An algorithm for belief revision
Renata Wassermann · 2000
In this paper we show that a particular construction of belief revision operator is equivalent to the standard method for computing consistency-based diagnosis. We show how a diagnosis problem can be translated into a problem of belief revision and show how kernel constructions for revision operators can be used for computing diagnosis. We also show how Reiter's algorithm for computing diagnosis can be adapted for being used in belief revision. 1 Introduction Belief revision (for an overview, see [ Gardenfors, 1988; Gardenfors and Rott, 1995 ] ) deals with the problem of how to accommodate new assertions into an existent body of knowledge. Traditionally, the body of knowledge is represented by a belief set, a set of formulas closed under logical implication. Instead of belief sets we are going to use belief bases to represent belief states. A belief base is a set not closed under logical consequence [ Fuhrmann, 1991; Hansson, 1989; Nebel, 1992 ] . For every belief base B...