Integral representation with respect to stopped continuous local martingales

We consider a Markovian jump process θ, with finite state space, feeding the parameters of a nonlinear diffusion process X. We observe θ and X in white noise, and—given a function f—we want to construct a finite filter for the f(X _t )-process. An algorithm is investigated which will produce a finite filter if it halts after a finite number of steps, and we give necessary and. sufficient conditions for this to happen.     

The Equilibrium Behavior of Reversible Coagulation-Fragmentation Processes

The coagulation-fragmentation process models the stochastic evolution of a population of N particles distributed into groups of different sizes that coagulate and fragment at given rates. The process arises in a variety of contexts and has been intensively studied for a long time. As a result, different approximations to the model were suggested. Our paper deals with the exact model which is viewed as a time-homogeneous interacting particle system on the state space _N, the set of all partitions of N. We obtain the stationary distribution (invariant measure) on _N for the whole class of reversible coagulation-fragmentation processes, and derive explicit expressions for important functionals of this measure, in particular, the expected numbers of groups of all sizes at the steady state. We also establish a characterization of the transition rates that guarantee the reversibility of the process. Finally, we make a comparative study of our exact solution and the approximation given by the steady-state solution of the coagulation-fragmentation integral equation, which is known in the literature. We show that in some cases the latter approximation can considerably deviate from the exact solution.     

Fragmentation Processes with an Initial Mass Converging to Infinity

We consider a family of fragmentation processes where the rate at which a particle splits is proportional to a function of its mass. Let F ₁^(m)(t),F ₂^(m)(t),… denote the decreasing rearrangement of the masses present at time t in a such process, starting from an initial mass m. Let then m→∞. Under an assumption of regular variation type on the dynamics of the fragmentation, we prove that the sequence (F ₂^(m),F ₃^(m),…) converges in distribution, with respect to the Skorohod topology, to a fragmentation with immigration process. This holds jointly with the convergence of m−F ₁^(m) to a stable subordinator. A continuum random tree counterpart of this result is also given: the continuum random tree describing the genealogy of a self-similar fragmentation satisfying the required assumption and starting from a mass converging to ∞ will converge to a tree with a spine coding a fragmentation with immigration. Research supported in part by EPSRC GR/T26368.     

Transition between Airy1 and Airy2 processes and TASEP fluctuations

We consider the totally asymmetric simple exclusion process, a model in the KPZ universality class. We focus on the fluctuations of particle positions, starting with certain deterministic initial conditions. For large time t, one has regions with constant and linearly decreasing density. The fluctuations on these two regions are given by the Airy₁ and Airy₂ processes, whose one‐point distributions are the GOE and GUE Tracy‐Widom distributions of random matrix theory. In this paper we analyze the transition region between these two regimes and obtain the transition process. Its one‐point distribution is a new interpolation between GOE and GUE edge distributions. © 2007 Wiley Periodicals, Inc.     

A system of Markov processes with random lifetimes

Summary Let E be a locally compact Hausdorff space with a countable base, and suppose {x_n} is a countable collection of points in E. Particles enter E at the site x _n according to a Poisson process N _n (t). Upon entrance to E, a typical particle moves through the space, independently of all other particles, according to the transition law of a Markov process, until its death, which occurs at some random time D. We prove several limit theorems for various functional of this infinite particle system. In particular, laws of large numbers, and central limit theorems are proved for occupation times of relatively compact Borel sets.Supported in part by Arizona State University Grant-in-Aid     

The central limit theorem for weighted empirical processes indexed by sets

Sufficient conditions are found for the weak convergence of a weighted empirical process {(ν_n(C)/q(P(C))) 1 _{[P(C) λn]}: C }, indexed by a class of sets and weighted by a function q of the size of each set. We find those functions q which allow weak convergence to a sample-continuous Gaussian process, and, given q, determine the fastest rate at which one may allow λ_n → 0.     

