Computing the Distribution of the Maximum of Gaussian Random Processes

The aim of this paper is to propose an Splus program to calculate bounds for the distribution of the maximum of a smooth Gaussian process on a fixed interval. We generalize the results given in Azaïs et al. (1999) to the case of the absolute value of the Gaussian process and to the non-homogeneous case. Our method relies on calculations of the first three terms of the Rice's series. Some applications are given to illustrate the method and the performances of the program. The corresponding Splus functions are available at the URL: http://www.lsp.ups-tlse.fr/Cdelmas/software.html.     

Dependence Between Extreme Values of Discrete and Continuous Time Locally Stationary Gaussian Processes

The maximum of a continuous, locally stationary Gaussian process which satisfies Bermans condition on the long range dependence is compared with the maximum of this process sampled at discrete time points. These two extreme values are asymptotically totally dependent if the grid of the discrete time points is sufficiently dense, and asymptotically independent if the the grid points are sparse.AMS 2000 Subject Classification. Primary—60F05, Secondary—60G15     

Extremes of Gaussian processes with a smooth random variance

Let ξ(t) be a standard stationary Gaussian process with covariance function r(t), and η(t), another smooth random process. We consider the probabilities of exceedances of ξ(t)η(t) above a high level u occurring in an interval [0,T] with T>0. We present asymptotically exact results for the probability of such events under certain smoothness conditions of this process ξ(t)η(t), which is called the random variance process. We derive also a large deviation result for a general class of conditional Gaussian processes X(t) given a random element Y.     

Discrete and Continuous Time Extremes of Gaussian Processes

Joint distribution of maximums of a Gaussian stationary process in continuous time and in uniform grid on the real axis is studied. When the grid is sufficiently sparse, maxima are asymptotically independent. When the grid is sufficiently tight, the maximums asymptotically coincide. In the boundary case which we call Pickands grid, the limit distribution is non-degenerate. It calculated in terms of a Pickands type constant.AMS 2000 Subject Classification. Primary—60G70, Secondary—60G15*Partially supported by the Scientific foundation of the Netherlands, RFFI grant 0401-00700 and grant DFG 436 RUS 113/722.     

Weak convergence of random polygonal lines to a Gaussian process

We obtain a limit theorem of convergence in distribution for random polygonal lines defined by sums of independent random variables with replacements. In a particular case, the limit is the Gaussian Ornstein-Uhlenbeck process.__________Translated from Lietuvos Matematikos Rinkinys, Vol. 45, No. 1, pp. 33–44, January–March, 2005.Translated by V. Mackeviius     

Asymptotic Poisson Character of Extremes in Non-Stationary Gaussian Models

Let X be a non-stationary Gaussian process, asymptotically centered with constant variance. Let u be a positive real. Define R_u(t) as the number of upcrossings of level u by the process X on the interval (0, t]. Under some conditions we prove that the sequence of point processes (R_u)_u>0 converges weakly, after normalization, to a standard Poisson process as u tends to infinity. In consequence of this study we obtain the weak convergence of the normalized supremum to a Gumbel distribution.AMS 2000 Subject Classifications Primary—60G70, 60G15     

Extremes of Normed Empirical Moment Generating Function Processes

We present a method for deriving the limiting distribution of the maximum of a normed empirical moment generating function process indexed by one parameter. We first extend slightly the results of Csörg et al. (1986b) to provide the rate of convergence for a Gaussian approximation to a non-Donsker empirical process. In cases we consider, the maximum tends to infinity in probability, but when appropriately scaled has a limiting Gumbel extreme value distribution.AMS 2000 Subject Classification Primary—62E20, 60G70*Author for correspondence: School of Mathematics and Statistics, University of Sydney, New South Wales 2006, Australia.     

Maximal inequalities for the Ornstein-Uhlenbeck process

Let be the Ornstein-Uhlenbeck velocity process solving

with , where 0$"> and is a standard Brownian motion. Then there exist universal constants 0$">and 0$"> such that

for all stopping times of . In particular, this yields the existence of universal constants 0$"> and 0$"> such that

for all stopping times of . This inequality may be viewed as a stopped law of iterated logarithm. The method of proof relies upon a variant of Lenglart's domination principle and makes use of Itô calculus.

     

Boundary Crossing for the Difference of Two Ordinary or Compound Poisson Processes

We consider the lower boundary crossing problem for the difference of two independent compound Poisson processes. This problem arises in the busy period analysis of single-server queueing models with work removals. The Laplace transform of the crossing time is derived as the unique solution of an integral equation and is shown to be given by a Neumann series. In the case of ±1 jumps, corresponding to queues with deterministic service times and work removals, we obtain explicit results and an approximation useful for numerical purposes. We also treat upper boundaries and two-sided stopping times, which allows to derive the conditional distribution of the maximum workload up to time t, given the busy period is longer than t.     

