Search

Much of the research on network modeling and analysis has focused on the design of end controllers and network algorithms with the objective of stability and convergence of the transmission rate. However, a network is typically composed of a mixture of both controlled elastic flows and uncontrolled real-time flows. In this paper, we study the effects of marking elasticity (which characterizes how aggressively the marking function responds to congestion) on queue overflow probability for uncontrolled real-time flows. First, we derive lower and upper bounds on the queue overflow probability at a router of a single bottleneck system. Using this, we quantify the trade-off between stability for controlled flows and queue overflow probability for uncontrolled real-time flows as a function of marking elasticity. Next, we compare the capacity required at a router with only FIFO scheduling versus a router with priority scheduling (priority given to the real-time flows) for supporting a given queue overflow probability. We quantify the "scheduling-gain" of priority scheduling over FIFO scheduling, as a function of marking elasticity. We show that this scheduling gain decreases with more elastic marking functions. Index Terms--Congestion control, queueing theory, stochastic processes.

In this paper, we propose a powerful trellis shaping for peak-to-average power ratio (PAR) reduction of pulse-shaped phase shift keying (PSK) systems. The proposed approach consists of 1) memory extension of trellis shaping encoder for capturing the full effect of pulse shaping filter impulse response, and 2) simple branch metrics that quantify envelope fluctuation of the filter output. Simulation results demonstrate that achieving near constant envelope is possible even if a pulse shaping filter with a roll-off factor as low as 0.1 is employed. Furthermore, due to the circular nature of the resulting signal trajectory, the proposed system also exhibits a robustness against sampling timing jitter at the receiver. Index Terms--Band-limited single-carrier communications, peak-to-average power ratio reduction, trellis shaping.

This technical note deals with multi-agent networks performing estimation tasks. Consider a network of mobile agents with sensors that can take measurements of a spatial stochastic process. Using the kriging statistical technique, a field estimate may be calculated over the environment, with an associated error variance at each point. We study a single-snapshot scenario, in which the spatial process mean is known and each agent can only take one measurement. We consider two optimization problems with respect to the measurement locations, using as objective functions the maximum error variance and the extended prediction variance. As the correlation between distinct locations vanishes, we show that circumcenter and incenter Voronoi configurations become network configurations that optimize the maximum error variance and the extended prediction variance, respectively. We also present distributed coordination algorithms that steer the network towards these configurations. Index Terms--Linear unbiased minimum variance estimator (LUMVE).

This paper is concerned with persistent identification of systems that involve deterministic unmodeled dynamics and stochastic observation disturbances, and whose unknown parameters switch values (possibly large jumps) that can be represented by a Markov chain. Two classes of problems are considered. In the first class, the switching parameters are stochastic processes modeled by irreducible and aperiodic Markov chains with transition rates sufficiently faster than adaptation rates of the identification algorithms. In this case, tracking real-time parameters by output observations becomes impossible and we show that an averaged behavior of the parameter process can be derived from the stationary measure of the Markov chain and can be estimated with periodic inputs and least-squares type algorithms. Upper and lower error bounds are established that explicitly show impact of unmodeled dynamics. In contrast, the second class of problems represents systems whose state transitions occur infrequently. An adaptive algorithm with variable step sizes is introduced for tracking the time-varying parameters. Convergence and error bounds are derived. Numerical results are presented to illustrate the performance of the algorithm. Index Terms--Error bounds, exogenous noise, parameter switching, parameter tracking, persistent identification, unmodeled dynamics.

In commonly used functional regression models, the regression of a scalar or functional response on the functional predictor is assumed to be linear. This means that the response is a linear function of the functional principal component scores of the predictor process. We relax the linearity assumption and propose to replace it by an additive structure, leading to a more widely applicable and much more flexible framework for functional regression models. The proposed functional additive regression models are suitable for both scalar and functional responses. The regularization needed for effective estimation of the regression parameter function is implemented through a projection on the eigenbasis of the covariance operator of the functional components in the model. The use of functional principal components in an additive rather than linear way leads to substantial broadening of the scope of functional regression models and emerges as a natural approach, because the uncorrelatedness of the functional principal components is shown to lead to a straightforward implementation of the functional additive model, based solely on a sequence of one-dimensional smoothing steps and without the need for backfilling. This facilitates the theoretical analysis, and we establish the asymptotic consistency of the estimates of the components of the functional additive model. We illustrate the empirical performance of the proposed modeling framework and estimation methods through simulation studies and in applications to gene expression time course data. KEY WORDS: Additive model: Asymptotics: Functional data analysis: Functional regression: Linear model; Principal components: Smoothing; Stochastic processes.

