To match the greatly increasing power of scientific engineering workstations with the slowly changing capability of the simulation practitioner, the authors examine the possibilities for a high-level output analysis interface. The goal is an interface which considers a parametric family of models and accepts commands stated in terms of understanding model characteristics over this family and for making practical decisions with respect to it. The interfaces considered must exploit the unique features of the simulation environment: the fact that additional data are readily available and that these data can be immediately incorporated into the analysis and presented to the user. Thus, many of the classical considerations of statistics involving making optimum use of limited data become eclipsed by questions of what additional data are needed to make the necessary decisions and how to present the results as the data are being acquired so as to maximize the power of the combination of practitioner and machine.<<ETX>>
[1]
David Edelman,et al.
The Five-Degree-of-Freedom Rule of Thumb for Fixed-Width Confidence Intervals for a Normal Mean
,
1991
.
[2]
I. Ibragimov,et al.
On Sequential Estimation
,
1975
.
[3]
N. Starr.
On the Asymptotic Efficiency of a Sequential Procedure for Estimating the Mean
,
1966
.
[4]
Anders Hald,et al.
Statistical Theory with Engineering Applications
,
1952
.
[5]
W. Cleveland,et al.
Locally Weighted Regression: An Approach to Regression Analysis by Local Fitting
,
1988
.
[6]
N. Starr.
The Performance of a Sequential Procedure for the Fixed-Width Interval Estimation of the Mean
,
1966
.
[7]
A. Nádas.
An Extension of a Theorem of Chow and Robbins on Sequential Confidence Intervals for the Mean
,
1969
.
[8]
Averill M. Law,et al.
Simulation Modeling and Analysis
,
1982
.
[9]
Michael Woodroofe,et al.
Asymptotic optimality in sequential interval estimation
,
1986
.