Uncertainty handling methodology implemented into the Analytic Platform

This report describes the uncertainty handling as part of the knowledge database for the Analytical Platform. After reviewing the decision-making under uncertainty and explaining the general methodology, we describe explicit applications of case-based methodology. A detailed construction of the knowledge database is presented and the definition of accident phases is addressed. In particular, our method provides an effective approach of offering categories even if the knowledge is missing or unclear. All the definitions and examples shown are closely connected with the Analytical Platform. The retrieval step, similarity function and k-nearest neighbors method, which are essential in our approach, are also described in detail for completeness. Moreover, we introduce the adaption and merging method, which enable us to reuse the strategies from historical events and fictitious scenarios and generate new strategies. Finally the outline of the work is listed in the conclusion and the future research plan is proposed.