Temporal regulation of origin activation is widely thought to explain the pattern of early- and late-replicating domains in the Saccharomyces cerevisiae genome. Recently, single-molecule analysis of replication suggested that stochastic processes acting on origins with different probabilities of activation could generate the observed kinetics of replication without requiring an underlying temporal order. To distinguish between these possibilities, we examined a clb5[DELTA] strain, where origin firing is largely limited to the first half of S phase, to ask whether all origins nonspecifically show decreased firing (as expected for disordered firing) or if only some origins ('late' origins) are affected. Approximately half the origins in the mutant genome show delayed replication while the remainder replicate largely on time. The delayed regions can encompass hundreds of kilobases and generally correspond to regions that replicate late in wild-type cells. Kinetic analysis of replication in wild-type cells reveals broad windows of origin firing for both early and late origins. Our results are consistent with a temporal model in which origins can show some heterogeneity in both time and probability of origin firing, but clustering of temporally like origins nevertheless yields a genome that is organized into blocks showing different replication times.

The well-known cumulative sum (CUSUM) sequential rule for abrupt model change detection in stochastic dynamic systems relies on the knowledge of the probability density functions of the system output variables conditional on their past values and on the system functioning mode at each time step. This paper shows how to build an asymptotically optimal detection rule under the common average run length (ARL) constraint when these densities are not available but can be consistently estimated. This is the case for nonlinear state--space systems observed through output variables: for such systems, a new class of particle filters based on convolution kernels allows to get consistent estimates of the conditional densities, leading to an optimal CUSUM-like filter detection rule (FDR). Index Terms--Average run length (ARL) constraint, convolution kernel filter, cumulative sum (CUSUM) rule, model change detection, particle filter, state--space systems.

Internet traffic exhibits self-similarity and long-range dependence (LRD) on various time scales. In this paper, we propose to use the Modified Allan Variance (MAVAR) and a Modified Hadamard Variance (MHVAR) to estimate the Hurst parameter H of the LRD traffic series or, more generally, the exponent [alpha] of data with 1/[f.sup.[alpha]]([alpha] [greater than or equal to] 0) power-law spectrum. MHVAR generalizes the principle of MAVAR, a time-domain quantity widely used for frequency stability characterization, to higher-order differences of input data. In our knowledge, this MHVAR has been mentioned in literature only few times and with little detail so far. The behaviour of MAVAR and MHVAR with power-law random processes and some common deter-ministic signals (viz. drifts, sine waves, steps) is studied by analysis and simulation. The MAVAR and MHVAR accuracy in estimating H is evaluated and compared to that of wavelet Logscale Diagram (LD). Extensive simulations show that MAVAR and MHVAR achieve significantly better confidence and no bias in H estimation. Moreover, MAVAR and MHVAR feature a number of other advantages, which make them valuable to complement other established techniques such as LD. Finally, MHVAR and LD are also applied to a real IP traffic trace. Index Terms--Communication system traffic, fractals, fractional noise, Internet, long-range dependence, serf-similarity, time domain analysis, wavelet transforms.

A noniterative implementation of the maximum likelihood (ML) estimator for the stochastic signal model is presented. The procedure based on Monte Carlo importance sampling allows estimating direction of arrival (DOA) at low angular separation and outperforms the iterative approaches for closely spaced narrow-band sources.

A study was conducted to examine the phenomenon of noise enhanced systems for a general parameter estimation problem and compare performance between the original estimators and noise enhanced estimators based on different criteria.

A study was conducted to examine the mechanism of the stochastic resonance (SR) effect for a fixed detector and also gain insight into the detector optimization problem with SR noise under both Neyman-Pearson and Bayesian criteria.

A study was conducted to explore the use of the variational Bayes (VB) approximation in Bayesian filtering to accelerate marginalized particle filtering and as a deterministic local (one-step) approximation. The performance of the various VB filtering schemes in relation to Gaussian model with a nonlinear substate and a hidden Markov model is also reviewed.

Several stochastic stability analyses are conducted to derive an easy-to-compute approximate bound for the range of step-sizes for the stable constant-modulus algorithm (CMA). The algorithm is shown to be more robust for smaller variation in the modulus of the transmitted symbols, exhibiting smaller steady-state misadjustment.

The stochastic nuclear quadrupole resonance (sNQR) data model is introduced and its benefits as compared to conventional NQR are discussed. The results have shown that the detectors are constant false alarm rate (CFAR) with respect to the noise power, when radio frequency interference (RFI) is absent, and approximately CFAR with respect to the unknown interference, when RFI is present.