Un modéle d'lnspections optimales

We consider the following model: we inspect the motion of a Markov process with which an “evolution cost” is associated. We inspect the process at times T ₁…, T _n ,…. If when we inspect, its value is in a given set A, it continues its evolution, otherwise we kill it. At each inspection we associate an "inspection cost" and a "killing cost". The problem consists of finding a sequence of optimal inspections. After the modelization we construct the value function by an iterative procedure as in impulse control theory, by using the theory of analytic functions and theorems of section. Thanks to the criteria of optimality we get a sequence of optimal inspections under very general hypotheses.     

Continuous sampling plans for Markov-dependent production processes under limited inspection capacity

A continuous sampling plan is a set of rules that provide a given Average Outgoing Quality (AOQ), ideally with the minimum of effort (as measured by the Average Fraction Inspected, or AFI). Most such plans are based on the assumption that the quality (either defective or not) of successive production units is uncorrelated. In this paper, we explore the impact of correlation in the production process on the design of a sampling plan when it is not possible to inspect long runs of production unit-by-unit. We shall generalize Dodge's continuous sampling plan on two counts, replacing Level 1 100% inspection by 100f_o% inspection, and considering the production process to be Markov dependent instead of consisting of independent Bernoulli trials. We derive formulae for the AOQ and AFI, and consider how best to choose the sampling plan parameters in the presence of nonzero correlation.     

On the maximum deviation in random walks

We consider a random walk with drift to the left. LetM _n denote the extreme position to the right of the particle during its firstn steps. An approximate expression for the characteristic function of the distribution of this random variable is evaluated. The numerical inversion of this characteristic function is performed with the aid of the Fast Fourier Transform.     

Asymptotic properties of some goodness-of-fit tests based on the L 1-norm

Some goodness-of-fit tests based on the L ₁-norm are considered. The asymptotic distribution of each statistic under the null hypothesis is the distribution of the L ₁-norm of the standard Wiener process on [0,1]. The distribution function, the density function and a table of some percentage points of the distribution are given. A result for the asymptotic tail probability of the L ₁-norm of a Gaussian process is also obtained. The result is useful for giving the approximate Bahadur efficiency of the test statistics whose asymptotic distributions are represented as the L ₁-norms of Gaussian processes.     

Asymptotic behaviour of time evolutions of infinite particle systems

Summary Motions of one-dimensional infinite particle systems are considered where the dynamics is given by systems of ordinary differential equations of first order. The aim of the paper is to show that under certain assumptions about the system of differential equations the distribution law P _tof the particle system at time t becomes more and more regular under the influence of such an interaction. Moreover, P _tis tending weakly toward a distribution describing a random particle system with equal successive spacings.     

A Globally Convergent Algorithm for a PDE-Constrained Optimization Problem Arising in Electrical Impedance Tomography

We study the convergence properties of an algorithm for the inverse problem of electrical impedance tomography, which can be reduced to a partial differential equation (PDE) constrained optimization problem. The direct problem consists of the potential equation div(??u) = 0 in a circle, with Neumann condition describing the behavior of the electrostatic potential in a medium with conductivity given by the function ?(x, y). We suppose that at each time a current ψ _i is applied to the boundary of the circle (Neumann's data), and that it is possible to measure the corresponding potential ? _i (Dirichlet data). The inverse problem is to find ?(x, y) given a finite number of Cauchy pairs measurements (? _i , ψ _i ), i = 1,…, N. The problem is formulated as a least squares problem, and the developed algorithm solves the continuous problem using descent iterations in its corresponding finite element approximations. Wolfe's conditions are used to ensure the global convergence of the optimization algorithm for the continuous problem. Although exact data are assumed, measurement errors in data and regularization methods shall be considered in a future work.     

Expansions for the multivariate chi-square distribution

Three classes of expansions for the distribution function of the χ_k²(d, R)-distribution are given, where k denotes the dimension, d the degree of freedom, and R the “accompanying correlation matrix.” The first class generalizes the orthogonal series with generalized Laguerre polynomials, originally given by Krishnamoorthy and Parthasarathy [12]. The second class contains always absolutely convergent representations of the distribution function by univariate chi-square distributions and the third class provides also the probabilities for any unbounded rectangular regions. In particular, simple formulas are given for the three-variate case including singular correlation matrices R, which simplify the computation of third order Bonferroni inequalities, e.g., for the tail probabilities of max{χ_i²|1 ≤ i ≤ k} (k > 3).     

