The weak revision as a new change operator for belief bases

Abstract

This thesis starts by exploring background areas of study such as incomplete and uncertain information in relational database, belief change in artificial intelligence. It also reminds of how some types of incompleteness and uncertainly can be modeled through disjunctive databases. The paper describes the language being used in addition to some update operators. This thesis then brings to light a new theory in belief change called the weak revision. It first shows the need for such a new operator through a real life example. Second it introduces this new operator by defining its general behavior. Third it proves the correctness of the operator. the thesis will also include a comparative study of the weak revision with the AGM (Alchourron, Gardenfors and Makinson) postulates.