Belief Revision for Deductive Databases.
Antonis C. Kakas · UPCommons institutional repository (Universitat Politècnica de Catalunya) · 1990
We study the problem of evolution of knowledge in a deductive database by formulating this in terms of belief revision and using abductive (hypothetical) reasoning. The problem of the existence of multiple different solutions to update requests on such databases is addressed by imposing naturally related preference orderings on these and choosing the preferable ones. This has highlighted and clarified further the role on integrity constraints in updating databases. A Truth Maintenance system associated to our abductive approach to belief revision is developed for recovering from wrong choices of solutions in the light of new information and for maintaining consistency as the database evolves.