Session Detail Information
Add this session to your itinerary

Cluster :  Simulation

Session Information  : Sunday Nov 09, 11:00 - 12:30

Title:  Simulation Modeling for Anaysis
Chair: Lee Schruben,University of California, Berkeley, Berkeley CA, United States of America, lees@berkeley.edu

Abstract Details

Title: Capturing the Impact of Input Model Uncertainty in Stochastic Models and Analysis: Lessons Learned
 Presenting Author: Russell Barton,Pennsylvania State University, 210E Business Building, Pennsylvania, United States of America, rrb2@psu.edu
 Co-Author: Barry Nelson,Northwestern University, 2145 Sheridan R.d. C210, Evanston IL 60091, United States of America, nelsonb@northwestern.edu
 Wei Xie,Rensselaer Polytechnic Institute, 110 8th Street, Center for Industrial Innovation 5207, Troy NY 12180-3590, United States of America, xiew3@rpi.edu
 
Abstract: Discrete event simulations are driven by input models fitted with finite data. This presentation will highlight the difficulties in characterizing the impact on output analysis for the general case, the progress that has been made in special cases, and the broader implications for many stochastic modeling and optimization settings.
  
Title: Optimal Selection of the Most Probable Multinomial Alternative
 Presenting Author: David Goldsman,Professor, Georgia Tech, Stewart School of Indust & Sys Engr, Atlanta GA 30332-0205, United States of America, Sman@isye.gatech.edu
 Co-Author: Anton Kleywegt,Georgia Institute of Technology, School of Industrial and Systems Enginee, Atlanta GA 30332-0205, United States of America, anton@isye.gatech.edu
 Eric Tollefson,United States Military Academy, Department of Systems Engineering, United States Military Academy, West Point NY, United States of America, eric.tollefson@usma.edu
 Craig Tovey,Georgia Tech, School of ISyE, Atlanta GA 3033, United States of America, ctovey@isye.gatech.edu
 
Abstract:  We present selection procedures based on linear and mixed-integer linear programs that find the multinomial cell having the highest probability. Our procedures are optimal in the sense that they minimize the number of observations taken to achieve a certain probability of correct selection.
  
Title: Vamp1re: A Single Criterion for Evaluating Confidence-Interval Procedures
 Presenting Author: Bruce Schmeiser,Emeritus Professor, Purdue University, School of Industrial Engineering, West Lafayette, United States of America, bruceschmeiser@gmail.com
 Co-Author: Ying-chieh Yeh,National Central University, School of Management, Chungli, Taiwan - ROC, yeh@mgt.ncu.edu.tw
 
Abstract: Confidence-interval procedures are classically evaluated using actual coverage probability and expected half width. We argue for a single criterion, based on Schruben's coverage function when sample size is constant. We extend the criterion to procedures with stopping rules.
  
Title: Effective Simulation Warm-up for a Neonatal Intensive Care Unit
 Presenting Author: Emily Lada,SAS Institute, World Headquarters, SAS Campus Drive, Cary NC 27513, United States of America, Emily.Lada@sas.com
 Co-Author: Anup Mokashi,SAS Institute, World Headquarters, SAS Campus Drive, Cary NC 27513, United States of America, Anup.Mokashi@sas.com
 James Wilson,North Carolina State University, Dept.of Industrial & Systems Engineering, 111 Lampe Drive, Daniels Hall, Raleigh NC 27695, United States of America, jwilson@ncsu.edu
 
Abstract: In simulating a neonatal intensive care unit subject to constraints on the available computing budget for generating certain key responses, effective warm-up is required to compute accurate point and confidence-interval estimates of, for example, the expected number of admissions per year as well as the long-run average length of a patient’s stay. Techniques for steady-state simulation analysis (N-Skart and SBatch) are adapted to this problem.