Implementation of an Automated Fire Support Planner

Abstract

Although the employment of fire support is a staple of modern military doctrine, today's constructive combat simulations depend on meticulous human input to generate any appropriate fire support plans. This status quo can be improved through AI techniques. We implement models of tactical risk, reduction of risk, and suppression effects in a representative combat simulation, as well as a greedy fire support planning algorithm that leverages these concepts. The algorithm is theoretically non-optimal, but testing shows that the resulting fire support plans are effective at improving simulated combat results and have some realistic emergent properties. The practical running time of the planner is less than 20 seconds for a company-sized unit, including navigation graph setup. The planner's best-first approach scales naturally in more time-constrained environments.