Application of Neural Networks for Pattern Recognition in Logistics Data

In this Phase I SBIR, Accurate Automation Corporation will prototype key component technologies for a neural network-based Logistic Support Assistant that will assist Navy Supply Management personnel in planning and forecasting of logistics efforts. Fully developed as part of a Phase II effort, this system will help ensure smooth continuity of performance even when personnel are reassigned. There are two very significant benefits that this system will offer. The first is that is will learn the logistics requirements for different tasks by using neural networks to identify complex patterns and correlations within the database. It will be able to suggest appropriate logistics/resupply actions. The system will also be able to learn the time requirements for brining in the different supplies, and will assist in timely scheduling of requisitions. The second great advantage of this system is that is will learn how to respond to different types of conditions or requirements. This ability to learn the different "contexts" of operation will make the system very effective in support of a wide range of Navy operating conditions. This system will operate in a LAN-connected PC environment using commercially available object-oriented development tools to support the interface, and will process large data bases on a client-server basis.