ +2 β« π’3/2π ππ’ 1β2/π₯ π₯ π₯ π₯π’ π₯π’ β 4π½ 2/π₯ πΏ + β« π3 ((ππ₯ | π = π’) , π) ππ’β€ (34) 2/π₯ π₯ π π 4 (π+π) 4π < + + < +π, 1β2/π₯ 10 10 5 5 β β β β β + β« π3 (ππ₯π’ π’+ππ₯(1βπ’) 1βπ’,π π’ 2/π₯ when π₯ is sufficiently large. Therefore, β 4π + π 1βπ’)ππ’By ( (15) and (29)) π:=lim sup ππ₯ β€ +π. (35) π₯ββ 5 4π½ 1β2/π₯ β β β β β From this equation and the arbitrariness of π>0,wecan β€ + β« π3 (ππ₯π’ π’+ππ₯(1βπ’) 1βπ’,π π’ π₯ 2/π₯ conclude π=0and β 1β2/π₯ π (π ,π)= π =0 β lim 3 π₯ lim π₯ (36) β β π₯ββ π₯ββ + ππ₯(1βπ’) 1βπ’)ππ’+β« π3 (π π’ 2/π₯ immediately. By (18), the theorem holds. (31) β β β β + ππ₯(1βπ’) 1βπ’,π π’+π 1βπ’)ππ’ 4π½ 1β2/π₯ Appendix β β β β€ + β« π3 (ππ₯π’ π’, π π’) ππ’ π₯ 2/π₯ Proof of Proposition 3 1β2/π₯ β β β From the process of generating the binary interval trees, it is + β« π3 (ππ₯(1βπ’) 1βπ’,π 1βπ’)ππ’ 2/π₯ obvious that, for given ππ₯ =π‘, π‘ β (0, π₯), 1β2/π₯ σ΅¨ 4π½ (ππ₯ β E [ππ₯])σ΅¨π =π‘ β β β π₯ = +2β« π3 (ππ₯π’ π’, π π’) ππ’ π₯ 2/π₯ (1) (1) (2) (2) {(ππ‘ β E [ππ‘ ]) + (ππ₯βπ‘ β E [ππ₯βπ‘]) , 1 β€ π‘ β€ π₯ β 1; { 4π½ 1β2/π₯ { (A.1) = +2β« π’3/2π (πβ ,π)ππ’ π (2) (2) 3 π₯π’ = {ππ₯βπ‘ β E [ππ₯βπ‘]β(4π‘ β 2) ,0<π‘<1; π₯ 2/π₯ { { (1) (1) 1β2/π₯ {π β E [π ]β[4 (π₯βπ‘) β2] , π₯β1<π‘<π₯, 4π½ 3/2 π‘ π‘ = +2β« π’ ππ₯π’ππ’. π₯ 2/π₯ where ππ₯ =π‘is the first point chosen from interval (0, π₯).For π₯β₯1, if we denote 5/2 (1) (1) Given π>0,letπΏ>0be small enough such that π½πΏ <π/8. ππ₯ := ππ₯ β E [ππ₯ ], For any fixed πΏ>0,whenπ₯ is sufficiently large, then (A.2) β (2) (2) ππ₯ := ππ₯ β E [ππ₯ ], 4π½ π < , then we have π₯ 10 β {ππ‘ +ππ₯βπ‘,1β€π‘β€π₯β1; 2 { <πΏ, { (32) σ΅¨ π β π₯ ππ₯σ΅¨π =π‘ = {ππ₯βπ‘ β (4π‘ β 2) ,0<π‘<1; (A.3) π₯ { { sup ππ₯π’ <π+π. πΏβ€π’β€1 {ππ‘ β [4 (π₯βπ‘) β2] , π₯β1<π‘<π₯. 6 Mathematical Problems in Engineering
3 4 We need to calculate Eππ₯ first before we get Eππ₯.Forπ₯>3, and when π₯>3,forthepartπ2,wehave we have π₯ 1 6 3 3 1 π2 =β β« [4 (π₯βπ‘) β2] Var [ππ‘]ππ‘ E [ππ₯] = E [E (ππ₯ |ππ₯)] = β« E [ππ₯βπ‘ π₯ π₯β1 π₯ 0 π₯ π₯ 6 3 1 =β β« [4 (π₯βπ‘) β2][(32 ln 2β20) π‘] ππ‘ β (4π‘ β 2)] ππ‘ + β« E {ππ‘ π₯ π₯β1 π₯ π₯β1 (A.7) 1 π₯β1 6 (32 ln 2β20) 3 1 β 3 =β β« (4π‘ β 2)(π₯βπ‘) ππ‘ β [4 (π₯βπ‘) β2]} ππ‘ + β« E [ππ‘ +ππ₯βπ‘] ππ‘ π₯ 0 π₯ 1 π₯ 2 (32 ln 2β20) 2 3 1 = . = β« E {ππ‘ β [4 (π₯βπ‘) β2]} ππ‘ + π₯ π₯ π₯β1 π₯ π₯β1 β 3 2 Therefore, β β« E [ππ‘ +ππ₯βπ‘] ππ‘ = (β 1 π₯ π₯ π₯ 3 2 3 2 (32 ln 2β20) 3 6 (A.4) E [ππ₯] = β« E [ππ‘ ] ππ‘ + , π₯>3. (A.8) β β« [4 (π₯βπ‘) β2] ππ‘ + π₯ 1 π₯ π₯β1 π₯ π₯ 2 6 That is, β β« [4 (π₯βπ‘) β2] E [ππ‘]ππ‘ β π₯β1 π₯ π₯ π₯ 3 3 2 2 π₯E [π ]=2β« E [π ]ππ‘+2 32 2β20 , β β« [4 (π₯βπ‘) β2] E [π ]ππ‘+ π₯ π‘ ( ln ) π‘ 1 π₯β1 π₯ (A.9) π₯ π₯β1 π₯>3. 