Rigorous Framework for Distributed Processing with Dynamic Stochastic Information

The vast and ever-increasing amount of networked information available creates a growing need for information processing systems that distribute decision making rather than centralize data. But distributed decision systems are notoriously hard to analyze. One reason is that the complexity of the design process encourages the use of ad-hoc methods, so it can easily become impossible to say how well a system actually performs vs. the best that could be done. Prism Informatix proposes to address the need for an optimal framework by extending the rigorous and established theory of distributed hypothesis testing networks. The basic framework has aroused considerable interest because it provides an optimal, analytically tractable, and highly intuitive approach to modeling distributed cooperative (team) decision making under uncertainty. To develop the theory into a practical information processing framework, we modify and extend it to handle real-world stochastic information dynamics, but in a manner that preserves the mathematical structure necessary to support analysis. We implement the new framework in a computational testbed where instrumented analysis studies allow its feasibility to be established.