We investigate revision programming, a logic-based mechanism for describing changes in databases and enforcing certain type of integrity constraints. We show that revisions justified by an initial database and a revision program can be computed by a sequential execution of the rules of the program (with subsequent check of the applicability of the rules). In general, a program may determine none, exactly one or many justified revisions of a given initial database. We exhibit two classes of programs, safe and stratified, with the property that for every initial database a unique justified revision exists. We study the complexity of basic problems associated with justified revisions. Although the existence problems are NP.-complete, for safe and stratified programs justified revisions can be computed in polynomial time.
CITATION STYLE
Wiktor Marek, V., & Truszczyński, M. (1995). Revision programming, database updates and integrity constraints. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 893, pp. 368–382). Springer Verlag. https://doi.org/10.1007/3-540-58907-4_28
Mendeley helps you to discover research relevant for your work.