Examining How Standby Assets Impact Optimal Dispatching Decisions within a Military Medical Evacuation System Via a Markov Decision Process Model

The Army medical evacuation (MEDEVAC) system ensures proper medical treatment is readily available to wounded soldiers on the battlefield. The objective of this research is to determine which MEDEVAC unit to task to an incoming 9-line MEDEVAC request and where to station a single standby unit to maximize patient survivability. A discounted, infinite-horizon continuous-time Markov decision process model is formulated to examine this problem. We design, develop, and test an approximate dynamic programming (ADP) technique that leverages a least squares policy evaluation value function approximation scheme within an approximate policy iteration algorithmic framework to solve practical-sized problem instances. A computational example is applied to a synthetically generated scenario in Iraq. The optimal policy and ADP-generated policies are compared to a commonly practiced (i.e., myopic) policy. Examining multiple courses of action determines the best location for the standby MEDEVAC unit, and sensitivity analysis reveals that the optimal and ADP policies are robust to standby unit mission preparation times. The best performing ADP-generated policy is within 2.62% of the optimal policy regarding a patient survivability metric. Moreover, the ADP policy outperforms the myopic policy in all cases, indicating the currently practiced dispatching policy can be improved