Efficient Modeling & Simulation of Biological Warfare Using Innovative Design of Experiments Methods

Overview

Innovative Design of Experiments (DOE) methods are used to significantly reduce the number of simulations required to model Biological Warfare (BW) attacks. The methods illustrated are applicable to almost any Modeling and Simulation (M&S) study involving large numbers of variables with many levels. The goal is to create a fast (seconds) surrogate metamodel of the simulation with which to make suitably accurate predictions of the response output for time consuming (hours) simulation trials not yet run. An example is shown that employs DOE methods that can be used when control variables are categorical and/or continuous. The analysis is based on a re-evaluation of data for all 648 possible combinations of settings of six variables in a BW attack case matrix completed for the military.