3 1 3 β β« E [ππ‘ ]ππ‘)+ β« (E [ππ‘] π₯β1 π₯ 1 Via differentiation with respect to π₯, we get the differential 2 β β 2 +3E [(ππ‘) ππ₯βπ‘]+3E [ππ‘ (ππ₯βπ‘) ] equation:
β 3 + E [π ] )ππ‘. σΈ 1 π₯βπ‘ (E [π3]) β E [π3]=0, π₯>3. π₯ π₯ π₯ (A.10) β In view of the independence between ππ‘ and ππ₯βπ‘ and that E[π ]=E[πβ]=0 1β€π‘β€π₯β1 π‘ π‘ holds for any ,wehave The solution to this differential equation is π₯ 3 2 3 3 E [ππ₯] =β β« [4 (π₯βπ‘) β2] ππ‘ E [π ]=ππ₯, π₯ > 3, π₯ π₯β1 π₯ 0 (A.11) 6 π₯ β β« 4 π₯βπ‘ β2 E [π2]ππ‘ [ ( ) ] π‘ where π0 is a constant real number. π₯ π₯β1 4 Similarly, for E[ππ₯] ,whenπ₯>4,wehave 2 π₯ 2 π₯β1 + β« E [π3]ππ‘+ β« E [π3]ππ‘ π‘ π‘ π₯ π₯ π₯β1 π₯ 1 2 E [π ]4 = β« E {π β [4 (π₯βπ‘) β2]}4 ππ‘ π₯ π₯ π₯ π‘ 2 3 π₯β1 =β β« [4 (π₯βπ‘) β2] ππ‘ (A.5) (A.12) π₯ π₯β1 1 π₯β1 + β« E [π +πβ ]4 ππ‘. π₯ π‘ π₯βπ‘ 6 2 π₯ 1 β β« [4 (π₯βπ‘) β2] E [ππ‘ ]ππ‘ π₯ π₯β1 π πβ Eπ =0 2 π₯ Because π‘ is independent of π₯βπ‘,and π‘ holds for any 3 1β€π‘β€π₯β1 + β« E [ππ‘ ]ππ‘ ,weget π₯ 1 2 π₯ 2 π₯ := π +π + β« E [π3]ππ‘. 4 4 1 2 π‘ E [ππ₯] = β« [4 (π₯βπ‘) β2] ππ‘ π₯ 1 π₯ π₯β1 12 π₯ It is easy to see that 2 2 + β« [4 (π₯βπ‘) β2] E [ππ‘ ]ππ‘ π₯ π₯β1 2 1 3 8 π₯ π1 =β β« π’ ππ’ = 0, (A.6) β β« [4 (π₯βπ‘) β2] E [π3]ππ‘ 2π₯ β2 π‘ π₯ π₯β1 Mathematical Problems in Engineering 7
π₯ 2 4 When π₯>4,forthepartπΌ5,wehave + β« E [ππ‘ ]ππ‘ π₯ π₯β1 π₯β1 π₯β1 6 β 2 1 4 β 4 β 2 πΌ = β« E [π π ] ππ‘ + β« (E [π ]+E [π ] +6E [π π ] )ππ‘ 5 π‘ π₯βπ‘ π‘ π₯βπ‘ π‘ π₯βπ‘ π₯ 1 π₯ 1 2 π₯ 2 = β« [4 (π₯βπ‘) β2]4 ππ‘ 6 2 β 2 = β« E [ππ‘ ] E [ππ₯βπ‘] ππ‘ π₯ π₯β1 π₯ 1 12 π₯ + β« [4 (π₯βπ‘) β2]2 E [π2]ππ‘ 6 π₯β1 π₯ π‘ 2 β 2 π₯β1 + β« E [ππ‘ ] E [ππ₯βπ‘] ππ‘ π₯ π₯β2 8 π₯ β β« [4 (π₯βπ‘) β2] E [π3]ππ‘ (A.17) π‘ π₯β2 π₯ π₯β1 6 2 β 2 + β« E [ππ‘ ] E [ππ₯βπ‘] ππ‘ π₯ π₯β1 π₯ 2 2 4 6 β 2 + β« E [ππ‘] ππ‘ + β« E [ππ‘ππ₯βπ‘] ππ‘ :=1 πΌ π₯ 1 π₯ 1 2 12 2 β 2 = β« E [ππ‘ ] E [ππ₯βπ‘] ππ‘ +πΌ2 +πΌ3 +πΌ4 +πΌ5. π₯ 1
