Applications of Artificial Neural Networks to combat simulations

As a result of reduced budgets and personnel levels, the Department of Defense has increased reliance on combat simulations for such diverse areas as training, testing, planning, and analysis. Each area has its own set of needs, goals, and objectives for designing future generations of combat simulation models. However, budget constraints alone mandate the development of multipleuse combat models. The bottom line is that future generations of combat models need to be faster, have higher fidelity and larger scale than current models. Research into emerging technologies for approaches to make computer simulations more effective and efficient is an essential ingredient to developing successful future generations of combat models. One emerging technology that has such potential is Artificial Neural Networks (ANN). Potential applications of ANN to combat simulation modeling are discussed. The main results of the authorʹs dissertation Artificial Neural Network Metamodels of Stochastic Computer Simulations [1] are discussed along with the ramifications on combat modeling and recommendations for future research.