On the deployment sequence optimality in model predictive control of military operations

Combat tasks can be planned and executed via controlling their risk of failure. The model predictive task commander (MPTC) described previously calculates the minimal effective force to be deployed in each mission of a task and dynamically adjusts it based on real time battle damage assessment data to maintain the desired level of risk. The risk comes from two main sources of uncertainty present in combat, namely the random effects of weapons and the unknown enemy intent. The MPTC uses probabilistic Markov models to capture the former, and game-theoretic optimization to cope with the latter. The paper reports on experiments suggesting that the so called initial deployment sequences are either optimal or almost optimal strategies in the game theoretic sense.