international series in operations research & management ... · management sciences for the...
TRANSCRIPT
International Series in OperationsResearch & Management Science
Volume 133
For other titles published in this series, go towww.springer.com/series/6161
George Samuel Fishman
Christos Alexopoulos · David Goldsman ·James R. WilsonEditors
Advancing the Frontiersof Simulation
A Festschrift in Honor ofGeorge Samuel Fishman
With Contributions byRussell Cheng, Soumyadip Ghosh, Shane G. Henderson,Peter W. Glynn, Eunji Lim, James O. Henriksen,Wolfgang Hormann, Josef Leydold, Jack P. C. Kleijnen,Pierre L’Ecuyer, Francois Panneton, Andrew F. Seila,Sally Brailsford, Steffen Straßburger, Thomas Schulze,Richard Fujimoto, Shing Chih Tsai, Barry L. Nelson,Jeremy Staum, Tuba Aktaran-Kalaycı
123
EditorsChristos AlexopoulosGeorgia Institute of TechnologyH. Milton Stewart School of Industrial
and Systems Engineering765 Ferst Drive NWAtlanta GA [email protected]
David GoldsmanGeorgia Institute of TechnologyH. Milton Stewart School of Industrial
and Systems Engineering765 Ferst Drive NWAtlanta GA [email protected]
James R. WilsonEdward P. Fitts Department of Industrial
and Systems EngineeringNorth Carolina State University111 Lampe DriveRaleigh NC [email protected]
ISSN 0884-8289ISBN 978-1-4419-0816-2 e-ISBN 978-1-4419-0817-9DOI 10.1007/b110059Springer Dordrecht Heidelberg London New York
Library of Congress Control Number: 2009932145
c© Springer Science+Business Media, LLC 2009All rights reserved. This work may not be translated or copied in whole or in part without the writtenpermission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York,NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use inconnection with any form of information storage and retrieval, electronic adaptation, computersoftware, or by similar or dissimilar methodology now known or hereafter developed is forbidden.The use in this publication of trade names, trademarks, service marks, and similar terms, even ifthey are not identified as such, is not to be taken as an expression of opinion as to whether or notthey are subject to proprietary rights.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
Preface
This Festschrift honors George Samuel Fishman, one of the founders of the field ofcomputer simulation and a leader of the disciplines of operations research and themanagement sciences for the past five decades, on the occasion of his seventiethbirthday. The papers in this volume span the theory, methodology, and applicationof computer simulation.
The lead article is appropriately titled “George Fishman’s Professional Career.”In this article we discuss George’s contributions to operations research and the man-agement sciences, with special emphasis on his role in the advancement of the fieldof simulation since the 1960s. We also include a brief personal biography togetherwith comments by several individuals about the extraordinary effect that George hashad on all his students, colleagues, and friends.
The second article, titled “A Conversation with George Fishman,” is the transcriptof an extended interview with George that we conducted in October 2007.
In the article titled “Computer Intensive Statistical Model Building,” RussellCheng studies resampling methods for building parsimonious multiple linear regres-sion models so as to represent accurately the behavior of the dependent variable interms of the smallest possible subset of explanatory (independent) variables. Theauthor shows how bootstrap resampling can be used not only for rapid identificationof good models but also for efficient comparison of competing models.
The next article is titled “Patchwork Distributions.” In this article, SoumyadipGhosh and Shane Henderson consider a class of multivariate probability distribu-tions that can be used to model a finite-dimensional random vector when the userhas specified all the marginal distributions of the random vector, the covariancematrix of the random vector, and the probabilities that the random vector lies incertain regions.
Peter Glynn and Eunji Lim examine the foundations of the batch-means methodfor steady-state simulation analysis in their article titled “Asymptotic Validity ofBatch Means Steady-State Confidence Intervals.” Although the large-sample valid-ity of many implementations of the batch-means method requires that the target out-put process must satisfy a functional central limit theorem, in this article the authorsestablish the validity of the batch-means method for Harris-recurrent Markov pro-cesses that satisfy a weaker (nonfunctional) central limit theorem.
v
vi Preface
In the article titled “Efficient Modeling of Delays in Discrete-Event Simula-tion,” Jim Henriksen considers the problem of efficiently implementing four typesof delays that are commonly encountered in simulation modeling, with a detaileddiscussion of the solution approach provided by the delay-modeling algorithms thathave been developed for the SLX simulation language.