This paper investigates vehicle-routing problems in which the travel times are random variables, and deliveries are made subject to soft time-window constraints. In particular, we model the travel time using a shifted gamma distribution. Penalties are incurred for deviations from the customers' time windows--early or late--and are developed using a fixed cost, a linear cost penalty, and/or a quadratic loss penalty. Alternatively, specifying a given probability of meeting the time-window constraints is considered. A tabu-search metaheuristic is developed, and computational results on test problems from the literature are reported. Keywords: stochastic vehicle routing; metaheuristics; time-window constraints

The optimal full-order linear filter in the linear minimum variance sense is derived for discrete-time stochastic linear systems with multiple packet dropouts. A sufficient condition is given for the existence of the steady-state filter and the asymptotic stability of the optimal filter is analyzed.

Optimal filters are derived on the sphere in order to detect compact objects embedded in a stochastic background process. A naive detection strategy is adopted to show the application of the new filter theory and it is shown to perform well, even at low signal-to-noise ratio.

The execution of the distributed random grouping (DRG) algorithm is discussed for the computation of average aggregate on a sensor network with random changing graphs by stochastic hybrid systems (SHSs).

For determining the stochastic response and stability of a strongly nonlinear single-degree-of-freedom system using the stochastic averaging technique, the size of excitations should be small such that the response of the system converges weakly to a Markov process. This condition is not often met with practical problems, and therefore, application of this method for obtaining their responses becomes difficult. Further, for systems with nonlinearities that cannot be integrated in closed form, stability analysis by examining the conditions of the two boundaries of the problem is not possible. A semianalytical method along with a weighted residual technique is presented here to circumvent these difficulties and to determine the response and stability of a strongly nonlinear system subjected to sizable stochastic excitation. The weighted residual technique is employed to correct the errors in averaged drift and diffusion coefficients resulting due to the size of the stochastic excitation. Two example problems are solved as illustrations of the method. DOI: 10.1061/(ASCE)0733-9399(2008)134:8(698) CE Database subject headings: Stochastic processes; Stability; Nonlinear systems; Markov chains.

A q-parameterized deterministic annealing expectation maximization (EM) (q-DAEM) algorithm based on nonextensive statistical mechanics is presented. Findings reveal the superior performance of the proposed algorithm.

Two types of linear frequency transforms for signals with nonuniform sampling times are examined.

The boundedness behavior of the spectral factorization mapping in the Wiener algebra is examined. Findings reveal that an increase in the degree of spectra worsens the boundedness constant of the spectral factorization mapping.

We analyze the problem of optimal location of a set of facilities in the presence of stochastic demand and congestion. Customers travel to the closest facility to obtain service; the [...]

An analysis of combined k-out-of-n, and consecutive [k.sub.c]-out-of-n systems is discussed. The reliabilities, lifetime distributions and mean lifetimes of the systems were calculated.

In this paper we consider the problem of estimating nonparametric panel data models with fixed effects. We introduce an iterative nonparametric kernel estimator. We also extend the estimation method to the case of a semiparametric partially linear fixed effects model. To determine whether a parametric, semiparametric or nonparametric model is appropriate, we propose test statistics to test between the three alternatives in practice. We further propose a test statistic for testing the null hypothesis of random effects against fixed effects in a nonparametric panel data regression model. Simulations are used to examine the finite sample performance of the proposed estimators and the test statistics. Keywords: Fixed effects models; Model specification tests; Nonparametric kernel method; Panel data; Partially linear model; Profile method; Random effects models; Semiparametric efficiency bound

We consider a singular stochastic control problem, which is called the monotone follower stochastic control problem, and give sufficient conditions for the existence and uniqueness of a local-time type optimal [...]

An independent-increments stochastic process is used to model the degradation path of a given device. An expectation-maximization (EM) algorithm for the model is developed to attain the maximum likelihood estimators .

This article proposes a class of asymptotically distribution-free specification tests for parametric conditional distributions. These tests are based on a martingale transfrom of a proper sequential empirical process of conditionally transformed data. Standard continuous functionals of this martingale provide omnibus tests while linear combinations of the orthogonal components in its spectral representation form a basis for directional tests. Finally, Neylnan-type smooth tests, a compromise between directional and omnibus tests, are discussed. As st special example we study in detail the construction of directional tests for the null hypothesis of conditional normality versus heteroskedastic contiguous alternatives. A small Monte Carlo study shows that our tests attain the nominal level already for small sample sizes. Keywords: Conditional models; Martingale transformation; Sequential empirical process; Specification tests

Constrained stochastic graph process (CSGP) is a novel analytical technique for the description of content distribution based on tree or meshes. Findings reveal the capability of the proposed technique in efficiently computing the mean and full distribution of metrics such as the file download times or the number of hops from the source to the receiving nodes.

