In concert with the I/ITSEC ‘99 theme, Synthetic Solutions for the 21st Century, this paper explores a variety of ways in which neural networks, synthetic models of human cognition, can be used to improve performance in a distributed simulation exercise. Specifically, it examines the use of neural networks in semi-automated forces (SAF) systems as a means of reducing network bandwidth and processing requirements. To address the first performance measure, reduction of network bandwidth requirements, this research investigates the use of neural networks in lieu of the current, Newtonian, DIS dead-reckoning models. While this concept is demonstrated in a SAF system, it is extendible to other types of players (e.g., manned modules or live/embedded systems) in a distributed simulation. To address the second performance measure, reduction of processing requirements, this research considers the use of neural networks in lieu of SAF behavioral models. This concept does not extend beyond SAF systems.
This paper motivates the need for this research by reviewing how SAF systems work and why they are limited by bandwidth and processor constraints. It also introduces the theory behind the neural networks' architecture and training algorithms as well as the specifics of how the networks were developed for this investigation. Lastly, it illustrates how the networks were integrated with SAF software, defines the networks' performance measures, presents the results of the scenarios considered in this investigation, and offers directions for future work.