Measurement system comparison approaches for military queuing model validation

Operationally validating military and defense queuing models requires the rigorous assessment of agreement between functional responses of the model and the system or process of interest. This article provides and contextualizes two distance-based validation methods for operationally validating complex transient-phase military and defense queuing models. The limits of agreement approach and the probability of agreement approach, both developed within the measurement system comparison literature, are contextualized through an illustrative M/M/1 queuing model application derived from the military air traffic control domain. The limits of agreement approach characterizes agreement through an evaluation of the differences between observations and predictions on the same entity, while the probability of agreement approach uses a mixed-effects structural model to characterize the relationship between observations and predictions. The procedures and results of both methods are juxtaposed against common Boolean-based statistical methods and are used to establish global predictive capabilities useful for calibrating military and defense queuing models over multiple settings of controllable input.