Variance Reduction using Nonlinear Control and Transformations
Lewis, Peter A.W.
Ressler, Richard L.
Wood, R. Kevin
MetadataShow full item record
Nonlinear regression-adjusted control variables are investigated for improving variance reduction in statistical and systems simulations. To this end, simple control variables are piecewise sectioned and then transformed using linear and nonlinear transformations. Optical parameters of these transformations are selected using linear or nonlinear least-squares regression algorithms. As an example, piecewise power-transformed variables are used in the estimation of the mean for the two-variable Anderson-Darling goodness-of-fit statistic W2. Substantial variance reduction over straightforward controls is obtained. These parametric transformations are compared against optimal, additive nonparametric transformations obtained by using the ACE algorithm and are sown, in comparison to the results from ACE, to be nearly optimal.
RightsThis publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.
NPS Report NumberNPS55-88-007
Showing items related by title, author, creator and subject.
Lewis, Peter A.W.; Ressler, Richard; Wood, R. Kevin (1987);Nonlinear regression-adjusted control variables are investigated for improving variance reduction in statistical and system simulations. Simple control variables are transformed using linear and nonlinear transformations, ...
An investigation of nonlinear controls and regression-adjusted estimators for variance reduction in computer simulation. Ressler, Richard L. (Monterey, California. Naval Postgraduate School, 1991-03);This dissertation develops new techniques for variance reduction in computer simulation. It demonstrates that applying nonlinear transformations to control variables can increase their effectiveness over linear controls. ...
Buttrey, Samuel E.; Whitaker, Lyn R. (2016-04-19); NPS-OR-16-003Clustering techniques divide observations into groups.Current techniques usually rely on measurements of dissimilarities between pairs of observations, between pairs of clusters, and between an observation and a cluster.For ...