Exact asymptotics for Boundary crossings of the brownian bridge with trend with application to the Kolmogorov test

We consider a boundary crossing probability of a Brownian bridgeB ₀ and a piecewise linear boundary functionu(t)−γh(t). The main result of this paper is an asymptotic expansion for γ→∞ of the boundary crossing probability thatB ₀ (t) is larger than the piecewise linear boundary functionu(t)−γh(t) for somet. Such probabilities occur for instance in the context of change point problems when the Kolmogorov test is used. Examples are discussed showing that the approximation is rather accurate even for small positive γ values. Supported by the Swiss National Science Foundation Grant 20-55586.98.     

The dependence of extreme values of discrete and continuous time strongly dependent Gaussian processes

In this note, the asymptotic relation between the maximum of a continuous strongly dependent stationary Gaussian process and the maximum of this process sampled at discrete time points is studied. It is shown that these two extreme values are asymptotically totally dependent no matter what the grid of the discrete time points is.     

Extremes of deterministic sub‐sampled moving averages with heavy‐tailed innovations

Let {X_k}_k?1 be a strictly stationary time series. For a strictly increasing sampling function g:?→? define Y_k=X_g(k) as the deterministic sub‐sampled time series. In this paper, the extreme value theory of {Y_k} is studied when X_k has representation as a moving average driven by heavy‐tailed innovations. Under mild conditions, convergence results for a sequence of point processes based on {Y_k} are proved and extremal properties of the deterministic sub‐sampled time series are derived. In particular, we obtain the limiting distribution of the maximum and the corresponding extremal index. Copyright © 2003 John Wiley & Sons, Ltd.     

线性过程方差变点的平方累计和SCUSUM估计(英文)

This paper considers the estimation problem of a variance change-point in linear process.Consistency of a SCUSUM type change-point estimator is proved and its rate of convergence is established.The mean-unknown case is also considered.     

一类N参数Gauss过程的异常震动点集合的Hausdorff维数

引进了一类N参数Gauss过程,它具有比N参数Wiener过程更为一般的性质．给出了此类N参数Gauss过程的异常震动点集的定义,并且定义了此异常震动点集的Hausdorff维数．研究了此类过程的异常震动点集Hausdorff维数,给出了它的一个确切的表达式,从而获得了与Zacharie (2001)的有关两参数Wiener过程的类似的结果．考虑的参数点集是一般的超长方体．而不是Zacharie (2001)考虑的超正方体．在此更为一般的情况下,首先建立了文中引进的过程的Fernique不等式．利用此不等式和Slepian引理,证明了过程的Lévy连续模定理．Zacharie(2001)关于Hausdorff维数公式的证明依赖于两参数Wiener过程的独立增量性,而这里引进的过程不具有这种性质,因此,必须采用新的证明途径．     

Weak Type Inequality for a Family of Singular Integral Operators Related with the Gaussian Measure

In this paper we study a family of singular integral operators that generalizes the higher order Gaussian Riesz Transforms and find the right weight w to make them continuous from L ¹(wdγ) into L ^1, ∞(dγ), being Some boundedness properties of these operators had already been derived by Urbina (Ann Scuola Norm Sup Pisa Cl Sci 17(4):531–567, 1990) and Pérez (J Geom Anal 11(3):491–507, 2001).      

On maxima of partial samples in Gaussian sequences with pseudo-stationary trends

We consider a Gaussian stationary sequence added by a pseudo-stationary trend and prove a limit theorem for joint distribution of its maximum and maximum of its subsequence. Supported by RFFI grants 07-01-00077 and 06-01-00454. Published in Lietuvos Matematikos Rinkinys, Vol. 47, No. 1, pp. 58–67, January–March, 2007.     

Compound Poisson Approximation of the Number of Exceedances in Gaussian Sequences

Consider a finite sequence of Gaussian random variables. Count the number of exceedances of some level a, i.e. the number of values exceeding the level. Let this level and the length of the sequence increase simultaneously so that the expected number of exceedances remains fixed. It is well-known that if the long-range dependence is not too strong, the number of exceeding points converges in distribution to a Poisson distribution. However, for sequences with some individual large correlations, the Poisson convergence is slow due to clumping. Using Steins method we show that, at least for m-dependent sequences, the rate of convergence is improved by using compound Poisson as approximating distribution. An explicit bound for the convergence rate is derived for the compound Poisson approximation, and also for a subclass of the compound Poisson distribution, where only clumps of size two are considered. Results from numerical calculations and simulations are also presented.     

On High-Level Exceedances of Gaussian Fields and the Spectrum of Random Hamiltonians

We study the almost sure asymptotic structure of high-level exceedances by Gaussian random field (x), xV with correlated values, where {V} is a family of -dimensional cubes increasing to Z . The results are applied to the study of the asymptotic behaviour of extreme eigenvalues of random Schrödinger operator in V.     

Chung-type Law of the Iterated Logarithm on l^P-valued Gaussian Processes

By estimating small ball probabilities for l^P-valued Gaussian processes, a Chung-type law of the iterated logarithm of l^P-valued Gaussian processes is given.