On a Laplace Transform Based Metric for Probabilities on a Hilbert Space

Let D be a closed subset of a real separable Hilbert space H. Let (D) denote the set of all Borel probability measures on D and (D) the set of all probabilities with integrable Laplace transform. A metric d, based on the Laplace transform, is defined on (D). Topological properties, viz., separability, connectedness, completeness, compactness and local compactness, of (D, d are investigated, and the d-topology is compared with the topology of weak convergence.     

Coxian representations of generalized Erlang distributions

This paper studies Coxian representations of generalized Erlang distributions. A nonlinear program is derived for computing the parameters of minimal Coxian representations of generalized Erlang distributions. The nonlinear program is also used to characterize the triangular order and the admissible region of generalized Erlang distributions. It is shown that the admissible region associated with a triangular order may not be convex. For generalized Erlang distributions of ME-order 3, a minimal Coxian representation is found explicitly. In addition, an algorithm is developed for computing a special type of ordered Coxian representations - the bivariate Coxian representation - for generalized Erlang distributions.     

Convergence in Distribution of Products of I.I.D. Nonnegative Matrices

Let (X _i) be a sequence of m × m i.i.d. stochastic matrices with distribution . Then ⁿ is the distribution of X _n X _n–1 ...X ₁. Simple sufficient conditions for the weak convergence of ( ⁿ ) are presented here. An extremely simple (and verifiable) necessary and sufficient condition is provided for m= 3. The method for m= 3 works for m> 3 even though calculations are more involved for higher values of m. We also discuss the purity of the limit distribution for m2.     

The Gerber–Shiu discounted penalty functions for a risk model with two classes of claims

In this paper, we consider the ruin problems for a risk model involving two independent classes of insurance risks. We assume that the claim number processes are independent Poisson and generalized Erlang(n) processes, respectively. When the generalized Lundberg equation has distinct roots with positive real parts, both of the Gerber–Shiu discounted penalty functions with zero initial surplus and the Laplace transforms of the Gerber–Shiu discounted penalty functions are obtained. Finally, some explicit expressions for the Gerber–Shiu discounted penalty functions with positive initial surplus are given when the claim size distributions belong to the rational family.     

Waiting-time tail probabilities in queues with long-tail service-time distributions

We consider the standardGI/G/1 queue with unlimited waiting room and the first-in first-out service discipline. We investigate the steady-state waiting-time tail probabilitiesP(W>x) when the service-time distribution has a long-tail distribution, i.e., when the service-time distribution fails to have a finite moment generating function. We have developed algorithms for computing the waiting-time distribution by Laplace transform inversion when the Laplace transforms of the interarrival-time and service-time distributions are known. One algorithm, exploiting Pollaczek's classical contourintegral representation of the Laplace transform, does not require that either of these transforms be rational. To facilitate such calculations, we introduce a convenient two-parameter family of long-tail distributions on the positive half line with explicit Laplace transforms. This family is a Pareto mixture of exponential (PME) distributions. These PME distributions have monotone densities and Pareto-like tails, i.e., are of orderx ^–r forr>1. We use this family of long-tail distributions to investigate the quality of approximations based on asymptotics forP(W>x) asx. We show that the asymptotic approximations with these long-tail service-time distributions can be remarkably inaccurate for typicalx values of interest. We also derive multi-term asymptotic expansions for the waiting-time tail probabilities in theM/G/1 queue. Even three terms of this expansion can be remarkably inaccurate for typicalx values of interest. Thus, we evidently must rely on numerical algorithms for determining the waiting-time tail probabilities in this case. When working with service-time data, we suggest using empirical Laplace transforms.     

Convergence of interval-type algorithms for generalized fractional programming

The purpose of this paper is to analyze the convergence of interval-type algorithms for solving the generalized fractional program. They are characterized by an interval [LB _k , UB _k ] including*, and the length of the interval is reduced at each iteration. A closer analysis of the bounds LB _k and UB _k allows to modify slightly the best known interval-type algorithm NEWMODM accordingly to prove its convergence and derive convergence rates similar to those for a Dinkelbach-type algorithm MAXMODM under the same conditions. Numerical results in the linear case indicate that the modifications to get convergence results are not obtained at the expense of the numerical efficiency since the modified version BFII is as efficient as NEWMODM and more efficient than MAXMODM.This research was supported by NSERC (Grant A8312) and FCAR (Grant 0899).     

An exact rate of convergence in the functional central limit theorem for special martingale difference arrays

Summary Since the topology of weak convergence of probability distributions on the Borel -field of the space C= C([0, 1]) is metrizable, it is natural to describe the speed of convergence in weak functional limit theorems by means of an appropriate metric. Using the metric proposed by Prokhorov it is shown that under suitable conditions the rate of convergence in the functional central limit theorem for C-valued partial sum processes based on martingale difference arrays is the same as in the special case of row-wise independent random variables where this rate is known to be an optimal one.     

