Sekhar Venkatraman

Bio: Sekhar Venkatraman is an academic researcher from Purdue University. The author has contributed to research in topics: Control variates & Covariance matrix. The author has an hindex of 4, co-authored 4 publications receiving 374 citations.

Least-squares estimation of distribution functions in johnson's translation system

Abstract: To summarize a set of data by a distribution function in Johnson's translation system, we use a least-squares approach to parameter estimation wherein we seek to minimize the distance between the vector of "uniformized" oeder statistics and the corresponding vector of expected values. We use the software package FITTRI to apply this technique to three problems arising respectively in medicine, applied statistics, and civil engineering. Compared to traditional methods of distribution fitting based on moment matching, percentile matchingL 1 estimation, and L ⌆ estimation, the least-squares technique is seen to yield fits of similar accuracy and to converge more rapidly and reliably to a set of acceptable parametre estimates.

Estimation procedures based on control variates with known covariance matrix

Abstract: This paper describes a new procedure for using control variates in multiresponse simulation when the covariance matrix of the controls is known. Assuming that the responses and the controls are jointly normal, we develop a new unbiased control-variates point estimator for the mean simulation response. We also compute the covariance matrix of this point estimator in order to construct an approximate confidence-region estimator for the mean response. If the covariances between the responses and the controls are unknown so that the optimal control coefficients must be estimated, then some of the potential efficiency improvement is lost. This loss is quantified in a new variance ratio. We summarize the results of an extensive experimental study in which we apply the proposed estimation procedure to closed queueing networks and stochastic activity networks.

Using path control variates in activity network simulation

Abstract: In the simulation of a stochastic activity network (SAN), the usual objective is to obtain point and confidence-interval estimators of the mean completion time for the network This paper presents a new procedure for using path control variates to improve the efficiency of such estimators Because each path control is the duration of an associated path in the network, the vector of selected path controls has both a known mean and a known covariance matrix All of this information is incorporated into point- and interval-estimation procedures for both normal and nonnormal responses To evaluate the performance of these procedures experimentally, we compare actual versus predicted reductions in point-estimator variance and confidence-interval half-length for a set of SANs in which the following characteristics are systematically varied: (a) the size of the network (number of nodes and activities); (b) the topology of the network; (c) the relative dominance (criticality index) of the critical path; and (d) the percentage of activities with exponentially distributed durations The experimental results indicate that large variance reductions can be achieved with these estimation procedures in a wide variety of networks

Generalized exponential distribution: different method of estimations

Abstract: Recently a new distribution, named as generalized exponential distribution has been introduced and studied quite extensively by the authors. Generalized exponential distribution can be used as an alternative to gamma or Weibull distribution in many situations. In a companion paper, the authors considered the maximum likelihood estimation of the different parameters of a generalized exponential distribution and discussed some of the testing of hypothesis problems. In this paper we mainly consider five other estimation procedures and compare their performances through numerical simulations.

Steady-state simulation of queueing processes: survey of problems and solutions

Abstract: For years computer-based stochastic simulation has been a commonly used tool in the performance evaluation of various systems. Unfortunately, the results of simulation studies quite often have little credibility, since they are presented without regard to their random nature and the need for proper statistical analysis of simulation output data.This paper discusses the main factors that can affect the accuracy of stochastic simulations designed to give insight into the steady-state behavior of queuing processes. The problems of correctly starting and stopping such simulation experiments to obtain the required statistical accuracy of the results are addressed. In this survey of possible solutions, the emphasis is put on possible applications in the sequential analysis of output data, which adaptively decides about continuing a simulation experiment until the required accuracy of results is reached. A suitable solution for deciding upon the starting point of a steady-state analysis and two techniques for obtaining the final simulation results to a required level of accuracy are presented, together with pseudocode implementations.

Sensitivity of computer simulation experiments to errors in input data

Abstract: This paper compares two methods of assessing variability in simulation output. The methods make specific allowance for two sources of variation: that caused by uncertainty in estimating unknown input parameters (parameter uncertainty), and that caused by the inclusion of random variation within the simulation model itself (simulation uncertainty). The first method is based on classical statistical differential analysis; we show explicitly that, under general conditions, the two sources contribute separately to the total variation. In the classical approach, certain sensitivity coefficients have to be estimated. The effort needed to do this becomes progressively more expensive, increasing linearly with the number of unknown parameters. Moreover there is an additional difficulty of detecting spurious variation when the number of parameters is large. It is shown that a parametric form of bootstrap sampling provides an alternative method which does not suffer from either problem. For illustration, simulation ...