Some experiments with agent-based combat models

The DoD has become increasingly reliant on models and their outputs. Given this reliance on models and their outputs, one might assume the models are accurate and faithfully represent the particular system of interest. Unfortunately, this is not the case, particularly when the systems of interest involve key elements of combat uncertainty. Agent-based simulation however potentially provides a means to capture and model the goal-directed behavior of combatants provided one can first build the simulations and then interpret the output. This paper presents two agent-based combat simulations, each focused on examining strategic effects. Both models and accompanying experiments are described. The emergent behavior of these agent models is then examined from a combat analysis perspective with extremely interesting results. This work breaks new ground in how to use agent-based simulations to gain insight into warfare dynamics.