Local convergence in a generalized Fermat-Weber problem

In the Fermat-Weber problem, the location of a source point in ^N is sought which minimizes the sum of weighted Euclidean distances to a set of destinations. A classical iterative algorithm known as the Weiszfeld procedure is used to find the optimal location. Kuhn proves global convergence except for a denumerable set of starting points, while Katz provides local convergence results for this algorithm. In this paper, we consider a generalized version of the Fermat-Weber problem, where distances are measured by anl _p norm and the parameterp takes on a value in the closed interval [1, 2]. This permits the choice of a continuum of distance measures from rectangular (p=1) to Euclidean (p=2). An extended version of the Weiszfeld procedure is presented and local convergence results obtained for the generalized problem. Linear asymptotic convergence rates are typically observed. However, in special cases where the optimal solution occurs at a singular point of the iteration functions, this rate can vary from sublinear to quadratic. It is also shown that for sufficiently large values ofp exceeding 2, convergence of the Weiszfeld algorithm will not occur in general.     

An exact FCFS waiting time analysis for a general class of G/G/s queueing systems

A closed form expression for the waiting time distribution under FCFS is derived for the queueing system MGE_k/MGE_m/s, where MGE_n is the class of mixed generalized Erlang probability density functions (pdfs) of ordern, which is a subset of the Coxian pdfs that have rational Laplace transform. Using the calculus of difference equations and based on previous results of the author, it is proved that the waiting time distribution is of the form 1- , under the assumption that the rootsU _j are distinct, i.e. belongs to the Coxian class of distributions of order . The present approach offers qualitative insight by providing exact and asymptotic expressions, generalizes and unifies the well known theories developed for the G/G/1,G/M/s systems and leads to an algorithm, which is polynomial if only one of the parameterss orm varies, and is exponential if both parameters vary. As an example, numerical results for the waiting time distribution of the MGE₂/MGE₂/s queueing system are presented.     

Gap, Excess and Bornological Convergence

Let be the nonempty subsets of a metric space 〈X, d〉. Some classical convergences in - such as convergence in Hausdorff distance, Attouch-Wets convergence and Wijsman convergence - have been shown to be compatible with the weak topology on induced by all gap and excess functionals with fixed left argument ranging in some bornology. Here we consider an arbitrary ideal of subsets of X and compare the gap and excess topology so generated with the corresponding convergence defined in terms of truncations by elements of the ideal. Dedicated to the memory of Flora Daniel.     

Fully discrete approximations of parabolic boundary-value problems with nonsmooth boundary data

We study a numerical scheme for the approximation of parabolic boundary-value problems with nonsmooth boundary data. This fully discrete scheme requires no boundary constraints on the approximating elements. Our principal result is the derivation of optimal convergence estimates in L_p[0,T; L₂()] norms for boundary data in L_p[0, T; L₂()], 1p . For the same algorithms, we also show that the convergence remains optimal even in higher norms. The techniques employed are based on the theory of analytic semigroups combined with singular integrals.This paper was written in 1990, when the author was in the Department of Mathematical Sciences, University of Cincinnati. A preliminary version of this research was presented at the SIAM Annual Meeting in July 1989.     

On the logistic midrange

Summary It is well-known that for a large family of distributions, the sample midrange is asymptotically logistic. In this article, the logistic midrange is closely examined. Its distribution function is derived using Dixon's formula (Bailey (1935,Generalized Hypergeometric Series, Cambridge University Press, p. 13)) for the generalized hypergeometric function with unit argument, together with appropriate techniques for the inversion of (bilateral) Laplace transforms. Several relationships in distribution are established between the midrange and sample median of the logistic and Laplace random variables. Possible applications in testing for outliers are also discussed.     

Spectral Convergence of Quasi-One-Dimensional Spaces

We consider a family of non-compact manifolds X_ε (“graph-like manifolds”) approaching a metric graph X₀ and establish convergence results of the related natural operators, namely the (Neumann) Laplacian and the generalized Neumann (Kirchhoff) Laplacian on the metric graph. In particular, we show the norm convergence of the resolvents, spectral projections and eigenfunctions. As a consequence, the essential and the discrete spectrum converge as well. Neither the manifolds nor the metric graph need to be compact, we only need some natural uniformity assumptions. We provide examples of manifolds having spectral gaps in the essential spectrum, discrete eigenvalues in the gaps or even manifolds approaching a fractal spectrum. The convergence results will be given in a completely abstract setting dealing with operators acting in different spaces, applicable also in other geometric situations. Communicated by Claude Alain Pillet Submitted: December 21, 2005 Accepted: January 30, 2006     

