International Series in Operations Research & Management Science

Transcription

1 International Series in Operations Research & Management Science Volume 133 For other titles published in this series, go to

2 George Samuel Fishman

3 Christos Alexopoulos David Goldsman James R. Wilson Editors Advancing the Frontiers of Simulation A Festschrift in Honor of George Samuel Fishman With Contributions by Russell Cheng, Soumyadip Ghosh, Shane G. Henderson, Peter W. Glynn, Eunji Lim, James O. Henriksen, Wolfgang Hörmann, Josef Leydold, Jack P. C. Kleijnen, Pierre L Ecuyer, François Panneton, Andrew F. Seila, Sally Brailsford, Steffen Straßburger, Thomas Schulze, Richard Fujimoto, Shing Chih Tsai, Barry L. Nelson, Jeremy Staum, Tûba Aktaran-Kalaycı 123

4 Editors Christos Alexopoulos Georgia Institute of Technology H. Milton Stewart School of Industrial and Systems Engineering 765 Ferst Drive NW Atlanta GA USA James R. Wilson Edward P. Fitts Department of Industrial and Systems Engineering North Carolina State University 111 Lampe Drive Raleigh NC USA David Goldsman Georgia Institute of Technology H. Milton Stewart School of Industrial and Systems Engineering 765 Ferst Drive NW Atlanta GA USA ISSN ISBN e-isbn DOI /b Springer Dordrecht Heidelberg London New York Library of Congress Control Number: c Springer Science+Business Media, LLC 2009 All rights reserved. This work may not be translated or copied in whole or in part without the written permission 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 in connection with any form of information storage and retrieval, electronic adaptation, computer software, 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 if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper Springer is part of Springer Science+Business Media (

5 Preface This Festschrift honors George Samuel Fishman, one of the founders of the field of computer simulation and a leader of the disciplines of operations research and the management sciences for the past five decades, on the occasion of his seventieth birthday. The papers in this volume span the theory, methodology, and application of 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 management sciences, with special emphasis on his role in the advancement of the field of simulation since the 1960s. We also include a brief personal biography together with comments by several individuals about the extraordinary effect that George has had on all his students, colleagues, and friends. The second article, titled A Conversation with George Fishman, is the transcript of an extended interview with George that we conducted in October In the article titled Computer Intensive Statistical Model Building, Russell Cheng studies resampling methods for building parsimonious multiple linear regression models so as to represent accurately the behavior of the dependent variable in terms of the smallest possible subset of explanatory (independent) variables. The author shows how bootstrap resampling can be used not only for rapid identification of good models but also for efficient comparison of competing models. The next article is titled Patchwork Distributions. In this article, Soumyadip Ghosh and Shane Henderson consider a class of multivariate probability distributions that can be used to model a finite-dimensional random vector when the user has specified all the marginal distributions of the random vector, the covariance matrix of the random vector, and the probabilities that the random vector lies in certain regions. Peter Glynn and Eunji Lim examine the foundations of the batch-means method for steady-state simulation analysis in their article titled Asymptotic Validity of Batch Means Steady-State Confidence Intervals. Although the large-sample validity of many implementations of the batch-means method requires that the target output process must satisfy a functional central limit theorem, in this article the authors establish the validity of the batch-means method for Harris-recurrent Markov processes that satisfy a weaker (nonfunctional) central limit theorem. v

6 vi Preface In the article titled Efficient Modeling of Delays in Discrete-Event Simulation, Jim Henriksen considers the problem of efficiently implementing four types of delays that are commonly encountered in simulation modeling, with a detailed discussion of the solution approach provided by the delay-modeling algorithms that have been developed for the SLX simulation language. In the article titled Sampling from Linear Multivariate Densities, Wolfgang Hörmann and Josef Leydold develop an efficient acceptance-rejection algorithm for generating a vector of dependent random variables whose joint density is linear with a domain that is bounded and symmetric about a point; and ultimately the authors extend their algorithm to generate random vectors from concave differentiable densities over point-symmetric domains. Jack Kleijnen s article, Factor Screening in Simulation Experiments: Review of Sequential Bifurcation, provides an overview of factor-screening methods for simulation experiments with special emphasis on sequential bifurcation, a method that is particularly effective in certain types of large-scale simulation studies. F 2 -Linear Random Number Generators, by Pierre L Ecuyer and François Panneton, is the ninth article in this volume. This article reviews various construction methods for random-number generators based on linear recurrence modulo 2, examines their theoretical properties, describes the relevant computational tools and algorithms, and ends with comparisons based on various qualitative criteria. The F 2 - linear class contains many long-period random-number generators including the Tausworthe, generalized feedback shift register (GSFR), twisted GSFR, Mersenne twister, WELL, and xorshift generators. The tenth article is titled Opportunities and Challenges in Health Care Simulation. In this article Andrew F. Seila and Sally Brailsford review successful applications of simulation in health care, examine the suitability of simulation models and methods for the analysis of health-care systems, and explore the reasons for the lack of adoption of simulation as a routine tool for health-care systems analysis. The authors end with an insightful list of ideas aimed at promoting wider adoption of simulation in health care. In the eleventh article, Future Trends in Distributed Simulation and Distributed Virtual Environments, Steffen Straßburger, Thomas Schulze, and Richard Fujimoto report the main results of a peer study of current trends in distributed simulation and distributed virtual environments. The survey assesses the current state of this methodology, its relevance to simulation practice, and the research challenges that must be addressed so as to facilitate the widespread use of this methodology in industry and government as well as in research organizations. In Combined Screening and Selection of the Best with Control Variates, Shing Chih Tsai, Barry L. Nelson, and Jeremy Staum formulate ranking-and-selection procedures with screening that exploit point estimators based on the method of control variates to gain greater statistical efficiency. Compared with previous ranking-andselection 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 Estimators for Steady-State Simulation, Tûba Aktaran-Kalaycı, Christos Alexopoulos, David Goldsman, and James R. Wilson seek to estimate the variance parameter of a simulation output process using optimal linear combinations of variance estimators