In the article titled “Sampling from Linear Multivariate Densities,” WolfgangHormann and Josef Leydold develop an efficient acceptance-rejection algorithm forgenerating a vector of dependent random variables whose joint density is linear witha domain that is bounded and symmetric about a point; and ultimately the authorsextend their algorithm to generate random vectors from concave differentiable den-sities over point-symmetric domains.
Jack Kleijnen’s article, “Factor Screening in Simulation Experiments: Reviewof Sequential Bifurcation,” provides an overview of factor-screening methods forsimulation experiments with special emphasis on sequential bifurcation, a methodthat is particularly effective in certain types of large-scale simulation studies.
“F2-Linear Random Number Generators,” by Pierre L’Ecuyer and FrancoisPanneton, is the ninth article in this volume. This article reviews various construc-tion methods for random-number generators based on linear recurrence modulo 2,examines their theoretical properties, describes the relevant computational tools andalgorithms, and ends with comparisons based on various qualitative criteria. The F2-linear class contains many long-period random-number generators—including theTausworthe, generalized feedback shift register (GSFR), twisted GSFR, Mersennetwister, WELL, and xorshift generators.
The tenth article is titled “Opportunities and Challenges in Health Care Simula-tion.” In this article Andrew F. Seila and Sally Brailsford review successful applica-tions of simulation in health care, examine the suitability of simulation models andmethods for the analysis of health-care systems, and explore the reasons for the lackof adoption of simulation as a “routine” tool for health-care systems analysis. Theauthors end with an insightful list of ideas aimed at promoting wider adoption ofsimulation in health care.
In the eleventh article, “Future Trends in Distributed Simulation and DistributedVirtual Environments,” Steffen Straßburger, Thomas Schulze, and Richard Fujimotoreport the main results of a peer study of current trends in distributed simulationand distributed virtual environments. The survey assesses the current state of thismethodology, its relevance to simulation practice, and the research challenges thatmust be addressed so as to facilitate the widespread use of this methodology inindustry and government as well as in research organizations.
In “Combined Screening and Selection of the Best with Control Variates,” ShingChih Tsai, Barry L. Nelson, and Jeremy Staum formulate ranking-and-selection pro-cedures with screening that exploit point estimators based on the method of controlvariates to gain greater statistical efficiency. Compared with previous ranking-and-selection procedures that incorporate screening and selection of the best alternative,substantial improvements in performance are achieved by the new procedures.
In the final article, “Optimal Linear Combinations of Overlapping Variance Esti-mators for Steady-State Simulation,” Tuba Aktaran-Kalaycı, Christos Alexopoulos,David Goldsman, and James R. Wilson seek to estimate the variance parameter of asimulation output process using optimal linear combinations of variance estimators
Preface vii
based on the methods of overlapping batch means and standardized time series.From the latter estimators they derive asymptotically valid confidence intervals forboth the mean and variance parameter of the target process.
We thank the authors who have contributed to this volume, especially for theirforbearance throughout the lengthy process of completing this book. Special thanksgo to Yvonne Smith (Georgia Tech), who produced the transcription of our interviewwith George, and to Carolyn Ford at Springer, who so ably supervised the productionof this book. We are also indebted to Fred Hillier (Stanford University) and GaryFolven (Springer) for patiently guiding us through the entire process.
