Edgeworth expansions for profiles of lattice branching random walks

被引:18
作者
Gruebel, Rudolf [1 ]
Kabluchko, Zakhar [2 ]
机构
[1] Leibniz Univ Hannover, Inst Math Stochast, Welfengarten 1, D-30167 Hannover, Germany
[2] Univ Munster, Inst Math Stat, Orleans Ring 10, D-48149 Munster, Germany
来源
ANNALES DE L INSTITUT HENRI POINCARE-PROBABILITES ET STATISTIQUES | 2017年 / 53卷 / 04期
关键词
Branching random walk; Edgeworth expansion; Central limit theorem; Profile; Biggins martingale; Random analytic function; Mod-phi-convergence; Height; Mode; BINARY SEARCH-TREES; CENTRAL-LIMIT-THEOREM; RANDOM RECURSIVE TREES; RANDOM ENVIRONMENT; RANDOM-VARIABLES; CONVERGENCE; GROWTH; WIDTH; ASYMPTOTICS; MARTINGALES;
D O I
10.1214/16-AIHP785
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Consider a branching random walk on Z in discrete time. Denote by L-n(k) the number of particles at site k is an element of Z at time n is an element of N-0. By the profile of the branching random walk (at time n) we mean the function k bar right arrow L-n(k). We establish the following asymptotic expansion of L-n(k), as n -> infinity: e(-phi(0))n L-n (k) = e(-1/2x2/n(k))/root 2 pi phi ''(0)n Sigma(r)(j=0) Fj(x(n)(k))/n(j/2) + 0(n(-r+1/2)) a.s., where r is an element of N-0 is arbitrary,phi(beta) = log Sigma(k is an element of Z)e(beta k)EL(1)(k) is the cumulant generating function of the intensity of the branching random walk and x(n) (k) = k-phi '(0)n/root phi ''(0)n The expansion is valid uniformly in k is an element of Z with probability 1 and the F-j's are polynomials whose random coefficients can be expressed through the derivatives of phi and the derivatives of the limit of the Biggins martingale at 0. Using exponential tilting, we also establish more general expansions covering the whole range of the branching random walk except its extreme values. As an application of this expansion for r = 0, 1, 2 we recover in a unified way a number of known results and establish several new limit theorems. In particular, we study the a.s. behavior of the individual occupation numbers L-n(k(n)), where k(n) is an element of Z depends on n in some regular way. We also prove a.s. limit theorems for the mode arg max(k is an element of Z) L-n(k) and the height max(k is an element of Z) L-n(k) of the profile. The asymptotic behavior of these quantities depends on whether the drift parameter. phi ' (0) is integer, non-integer rational, or irrational. Applications of our results to profiles of random trees including binary search trees and random recursive trees will be given in a separate paper.
引用
收藏
页码:2103 / 2134
页数:32
相关论文
共 40 条
[1]  
Asmussen S., 1976, Stochastic Processes & their Applications, V4, P1, DOI 10.1016/0304-4149(76)90022-3
[2]   ON DEVIATIONS OF THE SAMPLE-MEAN [J].
BAHADUR, RR ;
RAO, RR .
ANNALS OF MATHEMATICAL STATISTICS, 1960, 31 (04) :1015-1027
[3]  
Bhattacharya R. N., 1976, WILEY SERIES PROBABI
[4]  
BIGGINS J. D., 2010, London Math. Soc. Lecture Note Ser., V378, P113
[5]  
Biggins J.D., 1991, IMS LECT NOTES MONOG, V18, P159, DOI DOI 10.1214/LNMS/1215459294
[6]   UNIFORM-CONVERGENCE OF MARTINGALES IN THE BRANCHING RANDOM-WALK [J].
BIGGINS, JD .
ANNALS OF PROBABILITY, 1992, 20 (01) :137-151
[7]   GROWTH-RATES IN THE BRANCHING RANDOM-WALK [J].
BIGGINS, JD .
ZEITSCHRIFT FUR WAHRSCHEINLICHKEITSTHEORIE UND VERWANDTE GEBIETE, 1979, 48 (01) :17-34
[8]   A note on the growth of random trees [J].
Biggins, JD ;
Grey, DR .
STATISTICS & PROBABILITY LETTERS, 1997, 32 (04) :339-342
[9]   CHERNOFFS THEOREM IN BRANCHING RANDOM-WALK [J].
BIGGINS, JD .
JOURNAL OF APPLIED PROBABILITY, 1977, 14 (03) :630-636
[10]   THE CENTRAL-LIMIT-THEOREM FOR THE SUPERCRITICAL BRANCHING RANDOM-WALK, AND RELATED RESULTS [J].
BIGGINS, JD .
STOCHASTIC PROCESSES AND THEIR APPLICATIONS, 1990, 34 (02) :255-274