A sequential sampling rule for selecting the most probable multinomial event

Summary A sequential sampling rule is given for selecting the most probable event from a multinomial distribution withk cells. A random number of observations is taken from the given multinomial distribution at each stage of sampling, where the number is distributed according to a Poisson distribution with mean λ. The sampling is stopped when the count in any cell is greater than or equal to a given positive integerN. The cell with the highest count is selected for the most probable event. The mathematical analysis of the problem is simplified as a result of the statistical independence of the cell frequencies due to the randomization of the sample number. The expected value of the stage when the sampling terminates is decreasing in λ. The sequential sampling scheme in which one observation is taken at a time until the highest cell count is equal toN, corresponds to λ→0. A table is given showing some properties of the given selection procedure.     

Wigner quasi-probability densities and characterization of the Gaussian distribution

A new characterization of the Gaussian distribution was given by R. Hudson in 1974. A sharpening and some extensions of this result are proposed, and connections with other characterizations of the Gaussian distribution are given. Linnik's “ridge principle” is substantially used. Bibliography: 11 titles. Translated fromZapiski Nauchnykh Seminarov POMI, Vol. 228, 1996, pp. 142–153.     

Uniform Rate of Weak Convergence of the Minimum Contrast Estimator in the Ornstein–Uhlenbeck Process

The paper shows that the distribution of the normalized minimum contrast estimator of the drift parameter in the Ornstein–Uhlenbeck process observed over [0, T] converges to the standard normal distribution with an uniform error rate of the order O (T ^− 1/2). A precise estimate of the constant in the upper bound is also given.     

Fix-mahonian calculus III; a quadruple distribution

A four-variable distribution on permutations is derived, with two dual combinatorial interpretations. The first one includes the number of fixed points “fix”, the second the so-called “pix” statistic. This shows that the duality between derangements and desarrangements can be extended to the case of multivariable statistics. Several specializations are obtained, including the joint distribution of (des, exc), where “des” and “exc” stand for the number of descents and excedances, respectively. Authors’ addresses: Dominique Foata, Institut Lothaire, 1 rue Murner, F-67000 Strasbourg, France; Guo-Niu Han, Center for Combinatorics, LPMC, Nankai University, Tianjin 300071, P.R. China; I.R.M.A. UMR 7501, Université Louis Pasteur et CNRS, 7, rue René-Descartes, F-67084 Strasbourg, France     

On an analog of empirical distribution for multivariate censored data

A model of a “censored” experiment is formally defined. A functional is proposed which can be considered as an analog of the empirical distribution. Some properties of this functional are investigated. Translated fromStatisticheskie Metody Otsenivaniya i Proverki Gipotez, pp. 41–51, Perm, 1993.     

On the asymptotic independence of Dirichlet series

The joint limit distribution of functions given by Dirichlet series is studied. The necessary and sufficient condition when this distribution is a product of marginal distributions is found. An example of such Dirichlet series with linear independent systems of exponents is presented. Partially supported by the Lithuanian State Science and Studies Foundation. Vilnius University, Naugarduko 24, 2006 Vilnius; Šiauliai University, P. Višinskio 25, 5419 Šiauliai, Lithuania. Translated from Lietuvos Matematikos Rinkinys, Vol. 39, No. 1, pp. 65–73, January–March, 1999. Translated by A. Laurinčikas     

Heavy-traffic analysis for the GI/G/1 queue with heavy-tailed distributions

We consider a GI/G/1 queue in which the service time distribution and/or the interarrival time distribution has a heavy tail, i.e., a tail behaviour like t ^−ν with 1 < ν ⩽ 2 , so that the mean is finite but the variance is infinite. We prove a heavy-traffic limit theorem for the distribution of the stationary actual waiting time W. If the tail of the service time distribution is heavier than that of the interarrival time distribution, and the traffic load a → 1, then W, multiplied by an appropriate ‘coefficient of contraction’ that is a function of a, converges in distribution to the Kovalenko distribution. If the tail of the interarrival time distribution is heavier than that of the service time distribution, and the traffic load a → 1, then W, multiplied by another appropriate ‘coefficient of contraction’ that is a function of a, converges in distribution to the negative exponential distribution. This revised version was published online in June 2006 with corrections to the Cover Date.     

