Application of Rule-Based Computer Models to the Evaluation of Combat Training: A Feasibility Study

Abstract : This report describes the results of a study exploring the application of two modern computer techniques, knowledge-based modeling and adaptive programming technology to the analysis of small-unit tactical engagements. The underlying purpose of the study is to improve exercise evaluation in complex and realistic tactical training systems, such as the recently introduced MILES. The techniques are expected to lead to greater training effectiveness by providing trainers with computer aids which, on the basis of real-time exercise data, identify incorrect tactical behaviors in the exercising units, and suggest required training directions. Such computer aiding requires software that can describe in detail a variety of tactical situations, and can facilitate the related representation of performance data. In essence, the software must have the ability to combine training data from diverse sources into an integrated model of the simulated engagement. The focus of the present work was on examining the feasibility of a rule-based, event-driven, computer model for the representation of small-unit combat engagements and for subsequent performance evaluation.