Modeling dynamic tactical behaviors in COMBATXXI using Hierarchical Task Networks

Abstract

Defining accurate representations of group behaviors in simulations is an expensive, time-consuming task. One reason for this is that previously produced behaviors are often not reusable within other scenarios or simulations. Using Hierarchical Task Networks (HTNs) to model military behaviors is a promising technique for addressing this problem. HTNs provide a methodology for linking tactical behaviors, and offer a potential system for representing the military decision-making process at the tactical level. This thesis investigates the use of HTNs within the COMBATXXI model. COMBATXXI provides military planners a detailed representation of combat operations, and supports analysis efforts by providing insights into the effectiveness of weapon systems, unit organizations, and tactics. The use of HTNs within COMBATXXI is a relatively new concept; many aspects of HTN implementation have not been researched in depth. Work in this thesis involved development and testing of HTNs capable of executing a security formation behavior, and coordinating the execution of other ground combat related behaviors. The HTN-controlled behaviors were demonstrated in a simulated version of a United States Marine Corps live fire training range. The composable and dynamic aspects of these behaviors eased the scenario development process and added tactical realism to the test scenario.