Analysis of initial transient deletion for replicated steady-state simulations

@article{Glynn1991AnalysisOI,
  title={Analysis of initial transient deletion for replicated steady-state simulations},
  author={Peter W. Glynn and Philip Heidelberger},
  journal={Oper. Res. Lett.},
  year={1991},
  volume={10},
  pages={437-443}
}

Replicated batch means for steady-state simulations with initial transients

We provide asymptotic expressions for the expected value and variance of the replicated batch means variance estimator when the stochastic process being simulated has an additive initial transient.

Steady-State Simulation with Replication-Dependent Initial Transients: Analysis and Examples

TLDR
The replicated batch means RBM method for steady-state simulation output analysis generalizes both the independent replications IR and batch means BM methods and provides guidance on when IR, BM, or a combination thereof is the best choice, and also on effective choices of initial states for the replications.

A New Approach for Parallel Steady-State Simulations

  • Ming-hua Hsieh
  • Mathematics
    Proceedings of the 2006 Winter Simulation Conference
  • 2006
TLDR
A new procedure for building confidence interval estimators of steady-state parameters in discrete event simulations is proposed that can be improved by simply increasing the number of processors while the simulated time length is fixed on an appropriate level on each processor.

Experiments with initial transient deletion for parallel, replicated steady-state simulations

TLDR
The theory and the experimental results reported here strongly indicate that a nonstandard in the context of steady-state simulation, yet easy to apply, estimation procedure is required on highly parallel machines and this nonstandard estimator is a ratio estimator.

The efficiency of one long run versus independent replication in steady-state simulation

We evaluate the efficiency of one long run versus independent replications in steady-state discrete-event simulation, assuming that an initial portion of each replication will be deleted to allow the

New estimators for parallel steady-state simulations

TLDR
A class of estimators obtained by putting different weights on the sampling average across replications at selected time points is considered, chosen to maximize their Gaussian likelihood.

Analysis of parallel replicated simulations under a completion time constraint

TLDR
This work considers simulation experiments in which t is viewed as either being a strict constraint, or a guideline, in which case simulation beyond time t is permitted, and proposes an unbiased estimator for a simple mean value.

ARD: An automated replication-deletion method for simulation analysis

TLDR
Although ARD's CIs closely conformed to given coverage and precision requirements, ARD generally required a larger computing budget than single-run procedures.

To batch or not to batch?

TLDR
Batching can often do better than replication when it comes to point and confidence-interval estimation of the steady-state mean in the presence of an initial transient, and is a common rule of thumb in the folklore.

On an Initial Transient Deletion Rule with Rigorous Theoretical Support

TLDR
It is argued that the initial transient deletion rule has desirable properties both from a theoretical and practical standpoint and its use both in the single replication setting and in the multiple replications/parallel processing context is discussed.

References

SHOWING 1-10 OF 22 REFERENCES

Analysis of Initial Transient Deletion for Parallel Steady-State Simulations

This paper investigates theoretical properties of a simple method for using parallel processors in discrete event simulations: running independent replications, in parallel, on multiple processors

Limit theorems for the method of replication

The method of replication is frequently used by simulators to estimate steady-state quantities. In this paper, we obtain conditions under which this method yields asymptotically valid confidence

Experiments with initial transient deletion for parallel, replicated steady-state simulations

TLDR
The theory and the experimental results reported here strongly indicate that a nonstandard in the context of steady-state simulation, yet easy to apply, estimation procedure is required on highly parallel machines and this nonstandard estimator is a ratio estimator.

Analysis of parallel replicated simulations under a completion time constraint

TLDR
This work considers simulation experiments in which t is viewed as either being a strict constraint, or a guideline, in which case simulation beyond time t is permitted, and proposes an unbiased estimator for a simple mean value.

Optimal Tests for Initialization Bias in Simulation Output

TLDR
A family of tests for detecting initialization bias in the mean of a simulation output series using a hypothesis testing framework and appears to be both robust and powerful.

Simulation Run Length Control in the Presence of an Initial Transient

TLDR
A procedure based on Schruben's Brownian bridge model for the detection of nonstationarity and a spectral method for estimating the variance of the sample mean are explored for estimation of the steady state mean of an output sequence from a discrete event simulation.

A survey of research on the simulation startup problem

TLDR
This paper surveys research on the simulation startup problem and presents a procedure for evaluating startup policies when the mean of the truncated sample is to be used to estimate the steady-statemean of the underlying process.

A new approach for dealing with the startup problem in discrete event simulation

TLDR
A practical, data-based statistical procedure which can be used to reduce or remove bias owing to artificial startup conditions in simulations aimed at estimating steady-state means is proposed.

Evaluation of commonly used rules for detecting “steady state” in computer simulation

TLDR
Five heuristic rules for predicting the approximate end of transiency, four of which have been quoted extensively in the simulation literature, have been evaluated in the M/M/1 situation and performed poorly.

Control of initialization bias in multivariate simulation response

A procedure is proposed for controlling initialization bias in multiple response simulation output. The technique is an adaptation of multivariate quality control methods using a two-sample T 2