On weight distributions of perfect colorings and completely regular codes

A vertex coloring of a graph is called “perfect” if for any two colors a and b, the number of the color-b neighbors of a color-a vertex x does not depend on the choice of x, that is, depends only on a and b (the corresponding partition of the vertex set is known as “equitable”). A set of vertices is called “completely regular” if the coloring according to the distance from this set is perfect. By the “weight distribution” of some coloring with respect to some set we mean the information about the number of vertices of every color at every distance from the set. We study the weight distribution of a perfect coloring (equitable partition) of a graph with respect to a completely regular set (in particular, with respect to a vertex if the graph is distance-regular). We show how to compute this distribution by the knowledge of the color composition over the set. For some partial cases of completely regular sets, we derive explicit formulas of weight distributions. Since any (other) completely regular set itself generates a perfect coloring, this gives universal formulas for calculating the weight distribution of any completely regular set from its parameters. In the case of Hamming graphs, we prove a very simple formula for the weight enumerator of an arbitrary perfect coloring.     

Sensitivity of Bayes decisions for the success probability for samples from a nonbinomial population

Summary A family of generalised negative binomial distributions is employed to investigate inference robustness of the Bayes estimator of the unknown parameter of the binomial distribution. A zone of sensitivity for the test of significance is constructed to forewarn the pro-Jeffreys Bayesians against indiscriminate choice of the probability in favour of the null hypothesis. A few selected tables are presented to illustrate the effect of relaxation of the ‘binomiality’ assumption.     

Tail Calculus with Remainder,Applications to Tail Expansions for Infinite Order Moving Averages,Randomly Stopped Sums,and Related Topics

We derive asymptotic expansions for tails of infinite weighted convolutions of some heavy-tailed distributions. Applications are given to tail expansion of the marginal distribution of ARMA processes, randomly stopped sums, as well as limiting waiting time distribution. AMS 2000 Subject Classifications. Primary—62E99, Secondary—41A60, 44A35, 60G50, 60K25     

Method of primers in calculating the probability of destruction of a unidirectional composite material in tension

A method of primers is elaborated which allows one to calculate the distribution function of durability of a composite material in tension in the reinforcement direction. Integral and differential equations for calculating the probabilities of formation of primers and destruction of a material caused by their formation are presented. Distribution functions of material strength for the parameter of Weibull distribution of fiber strength on the interval 2.1 ≤ β _f ≤ 50.1 are calculated. From the functions, the average values and root-mean-square deviations of material strength are found. The results obtained agree well with calculations by using the structural-imitation simulation. The distribution functions of material strength with a high precision are approximated by the three-parameter Weibull distributions. The distribution parameters are approximated by the linear functions of ln (β_f). __________ Translated from Mekhanika Kompozitnykh Materialov, Vol. 43, No. 6, pp. 823–838, November–December, 2007.     

Principle of local energy minimum for donor-acceptor systems with low compensation

The charged states and structures of one-acceptor local complexes (OLC) are studied as functions of external potentials for donor-acceptor systems with low compensations K. The criterion of a two-acceptor local complex (TLC) is formulated. It is used to derive stochastic equations for self-consistent fluctuating potentials (FP) in the OLC ensemble and to find an expression for the Fermi level μ(K) of impurity electrons via the one-site distribution of FP. It is shown in the “smeared site” approximation that the asymptotic behavior of μ(K) for small K is determined by the average size of the TLC and by self-consistent screening effects in the OLC ensemble. Translated from Teoreticheskaya i Matematicheskaya Fizika, Vol. 111, No. 3, pp. 452–461, June, 1997.     

Bivariate inverse gaussian distribution

Summary A bivariate inverse Gaussian (IG) density function is constructed. Relations of the bivariate IG distribution to the normal and χ² distributions are established. The corresponding bivariate random walk (RW) density function is obtained. The properties and behaviour of bivariate IG distribution are studied for large parametric values. Moment estimates of the five parameters are given and applications are pointed out. A generalization to the multivariate IG distribution is proposed.     

