|
Activity Number:
|
433
|
|
Type:
|
Roundtables
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 12:30 PM to 1:50 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #308240 |
|
Title:
|
Verifying Computer Experiments with Physical Experiments
|
|
Author(s):
|
William Notz*+
|
|
Companies:
|
The Ohio State University
|
|
Address:
|
Dept. of Statistics, Columbus, OH, 43210-1247,
|
|
Keywords:
|
computer experiments ; validation
|
|
Abstract:
|
Complex computer simulations have become popular substitutes for physical processes that are expensive or difficult to observe. Experimenting with the computer simulation as a means of studying the physical process is called a computer experiment. For this approach to yield meaningful results, the computer simulation must be a reasonable approximation to the physical experiment. Verifying this is sometimes referred to as validating the computer experiment. This session will consider methods for validating computer experiments. There is only a modest literature on validating computer experiments. We will discuss the current state of the literature on this topic, problems that are related (such as calibration and combining data from multiple sources), and possible research directions. A bibliography of some of the literature that exists will be provided.
|