Atlanta, GA, USA Christos AlexopoulosAtlanta, GA, USA David GoldsmanRaleigh, NC, USA James R. WilsonMarch 2009
Contents
George Fishman’s Professional Career . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Christos Alexopoulos, David Goldsman, and James R. Wilson
A Conversation with George Fishman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Christos Alexopoulos, David Goldsman, and James R. Wilson
Computer Intensive Statistical Model Building . . . . . . . . . . . . . . . . . . . . . . . . 43Russell Cheng
Patchwork Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Soumyadip Ghosh and Shane G. Henderson
Asymptotic Validity of Batch Means Steady-State Confidence Intervals . . . 87Peter W. Glynn and Eunji Lim
Efficient Modeling of Delays in Discrete-Event Simulation . . . . . . . . . . . . . . 105James O. Henriksen
Sampling from Linear Multivariate Densities . . . . . . . . . . . . . . . . . . . . . . . . . . 143Wolfgang Hormann and Josef Leydold
Factor Screening in Simulation Experiments: Review of SequentialBifurcation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153Jack P. C. Kleijnen
F2-Linear Random Number Generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169Pierre L’Ecuyer and Francois Panneton
Opportunities and Challenges in Health Care Simulation . . . . . . . . . . . . . . . 195Andrew F. Seila and Sally Brailsford
ix
x Contents
Future Trends in Distributed Simulationand Distributed Virtual Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231Steffen Straßburger, Thomas Schulze, and Richard Fujimoto
Combined Screening and Selection of the Best with Control Variates . . . . . 263Shing Chih Tsai, Barry L. Nelson, and Jeremy Staum
Optimal Linear Combinations of Overlapping Variance Estimators forSteady-State Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291Tuba Aktaran-Kalaycı, Christos Alexopoulos,David Goldsman, and James R. Wilson
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Contributors
Tuba Aktaran-Kalaycı H. Milton Stewart School of Industrial and SystemsEngineering, Georgia Institute of Technology, Atlanta, GA 30332, USA,[email protected]
Christos Alexopoulos H. Milton Stewart School of Industrial and SystemsEngineering, Georgia Institute of Technology, Atlanta, GA 30332, USA,[email protected]
Sally Brailsford School of Management, University of Southampton, HighfieldCampus, Southampton S017 1BJ, UK, [email protected]
Russell Cheng School of Mathematics, University of Southampton, Southampton,SO17 1BJ, UK, [email protected]
Richard Fujimoto Computational Science and Engineering Division, College ofComputing, Georgia Institute of Technology, Atlanta, GA 30332-0765, USA,[email protected]
Soumyadip Ghosh IBM T. J. Watson Research Center, Yorktown Heights, NY10598, USA, [email protected]
Peter W. Glynn Department of Management Science and Engineering, StanfordUniversity, Stanford, CA 94305, USA, [email protected]
David Goldsman H. Milton Stewart School of Industrial and Systems Engineering,Georgia Institute of Technology, Atlanta, GA 30332, USA, [email protected]
Shane G. Henderson School of Operations Research and Information Engineering,Cornell University, Ithaca, NY 14853, USA, [email protected]
James O. Henriksen Wolverine Software Corporation, Alexandria, VA 22305,USA, [email protected]
Wolfgang Hormann Department of Industrial Engineering, Bogazici University,34342 Bebek-Istanbul, Turkey, [email protected]
Jack P. C. Kleijnen Department of Information Management, Tilburg University,5000 LE Tilburg, Netherlands, [email protected]
xi
xii Contributors
Josef Leydold Department of Statistics and Mathematics, WU Wien, Augasse 2–6,A-1090 Vienna, Austria, [email protected]
Pierre L’Ecuyer Departement d’Informatique et de Recherche Operationnelle(DIRO), Universite de Montreal, C.P. 6128, Succ. Centre-Ville, Montreal, H3C 3J7,Canada, [email protected]
Eunji Lim Department of Industrial Engineering, University of Miami, Miami, FL33124, USA, [email protected]
Barry L. Nelson Department of Industrial Engineering & Management Sciences,Northwestern University, Evanston, IL 60208, USA,[email protected]
Francois Panneton Standard Life Investments Inc., 1001 de Maisonneuve Ouest,bureau 700, Montreal, Quebec, H3A 3C8, Canada,[email protected]
Thomas Schulze School of Computer Science, Otto-von-Guericke University ofMagdeburg, 39016 Magdeburg, Germany, [email protected]
Andrew F. Seila University of Georgia, 1240 Saint Andrews Drive, Bogart, Georgia30602, USA, [email protected]
Jeremy Staum Department of Industrial Engineering & Management Sciences,Northwestern University, Evanston, IL 60208, USA, [email protected]
Steffen Straßburger School of Economic Sciences, Ilmenau University ofTechnology, 98684 Ilmenau, Germany, [email protected]
Shing Chih Tsai Department of Industrial and Information Management, NationalCheng Kung University, Tainan, Taiwan, [email protected]
James R. Wilson Edward P. Fitts Department of Industrial and SystemsEngineering, North Carolina State University, Raleigh, NC 27695-7906, USA,[email protected]
George Fishman Receives the INFORMS Simulation Society’s LifetimeProfessional Achievement Award from R. G. Sargent (2003)