Knowledge acquisition for search and rescue planning

There is an increasing adoption of knowledge-level modelling within expert system development. However, it has had less impact in the generic areas of planning, scheduling and resource allocation. In this paper, we outline the development of a knowledge-level modelling approach within the domain of planning for search and rescue (SAR). Existing problem solving models for planning are almost exclusively derived from an analysis of the functional architectures of classic AI planners such as TWEAK and NONLIN. We argue that this makes their suitability for directly assisting knowledge acquisition questionable. Our approach makes a clear distinction between domain-derived knowledge-level models and those derived from computational architectures. We describe how the combination of these two types of models can achieve clear benefits within the course of KBS development. The paper includes extensive descriptions of the SAR domain, which illustrate the practical knowledge engineering problems that our approach attempts to address.