A fuzzy goal programming and meta heuristic algorithms for solving integrated production: distribution planning problem

Integrated production–distribution planning is one of the most important issues in supply chain management (SCM). We consider a supply chain (SC) network to consist of a manufacturer, with multiple plants, products, distribution centers (DCs), retailers and customers. A multi-objective linear programming problem for integrating production–distribution, which considers various simultaneously conflicting objectives, is developed. The decision maker’s imprecise aspiration levels of goals are incorporated into the model using a fuzzy goal programming approach. Due to complexity of the considered problem we propose three meta-heuristics to tackle the problem. A simple genetic algorithm and a particle swarm optimization (PSO) algorithm with a new fitness function, and an improved hybrid genetic algorithm are developed. In order to show the efficiency of the proposed methods, two classes of problems are considered and their instances are solved using all methods. The obtained results show that the improved hybrid genetic algorithm gives us the best solutions in a reasonable computational time.     

Tail fit and the Zipf–Pareto law

A limit theorem with bounds on the rate of convergence is proven. The joint distribution of a fixed number of relative decrements of the top order statistics from a random sample converges to the limit as the sample size increases if and only if the underlying distribution is in essence a Pareto. In conjunction with a chi-square test of fit, it provides an asymptotically distribution-free test of fit to the family of distributions with regularly varying tails at infinity. When the limit distribution holds, rank-size plots obey Zipf’s law. The test can be used to detect departures from this Zipf–Pareto law.      

Distribution systems design with two-level routing considerations

In this study, we formulate and analyze a strategic design model for three-echelon distribution systems with two-level routing considerations. The key design decisions considered are: the number and locations of distribution centers (DC’s), which big clients (clients with larger demand) should be included in the first level routing (the routing between plants and DC’s), the first-level routing between plants, DC’s and big clients, and the second-level routing between DC’s and other clients not included in the first-level routing. A hybrid genetic algorithm embedded with a routing heuristic is developed to efficiently find near-optimal solutions. The quality of the solution to a series of small test problems is evaluated—by comparison with the optimal solution solved using LINGO 9.0. In test problems for which exact solutions are available, the heuristic solution is within 1% of optimal. At last, the model is applied to design a national finished goods distribution system for a Taiwan label-stock manufacturer. Through the case study, we find that the inclusion of big clients in the first-level routing in the analysis leads to a better network design in terms of total logistic costs.     

Iterative Approximation of Statistical Distributions and Relation to Information Geometry

The optimal control of stochastic processes through sensor estimation of probability density functions is given a geometric setting via information theory and the information metric. Information theory identifies the exponential distribution as the maximum entropy distribution if only the mean is known and the Γ distribution if also the mean logarithm is known. The surface representing Γ models has a natural Riemannian information metric. The exponential distributions form a one-dimensional subspace of the two-dimensional space of all Γ distributions, so we have an isometric embedding of the random model as a subspace of the Γ models. This geometry provides an appropriate structure on which to represent the dynamics of a process and algorithms to control it. This short paper presents a comparative study on the parameter estimation performance between the geodesic equation and the B-spline function approximations when they are used to optimize the parameters of the Γ family distributions. In this case, the B-spline functions are first used to approximate the Γ probability density function on a fixed length interval; then the coefficients of the approximation are related, through mean and variance calculations, to the two parameters (i.e. μ and β) in Γ distributions. A gradient based parameter tuning method has been used to produce the trajectories for (μ, β) when B-spline functions are used, and desired results have been obtained which are comparable to the trajectories obtained from the geodesic equation. This revised version was published online in August 2006 with corrections to the Cover Date.     

Timeout control scheme for overloaded M/E2/1 queue

A timeout scheme is considered for controlling an infinite ‘firstcome, first served’ overloaded single-server queue. Inthe overload situation, a customer-rejection mechanism is usedfor timing out ‘older’ customers in the queue, i.e.excluding those who have waited longer than a certain time.Applying ‘level-crossing analysis’ to an M/E₂/1queue, exact analytic expressions of performance such as thedensity and distribution functions of waiting time of the customerswho get served, the mean delay of customers, successful throughput,and ‘goodput’ are determined for this queue.