Methodology for Conducting Composite Behavior Model Verification in a Combat Simulation

摘要

The United States Army's One Semi-Automated Forces (OneSAF) Objective System (OOS) is the next generation of Army high resolution combat models. Its development has leveraged the ever-increasing computing power available today to represent highly complex battlefield phenomena, particularly human behavior. In the fall of 2005, the Product Manager (PM) OneSAF asked us to conduct a verification of the orderable, composite behavior models within OOS. As a result, we developed and executed a unique process to verify those behaviors under tight resource constraints. Our methodology and test designs allowed us to evaluate the behaviors thoroughly with a minimum number of scenarios. Based upon our work, we were able to verify a number of composite behaviors and to provide valuable feedback to the PM OneSAF. In this paper, we provide an overview of the problem, a description of the methodology we developed, and a summary of our challenges and results.