Minimum nastiness curve fitting

The ‘nastiness’ of a function φ(x) is defined. We then discuss minimum nastiness interpolation to a set of given points (x _k, φ_k), as well as minimum nastiness curve fitting, where the given valuesφ _k have errorsδ _k.     

SDDP for some interstage dependent risk-averse problems and application to hydro-thermal planning

We consider interstage dependent stochastic linear programs where both the random right-hand side and the model of the underlying stochastic process have a special structure. Namely, for equality constraints (resp. inequality constraints) the right-hand side is an affine function (resp. a given function b _t ) of the process value for the current time step t. As for m-th component of the process at time step t, it depends on previous values of the process through a function h _tm . For this type of problem, to obtain an approximate policy under some assumptions for functions b _t and h _tm , we detail a stochastic dual dynamic programming algorithm. Our analysis includes some enhancements of this algorithm such as the definition of a state vector of minimal size, the computation of feasibility cuts without the assumption of relatively complete recourse, as well as efficient formulas for sharing optimality and feasibility cuts between nodes of the same stage. The algorithm is given for both a non-risk-averse and a risk-averse model. We finally provide preliminary results comparing the performances of the recourse functions corresponding to these two models for a real-life application.     

Markov process with creation and annihilation

Summary A simple way is given to construct probabilistically a strong Markov process (y _i, P _x, of an extended sense such that the expectation u = E_x[f(y _t)] provides the solution of (1.1) for a Borel measurable function c (i.e. with both terms of creation and annihilation of mass), where and stand for the dates of birth and death of a particle, respectively. Actually the (non-probability) measure P _x is given as the sum of induced measures of probability measure (of Brownian motion with age) by a set of mappings. The probability measure is obtained by making killed processes of Brownian motion and piecing them together in a specific way.     

A globally convergent method for solving nonlinear equations without the differentiability condition

We give a general iterative method which computes the maximal real rootx _max of a one variable Lipschitzian function in a given interval. The method generates a monotonically decreasing sequence which converges towardsx _max or demonstrates the non-existence of a real root in the considered interval. We show that the method is globally convergent and locally linearly convergent. We also compute the number of iterations needed to reach the given accuracy.     

Age-dependent branching processes in random environments

We consider an age-dependent branching process in random environments. The environments are represented by a stationary and ergodic sequence ξ = (ξ0,ξ1,...) of random variables. Given an environment ξ, the process is a non-homogenous Galton-Watson process, whose particles in n-th generation have a life length distribution G(ξn) on R , and reproduce independently new particles according to a probability law p(ξn) on N. Let Z(t) be the number of particles alive at time t. We first find a characterization of the conditional probability generating function of Z(t) (given the environment ξ) via a functional equation, and obtain a criterion for almost certain extinction of the process by comparing it with an embedded Galton-Watson process. We then get expressions of the conditional mean EξZ(t) and the global mean EZ(t), and show their exponential growth rates by studying a renewal equation in random environments.     

A free boundary problem for three-dimensional harmonic vector fields

We study in this paper a free boundary value problem ( FB ), where a region G_o in R ³ is determined by the condition that there exists a vector field v_o in G_o which satisfies div v_o = e_o, curl v_o = g_o in G_o and v_o = E on the boundary ?G_o with a given scalar function e_o and given vector fields g_o and E. We give two equivalent formulations for this problem. Then we characterize the solutions by a non-linear integral equation. In order to solve the latter by a Newton method we linearize this equation. We investigate the ensuing linear integral equation. In case of axisymmetric configurations this is a singular integral equation whose index can be easily determined from the given data. We obtain a related equation, if we try to construct a field v in a region G which is on the boundary perpendicular to a given field B . Finally we use this method to investigate an astrophysical problem, which arises in the theory of pulsar magnetospheres.