(A.13) π₯β2 6 2 β 2 + β« E [ππ‘ ] E [ππ₯βπ‘] ππ‘. π₯ 2 In particular, for the part πΌ1,wehave 2 Noting that E[ππ‘ ]=Var[ππ‘] and (6), we can see that 32 πΌ = . 1 (A.14) π₯ 5π₯ 2 β 2 β« E [ππ‘ ] E [ππ₯βπ‘] ππ‘ 1 When π₯>3,forthepartπΌ2,wehave 2 2 =2β« (32π‘ ln π‘β16π‘ +8π‘+8) π₯ 1 12 2 2 πΌ2 = β« [4 (π₯βπ‘) β2] E [ππ‘ ]ππ‘ π₯ π₯β1 β ((32 ln 2β20)(π₯βπ‘)) ππ‘ 12 π₯ = β« [4 (π₯βπ‘) β2]2 Var [π ]ππ‘ π‘ π₯β2 π₯ π₯β1 + β« ((32 ln 2β20) π‘) 12 π₯ 2 = β« [4 (π₯βπ‘) β2]2 [(32 2β20) π‘] ππ‘ (A.18) π₯ ln π₯β1 1 (A.15) β ((32 2β20)(π₯βπ‘)) ππ‘ = (32 2β20)2 12 (32 2β20) 1 ln 6 ln = ln β« (4π‘ β 2)2 (π₯βπ‘) ππ‘ π₯ 0 3 4 2 2 1 48 24 (32 2β20) β π₯ β (32 ln 2β20) π₯ + (32 ln 2β20) = (32 2β20) β ln 3 3 3 ln 3π₯ 16 6π β (256 2 β 168) π₯+ (32 2β20) := π π₯3 := 12π β 1 , ln 9 ln 3 1 π₯ 2 +π2π₯ +π1π₯+π0, where π1 := 4(32 ln 2 β 20)/3 is a constant. π₯>4 πΌ When ,forthepart 3,wehave where
π₯ 16 8 3 π0 = (32 ln 2β20) ; πΌ3 =β β« [4 (π₯βπ‘) β2] E [ππ‘ ]ππ‘ 9 π₯ π₯β1 8 π₯ 1 =β β« [4 (π₯βπ‘) β2] π0π‘ππ‘ (A.16) π1 = (32 ln 2β20)(256 ln 2 β 168) ; π₯ π₯β1 3 (A.19) 8π 1 8π =β 0 β« (4π‘ β 2)(π₯βπ‘) ππ‘ = 0 , 4 2 π2 =β (32 ln 2β20) ; π₯ 0 3π₯ 3
1 2 π3 = (32 ln 2β20) . where π0 isthesameasthatin(A.11). 6 8 Mathematical Problems in Engineering
Therefore, Education Institutions of China (no. 15KJA110001), Qing Lan 2 π₯ Project, PAPD, Program of Excellent Science and Technology E [π ]4 = β« E [π ]4 ππ‘ +πΌ +πΌ +πΌ +πΌ, Innovation Team of the Jiangsu Higher Education Institu- π₯ π₯ π‘ 1 2 3 5 1 (A.20) tions of China, Project of Construction for Superior Subjects π₯>4. of Statistics of Jiangsu Higher Education Institutions, and ProjectoftheKeyLabofFinancialEngineeringofJiangsu That is, Province. π₯ 4 4 π₯E [ππ₯]=2β« E [ππ‘] ππ‘ +1 (πΌ +πΌ2 +πΌ3 +πΌ5)π₯, References 1 (A.21) π₯>4. [1] M. Drmota, Random Trees, Springer, New York, NY,USA, 2009. [2] L. Devroye and S. Janson, βProtected nodes and fringe subtrees Via differentiation with respect to π₯, we get the differential in some random trees,β Electronic Communications in Probabil- equation: ity,vol.19,article6,2014. [3] Q. Feng and Z. Hu, βPhase changes in the topological indices of σΈ 1 π 1 (Eπ4) β Eπ4 =12 1 +6(3ππ₯+2π +π ), scale-free trees,β Journal of Applied Probability,vol.50,no.2,pp. π₯ π₯ π₯ π₯ 3 2 1 π₯ 516β532, 2013. (A.22) π₯>4. [4] N. Broutin and P.Flajolet, βThe distribution of height and diam- eter in random non-plane binary trees,β Random Structures and Algorithms,vol.41,no.2,pp.215β252,2012. The solution to this differential equation is [5] F. Dennert and R. Grubel,Β¨ βOn the subtree size profile of binary 4 2 search trees,β Combinatorics, Probability and Computing,vol.19, Eππ₯ =18π3π₯ +12π2π₯ ln π₯+ππ₯β6π1 β 12π1, (A.23) no. 4, pp. 561β578, 