Prospective Updating of Theories with Preferences

This work focuses on updating and revising theories with preferences within the context of logic programming. This aim is achieved by first exploiting preferences to reduce the number of abductive extensions of the initial theory, then by using the observations to confirm or deny the abduced hypotheses. In case the observations disconfirm the preferred abduced hypotheses, a revision process is launched in order to revise the theory’s preferences with respect to the new acquired observations. A methodology for model-based diagnosis is also proffered as an application of preferential theory revision, using observations to disambiguate among possible relevant revision scenarios. 1