Capturing the Effects of C4I in a Campaign Context: A Practical Approach to Calibrating Analytical Simulations

The Air Force modeling and simulation community needs improved capabilities for measuring the effectiveness of command and control (C2) networks and processes in campaign-level analyses. A major modeling problem is capturing complex relationships between C2 network states and performance in a manner that is both traceable to empirical evidence and available in the timeframe required by an analytical simulation. Neural network technology offers a model abstraction technique with the potential to meet these criteria. Here campaign-level cause and effect relationships are captured using a custom neural network to help determine the effects of C2 network states on military operations. This neural network sub-model was then integrated into the Air Force 5 THUNDER simulation as a proof-of-concept. The resulting simulation showed sensitivity to state changes as provided by the neural network. In contrast to other C2 abstraction techniques, this neural network implementation was more credible because its values were directly derived from, and therefore more clearly traceable to, source data.