Problems and tools to model uncertainty in expert- and decision support systems

Various sources of uncertainty are described in this contribution and ways and tools to model uncertainties in Expert Systems are discussed.

Abstract

Already in Decision Support Systems, which are more directed towards well-structured problem areas, uncertainties play a certain role. Modelling of uncertainty, however, is much more important in Expert Systems which are intended to support decisions, diagnosis etc. in ill-structured situations. Until recently uncertainty was primarily modelled via probabilities. In the meantime, one has become aware of the fact that uncertainties are of various types. To model uncertainty properly theories have been developed which supplement probability theory. Various sources of uncertainty are described in this contribution. Also ways and tools to model uncertainties in Expert Systems are discussed.