|
Activity Number:
|
542
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Government Statistics
|
| Abstract - #309242 |
|
Title:
|
Statistical 'Pre- and Post-Processing' as Integrated into Computer Simulations
|
|
Author(s):
|
Turkan Gardenier*+
|
|
Companies:
|
Pragmatica Corp.
|
|
Address:
|
115 Saint Andrews Dr. NE, Vienna, VA, 22180,
|
|
Keywords:
|
Pre-processing ; post-processing ; simulation ; Hadamard matrix ; fractional factorial ; response surface
|
|
Abstract:
|
Large scale computer simulations demand the generation of data in order to evaluate the effects of input parameters (singly and jointly) upon results. Design and selection of input data using Hadamard matrices and fractional factorials, also denoted as "pre- processing," enhance the reliability of outputs analyzed. Results may be expressed as a "meta-model" or submitted to stagewise refinement using statistical response surface methodology. In computerized simulation studies the process has sometimes been denoted as "post-processing." This presentation will illustrate the integration of a compendium of these statistical methods into large scale military simulations and their impact on cost reduction and data efficiency.
|