2010. π₯>4, [6] H. M. Mahmoud, βOne-sided variations on binary search trees,β Annals of the Institute of Statistical Mathematics,vol.55,no.4, where π is a constant and the constants π1,π1,π2,π3 are real pp. 885β900, 2003. numbers as defined before. From this equation, Proposition 3 [7]Q.FengandZ.Hu,βOntheZagrebindexofrandomrecursive follows. trees,β Journal of Applied Probability,vol.48,no.4,pp.1189β1196, 2011. Consent [8] R. GrubelΒ¨ and I. Michailow, βRandom recursive trees: a bound- ary theory approach,β Electronic Journal of Probability,vol.20, Informed consent was obtained from all individual partici- article 37, 2015. pants included in the study. [9] H. M. Mahmoud and M. D. Ward, βAsymptotic properties of protected nodes in random recursive trees,β JournalofApplied Conflict of Interests Probability, vol. 52, no. 1, pp. 290β297, 2015. [10] M. Sibuya and Y. Itoh, βRandom sequential bisection and its The authors declare that they have no conflict of interests. associated binary tree,β Annals of the Institute of Statistical Mathematics,vol.39,no.1,pp.69β84,1987. Acknowledgments [11] H. Prodinger, βHow to select a loser,β Discrete Mathematics,vol. 120, no. 1β3, pp. 149β159, 1993. The authors are most grateful to the referee and the editor [12] J. A. Fill, H. M. Mahmoud, and W. Szpankowski, βOn the for their very thorough reading of the paper and valuable distribution for the duration of a randomized leader election suggestions, which greatly improve the original results and algorithm,β The Annals of Applied Probability,vol.6,no.4,pp. presentation of this paper. Jie Liuβs work was supported by 1260β1283, 1996. the National Natural Science Foundation of China (nos. [13] Y. Itoh and H. M. Mahmoud, βOne-sided variations on interval 11101394, 71471168, and 71520107002), China Postdoctoral trees,β Journal of Applied Probability, vol. 40, no. 3, pp. 654β670, Science Foundation Funded Project (nos. 201104312 and 2003. 20100480688), Fund for the Doctoral Program of Higher [14] S. Janson, βOne-sided interval trees,β Journal of the Iranian Education Foundation (no. 20113402120005), and the Fun- Statistical Society,vol.3,pp.149β164,2004. damental Research Funds for the Central Universities of [15] S. Janson and R. Neininger, βThe size of random fragmentation China(no.WK2040160008).YangYangβsworkwassupported trees,β Probability Theory and Related Fields,vol.142,no.3-4,pp. by the National Natural Science Foundation of China (no. 399β442, 2008. 