Many compelling EIS/ESS applications in the military and in industry remain infeasible due to limitations in computer technology. Many interesting problems are too large or too complex for practical solutions by traditional approaches. The complexity may arise from unknown, non-linear, or time-varying relationships between variables. Neural network technology has demonstrated significant capabilities in similar applications. Neural nets "learn by example", providing adaptive capabilities unlimited by the problem size or the nature of relationships between variables. These capabilities merit investigation in the context of army EIS/ESS applications. We will explore potential applications in the following areas: military procurement, logistics, operations, intelligence, and training. We will qualify these applications on the basis of technical feasibility, resource requirements, and benefits to the army and the intended user. We hope to then prepare phase ii proposals for implementation of the most compelling applications.
