Preferential Default Reasoning on the Semantic Web
|Preferential Default Reasoning on the Semantic Web|
|Student name:||student name::Rafael Kiesel|
|Second supervisor:||Erman Acar|
On the Semantic Web one is sometimes faced with the task of handling inconsistent knowledge. One approach to deal with such inconsistency is to replace some of the given axioms, which usually always have to hold, with defaults which can only be used to derive new knowledge if it does not lead to inconsistency. A recent approach provides model-theoretic semantics for free defaults in the context of description logics. While satisfiability and entailment in this framework are decidable, the complexity of checking them is too high to be feasible for practical use. This thesis extends the semantics to the case where the defaults are ordered with a preference relation with respect to fulfilment. The focus is then on whether the introduction of such an ordering reduces the complexity of satisfiability and entailment checks and whether it increases the amount of useful information derivable.