Model Predictive Risk Control of Military Operations

Work reported in this paper was done as part of the DARPA Joint Force Air Component Commander (JFACC) project. Its objective was to investigate the possibility of improving the stability and agility of military operations using the concepts of modern control and game theories within the framework of state-of-the-art computer science and operations research. Military operations are always defined and executed within the context of a command and control (C2) hierarchy. The questions we have studied at the Task level were twofold. What is the minimal effective task force size and composition for a given task, and how to guarantee successful task execution in the presence of uncertainties in combat, both due to random effects of weapons and an intelligent adversary? At the Task Group level, we asked how to optimally allocate and schedule available resources to satisfy the force size requirements for as many concurrent tasks as possible. We have developed probabilistic Markov models of combat dynamics, and then used them to build the Model Predictive Task Commander and Model Predictive Resource Allocator systems, which are briefly described in the paper along with experimental results showing their performance in simulated battles.