Probabilistic and Relational Inferences in Dynamic Environments (PRIDE)

Uncertainty on the modern battlefield makes mission planning and decision making extremely complex for field commanders. Mission planners can be aided by initial and ongoing situation assessment, future state estimation, assessment of the current plan"s probability of success, what-if analysis, and suggestions for decisions. At present, few automated tools exist to provide mission planners with the required information. An automated tool to assist mission planners must deal with both the inherent uncertainty and the complexity of the situation. Charles River Analytics proposes a system for Probabilistic and Relational Inferences in Dynamic Environments (PRIDE), an approach based on probabilistic relational models (PRMs). PRMs use probabilities to handle uncertainty while capturing the logical and relational structure of a situation to handle complexity. We base our implementation of PRMs on probabilistic programming (PP), which provides a powerful and flexible way to represent probabilistic models using the power of programming languages. We propose to design and develop a full-fledged prototype of our PRM and PP engine, to design and develop temporal reasoning capabilities, and to design and develop decision-theoretic reasoning capabilities.