Differentiability of Inverse Operators and Limit Theorems for Inverse Functions

Let H be a map from a set SR ^d to R ^d . For tR ^d let _H (t) denote the distance from t to the set H(S). Consider sequences {s _n} _n1 in S such that . Any limit point of any such sequence (finite or infinite) is considered as a possible value of the inverse H ^–1(t). Any map defined in such a way will be called an SC-inverse (a selected closest inverse) to H. In the paper we study differentiability of the nonlinear operator at H=G, where G is a one-to-one map from S onto a set TR ^d with good analytic properties (specifically, a diffeomorphism). We establish compact differentiability of this operator tangentially to continuous functions and introduce a family of norms such that it is Fréchet differentiable with respect to them. We also obtain optimal bounds for the remainder of the differentiation, extending to the multivariate case recent results of Dudley. These differentiability results are applied to random maps , which could be statistical estimators of an unknown map G. For a function J on R ^d , let (J) _T be its restriction to T. It is shown that for a diffeomorphism G and for an increasing sequence of positive numbers {a _n } _n1 weak convergence of the sequence {a _n (G _n – G)} _n1 (locally in S) is equivalent to weak convergence of the sequence (locally in T) along with the convergence of the sequence to 0 in probability (locally uniformly in S). The equivalence holds for all SC-inverses and all double SC-inverses and it extends to the multivariate case a theorem of Vervaat. Moreover, each of these equivalent statements implies a kind of Taylor expansion of the SC-inverse at G (locally uniformly in T) where inv(A) denotes the inverse of a nonsingular linear transformation A in R ^d . Such limit theorems for functional inverses can be used to study asymptotic behavior of statistical estimators defined implicitly (as solutions of equations involving the empirical distribution P _n ). We show how to apply this approach to get asymptotic normality of M-estimators in the multivariate case under minimal assumptions. We consider an extension of the quantile function to the multivariate case related to M-parameters of a distribution P in R ^d (an M-quantile function)and use limit theorems for functional inverses to study limit behavior of the empirical M-quantile process. We also show how to use these theorems to study asymptotics of regression quantiles.     

Characterization of infinitely divisible multivariate gamma distributions

A particular class of p-dimensional exponential distributions have Laplace transforms |I + VT|^?1, V positive definite or positive semi-definite and T = diagonal (t₁,…, t_p). A characterization is given of when these Laplace transforms are infinitely divisible.     

Two-step and three-step Q-superlinear convergence of SQP methods

This paper investigates the convergence rates of the variable-multiplier pair (x, ) in sequential quadratic programming methods for equality constrained optimization. The two main results of the paper are that the Q-superlinear convergence of {x _k } implies two-step Q-superlinear convergence for {(x _k , _k )} and that the two-step Q-superlinear convergence of {x _k } implies three-step Q-superlinear convergence for {(x _k , _k )}.The author is indebted to Professor Richard Tapia for many helpful comments and suggestions on the paper. The comments by Professors A. R. Conn and N. I. M. Gould on an earlier version are also acknowledged. This research was funded by SERC and ESRC research contracts.     

Alternative Approaches to the Two-Scale Convergence

Two-scale convergence is a special weak convergence used in homogenization theory. Besides the original definition by Nguetseng and Allaire two alternative definitions are introduced and compared. They enable us to weaken requirements on the admissibility of test functions (x, y). Properties and examples are added.     

On Asymptotic Proximity of Distributions

We consider some general facts concerning the convergence

Weak convergence using higher-order cumulants

Denote byc _j (F) thejth cumulant (or semi-invariant) of the distribution functionF. We say thatF is specified by its higher-order cumulants if it is the unique distribution functionG having the following property: there exists a positive integerJ such thatc _j (G)=c _j (F) forj=1,2 andjJ. Let (F _n n1) be a sequence of distribution functions, and suppose that there existsJ such thatc _j (F _n )c _j (F) asn, forj=1,2 andjJ. It is proved thatF _n F so long asF is specified by its higher-order cumulants. It is an open problem to characterize the family of distributions which are specified by their higher-order cumulants.     

Some limit theorems for the empirical process indexed by functions

Summary Let,n1, be a sequence of classes of real-valued measurable functions defined on a probability space (S,,P). Under weak metric entropy conditions on,n1, and under growth conditions on we show that there are non-zero numerical constantsC ₁ andC ₂ such that where (n) is a non-decreasing function ofn related to the metric entropy of. A few applications of this general result are considered: we obtain a.s. rates of uniform convergence for the empirical process indexed by intervals as well as a.s. rates of uniform convergence for the empirical characteristic function over expanding intervals.Portions of this article were presented during the conference on Mathematical Stochastics (February 19–25, 1984) at Oberwolfach, West Germany