7 Preface vii based on the methods of overlapping batch means and standardized time series. From the latter estimators they derive asymptotically valid confidence intervals for both the mean and variance parameter of the target process. We thank the authors who have contributed to this volume, especially for their forbearance throughout the lengthy process of completing this book. Special thanks go to Yvonne Smith (Georgia Tech), who produced the transcription of our interview with George, and to Carolyn Ford at Springer, who so ably supervised the production of this book. We are also indebted to Fred Hillier (Stanford University) and Gary Folven (Springer) for patiently guiding us through the entire process. Atlanta, GA, USA Atlanta, GA, USA Raleigh, NC, USA March 2009 Christos Alexopoulos David Goldsman James R. Wilson

8 Contents George Fishman s Professional Career... 1 Christos Alexopoulos, David Goldsman, and James R. Wilson A Conversation with George Fishman Christos Alexopoulos, David Goldsman, and James R. Wilson Computer Intensive Statistical Model Building Russell Cheng Patchwork Distributions Soumyadip Ghosh and Shane G. Henderson Asymptotic Validity of Batch Means Steady-State Confidence Intervals Peter W. Glynn and Eunji Lim Efficient Modeling of Delays in Discrete-Event Simulation James O. Henriksen Sampling from Linear Multivariate Densities Wolfgang Hörmann and Josef Leydold Factor Screening in Simulation Experiments: Review of Sequential Bifurcation Jack P. C. Kleijnen F 2 -Linear Random Number Generators Pierre L Ecuyer and François Panneton Opportunities and Challenges in Health Care Simulation Andrew F. Seila and Sally Brailsford ix

9 x Contents Future Trends in Distributed Simulation and Distributed Virtual Environments Steffen Straßburger, Thomas Schulze, and Richard Fujimoto Combined Screening and Selection of the Best with Control Variates Shing Chih Tsai, Barry L. Nelson, and Jeremy Staum Optimal Linear Combinations of Overlapping Variance Estimators for Steady-State Simulation Tûba Aktaran-Kalaycı, Christos Alexopoulos, David Goldsman, and James R. Wilson Author Index...329

10 Contributors Tûba Aktaran-Kalaycı H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA, Christos Alexopoulos H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA, Sally Brailsford School of Management, University of Southampton, Highfield Campus, Southampton S017 1BJ, UK, Russell Cheng School of Mathematics, University of Southampton, Southampton, SO17 1BJ, UK, Richard Fujimoto Computational Science and Engineering Division, College of Computing, Georgia Institute of Technology, Atlanta, GA , USA, Soumyadip Ghosh IBM T. J. Watson Research Center, Yorktown Heights, NY 10598, USA, Peter W. Glynn Department of Management Science and Engineering, Stanford University, Stanford, CA 94305, USA, David Goldsman H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA, Shane G. Henderson School of Operations Research and Information Engineering, Cornell University, Ithaca, NY 14853, USA, James O. Henriksen Wolverine Software Corporation, Alexandria, VA 22305, USA, Wolfgang Hörmann Department of Industrial Engineering, Boğaziçi University, Bebek-Istanbul, Turkey, Jack P. C. Kleijnen Department of Information Management, Tilburg University, 5000 LE Tilburg, Netherlands, xi

11 xii Contributors Josef Leydold Department of Statistics and Mathematics, WU Wien, Augasse 2 6, A-1090 Vienna, Austria, josef.leydold@wu-wien.ac.at Pierre L Ecuyer Département d Informatique et de Recherche Opérationnelle (DIRO),Université de Montréal, C.P. 6128, Succ. Centre-Ville, Montréal, H3C 3J7, Canada, lecuyer@iro.umontreal.ca Eunji Lim Department of Industrial Engineering, University of Miami, Miami, FL 33124, USA, lim@miami.edu Barry L. Nelson Department of Industrial Engineering & Management Sciences, Northwestern University, Evanston, IL 60208, USA, nelsonb@iems.northwestern.edu François Panneton Standard Life Investments Inc., 1001 de Maisonneuve Ouest, bureau 700, Montréal, Québec, H3A 3C8, Canada, Francois.Panneton@standardlife.ca Thomas Schulze School of Computer Science, Otto-von-Guericke University of Magdeburg, Magdeburg, Germany, tom@iti.cs.uni-magdeburg.de Andrew F. Seila University of Georgia, 1240 Saint Andrews Drive, Bogart, Georgia 30602, USA, aseila@gmail.com Jeremy Staum Department of Industrial Engineering & Management Sciences, Northwestern University, Evanston, IL 60208, USA, staum@iems.northwestern.edu Steffen Straßburger School of Economic Sciences, Ilmenau University of Technology, Ilmenau, Germany, Steffen.Strassburger@tu-ilmenau.de Shing Chih Tsai Department of Industrial and Information Management, National Cheng Kung University, Tainan, Taiwan, str6969@mail.ncku.edu.tw James R. Wilson Edward P. Fitts Department of Industrial and Systems Engineering, North Carolina State University, Raleigh, NC , USA, jwilson@ncsu.edu

12 George Fishman Receives the INFORMS Simulation Society s Lifetime Professional Achievement Award from R. G. Sargent (2003)