Adaptive Model-based Adversarial Reasoning System (AMARS) for Enhanced Synthetic Battlespaces

Recent military operations have demonstrated the use by adversaries of non-traditional or asymmetric military tactics to offset US military might. This issue has thus come into the forefront of national security. Rogue nations with links to trans-national terrorists have created a highly unpredictable and potential dangerous environment for US military operations. In his testimony to the US Senate, Vice Admiral Wilson, Director of the Defense Intelligence Agency (DIA) identifies several characteristics of these threats including extremism in beliefs, global in nature, non-state oriented, and highly networked and adaptive, thus making these adversaries less vulnerable to conventional military approaches. Additionally, US forces must also contend with more traditional state-based threats that are further evolving their military fighting strategies and capabilities. What are needed are solutions to assist our forces in the prosecution of operations against these diverse threat types and their atypical strategies and tactics. Here, we propose an Adaptive Model-based Adversarial Reasoning System (AMARS) that supports both training and simulation based acquisition requirements for effective responses to enemy asymmetric tactics and strategies. The proposed system merges model-based reasoning about individual, group, and organizational behavior within a genetic algorithm-based model adaptation and development environment.