Evaluating complex systems when numerical information is sparse

Analyzing complex systems for which there is insufficient information for a complete quantitative characterization is a common problem encountered in military and research applications. As a result of repeated experience with this situation, we developed an approach that uses integrated logic modeling and approximate reasoning to make sophisticated and complicated predictions and decisions about systems with significant gaps in quantitative understanding. We describe how a process tree can be used to gain better understanding of complex physical or operational processes. We show how this understanding can be used to develop an approximate reasoning decision model that efficiently uses experience and expert judgment to make reasonable decisions.