A framework for the optimization of doctrine and systems in Army Air Defense units using predictive models of stochastic computer simulations

Abstract

This thesis presents a new methodology that can be used to address large-scale raids made up of different types of Theater Ballistic Missiles (TBMs) and Cruise Missiles (CMs) that attempt to overwhelm the Air Defense Artillery (ADA) systems at a particular location. The primary focus will be on how existing ADA systems can adjust their tactics in order to minimize the damage caused by threats that are not shot down and impact friendly forces. Nearly all the literature to date optimizes systems and tactics to reduce the number of leakers — threats not shot down — that impact the ground. However, simply counting the number of leakers does not adequately describe the effects to friendly forces. Instead, the first part of this thesis combines existing methods for external ballistics, concrete penetration, explosive cratering, and weapon blast and fragmentation damage in order to create an integrated program that can describe the damage to an airfield runway, infrastructure, and parked aircraft. The second part of this thesis focuses on modeling the ADA missile engagements using an accredited Department of Defense ADA simulation model called the Extended Air Defense Simulation (EADSIM). Both the airfield damage model and ADA simulation have run times ranging from minutes to hours. They are also stochastic; so a large number of runs are required for each input vector in order to properly understand the output range. In order to reduce the computation time to allow for later optimization, the methods of design of experiments and machine learning were used to create fast running models that predict the outputs of these simulations. The final part of this work uses these prediction algorithms to first optimize the TBM and CM fire plan, then optimize the ADA defense tactics, and finally optimize the ADA defense tactics with a new interceptor missile system.