Automated war gaming: an overview of the RAND Strategy Assessment Center
This paper outlines the origins, objectives, and architecture of the RAND Strategy Assessment Center (RSAC), a highly ambitious multiyear research program to develop a system and methodology combining the systematic reproducibility of analytic modeling/simulation with the contextual and behavioral richness of war gaming. The system integrates an adaptive simulation model of military operations with three rule-based artificial intelligence agents that represent Red, Blue, and nonsuperpower behavior. The paper describes the hierarchical decisionmaking structure within which superpower decisions are made as well as the interaction among the rule-based agents and the simulation model. It describes the high-level decisionmaking procedure, and provides an example to illustrate the mechanisms used at the lowest decisionmaking level. It considers the issue of human intervention and outlines the next steps in the RSAC system's development and transfer to the U.S. government.