71471090), the Humanities and Social Sciences Foundation [16] M. Javanian, H. M. Mahmoud, and M. Vahidi-Asl, βPaths in m- of the Ministry of Education of China (no. 14YJCZH182), ary interval trees,β Discrete Mathematics,vol.287,no.1β3,pp. China Postdoctoral Science Foundation (nos. 2014T70449 45β53, 2004. and 2012M520964), Natural Science Foundation of Jiangsu [17]C.Su,J.Liu,andZ.S.Hu,βTheLawsoflargenumbersforthe Province of China (no. BK20131339), the Major Research size of complete interval trees,β Advances in Mathematics,vol. Plan of Natural Science Foundation of the Jiangsu Higher 36, no. 2, pp. 181β188, 2007. Mathematical Problems in Engineering 9
[18] W. Pan, S.-K. Li, X. Cai, and L. Zeng, βThe construction algo- rithm of binary interval tree nodes based on lattice partition,β JournalofComputer-AidedDesign&ComputerGraphics,vol. 23, pp. 1115β1122, 2011. [19] V. M. Zolotarev, Modern Theory of Summation of Random Variables,VSP,Utrecht,TheNetherlands,1997. [20] V. M. Zolotarev, βApproximation of distributions of sums of independent random variables with values in infinite- dimensional spaces,β Theory of Probability and Its Applications, vol. 21, no. 4, pp. 721β737, 1976. [21] S. T. Rachev and L. Ruschendorf,Β¨ βProbability metrics and recursive algorithms,β Advances in Applied Probability,vol.27, no. 3, pp. 770β799, 1995. Advances in Advances in Journal of Journal of Operations Research Decision Sciences Applied Mathematics Algebra Probability and Statistics Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014
The Scientific International Journal of World Journal Differential Equations Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014
Submit your manuscripts at http://www.hindawi.com
International Journal of Advances in Combinatorics Mathematical Physics Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014
Journal of Journal of Mathematical Problems Abstract and Discrete Dynamics in Complex Analysis Mathematics in Engineering Applied Analysis Nature and Society Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014
International Journal of Journal of Mathematics and Mathematical Discrete Mathematics Sciences
Journal of International Journal of Journal of Function Spaces Stochastic Analysis Optimization Hindawi Publishing Corporation Hindawi Publishing Corporation Volume 2014 Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014