This note is concerned with stability analysis and stabilization of randomly switched systems. These systems may be regarded as piecewise deterministic stochastic systems: the discrete switches are triggered by a stochastic process which is independent of the state of the system, and between two consecutive switching instants the dynamics are deterministic. Our results provide sufficient conditions for almost sure stability and stability in the mean using Lyapunov-based methods when individual subsystems are stable and a certain "slow switching" condition holds. This slow switching condition takes the form of an asymptotic upper bound on the probability mass function of the number of switches that occur between the initial and current time instants. This condition is shown to hold for switching signals coming from the states of finite-dimensional continuous-time Markov chains; our results, therefore, hold for Markovian jump systems in particular. For systems with control inputs, we provide explicit control schemes for feedback stabilization using the universal formula for stabilization of nonlinear systems. Index Terms--Almost sure and mean stochastic stability, random switching, stabilization, switched systems.

The development of space-time covariance functions capable of producing flexible space-time asymmetries is reported. It could be applied to a dataset of total column ozone levels in the Northern hemisphere.

The article studies an algorithm for simulating improper complex-valued second-order stationary stochastic process having specified second-order properties. Results suggest that the generated process obey the necessary distributional properties.

The stochastic maximum likelihood signal estimate properties are studied under a misspecified number of signals. Maximum likelihood estimates provide valuable information about true parameters.

This paper examines the relationship between dividend-based and earnings-based cost of equity capital estimates under temporary competitive advantage. A new insight is that the true cost of equity is shown to approach the geometric mean of the supernormal dividend-based and simple earnings-based estimates, as the mean period of competitive advantage approaches the supernormal payback period. We proceed to analyse the international shipping sector across 14 countries from 1998 to 2005. Although the mean cost of equity is around 10-11%, there is high variability and non-normality. The cost of equity measures under temporary competitive advantage provide estimates statistically independent of those based on stock market risk. doi: 10.1057/palgrave.jors.2602285 Keywords: finance; shipping; stochastic processes; competitive advantage

The spectral representation of stationary stochastic processes via the Karhunen-Loeve (KL) expansion is examined from a numerical efficiency perspective. Attention is focused on processes which have commonly been characterized by covariance kernels decaying exponentially versus the position/time delay variable. By introducing a slight modification in the mathematical description of this covariance kernel, the nondifferentiability at its peak is eliminated, whereas most of its advantageous properties are retained. It is shown that compared to the common exponential model, the requisite number of terms for representing the process in context with the modified kernel is significantly smaller. The effect is demonstrated by means of a specific numerical example. This is done by first determining the eigenfunctions/eigenvalues associated with the KL expansion for the modified kernel model, and by afterwards estimating the approximation errors corresponding to the two kernels considered for specific numerical values. Clearly, the enhanced computational efficiency of the KL expansion associated with the modified kernel can significantly expedite its incorporation in stochastic finite elements and other areas of stochastic mechanics. DOI: 10.1061/(ASCE)0733-9399(2007)133:7(773) CE Database subject headings: Stochastic processes; Spectral analysis; Monte Carlo method; Computer aided simulation; Numerical analysis.

Estimating the variance of the sample mean is a classical problem of steady-state simulation output analysis. Traditional batch means estimators require specification of the simulation run length a priori. To our knowledge, the Dynamic Non-overlapping Batch Means (DNBM) estimator is the only existing variance estimator that requires a constant storage space for any sample size. In this paper, we develop the Dynamic Partial-overlapping Batch Means (DPBM) algorithm, that also requires a constant storage space. In terms of the mean squared error, the statistical performance of the DPBM estimators is superior to that of the DNBM estimators. Keywords: Variance of the sample mean, batch means estimators, mean squared error, 1. Introduction In stochastic simulation experiments, output data [Y.sub.1], [Y.sub.2],... are collected to infer various measures of the performance of the system. Often the measure is a population mean E(Y) [...]

The problem of dynamically choosing which group of sensors should communicate with the access point (AP) to maximize network lifetime as a stochastic shortest path Markov decision process is formulated.

In this letter, we propose a systematic technique for optimum data-sequence design for data-aided (DA) estimation of synchronization parameters. It is assumed that the system operates in burst mode, transmitting a linearly modulated information signal over an additive white Gaussian noise channel. The method is based on the minimization of the corresponding Cramer--Rao lower bound subject to a power constraint and is applied in four practical receiver synchronization contexts: estimation of the symbol timing with known carrier frequency offset and either known or unknown carrier phase, joint estimation of the symbol timing and carrier phase with known carrier frequency offset, and joint estimation of the carrier phase and carrier frequency offset with known symbol timing. Index Terms--Cramer-Rao lower bound (CRLB), data-aided (DA) estimation, synchronization, training sequence design.